From 6242362af01671d9a7f38a45317c8296c654e012 Mon Sep 17 00:00:00 2001 From: 0xrushi <6279035+0xrushi@users.noreply.github.com> Date: Fri, 6 Feb 2026 23:23:57 -0500 Subject: [PATCH 01/11] Enhance model management and transcription provider setup - Added `add_or_update_model` method in `ConfigManager` to facilitate adding or updating models in the configuration. - Updated `ChronicleSetup` to support a new OpenAI-Compatible transcription provider, allowing users to configure custom endpoints and API keys. - Enhanced user prompts for API base URL and model name during setup, improving the configuration experience. - Introduced unit tests for the new model management functionality and transcription provider setup, ensuring robust validation of the changes. - Improved Docker configurations for ASR services, including support for customizable CUDA versions and DNS settings. --- backends/advanced/init.py | 82 ++++- .../tests/test_setup_llm_custom_provider.py | 286 ++++++++++++++++++ .../tests/test_transcription_url_config.py | 183 +++++++++++ config_manager.py | 19 ++ extras/asr-services/docker-compose.yml | 5 + .../providers/vibevoice/Dockerfile | 4 +- .../tests/test_cuda_version_config.py | 220 ++++++++++++++ 7 files changed, 797 insertions(+), 2 deletions(-) create mode 100644 backends/advanced/tests/test_setup_llm_custom_provider.py create mode 100644 backends/advanced/tests/test_transcription_url_config.py create mode 100644 extras/asr-services/tests/test_cuda_version_config.py diff --git a/backends/advanced/init.py b/backends/advanced/init.py index aad7ff0e..e70a1369 100644 --- a/backends/advanced/init.py +++ b/backends/advanced/init.py @@ -309,7 +309,8 @@ def setup_llm(self): choices = { "1": "OpenAI (GPT-4, GPT-3.5 - requires API key)", "2": "Ollama (local models - runs locally)", - "3": "Skip (no memory extraction)" + "3": "OpenAI-Compatible (custom endpoint - Groq, Together AI, LM Studio, etc.)", + "4": "Skip (no memory extraction)" } choice = self.prompt_choice("Which LLM provider will you use?", choices, "1") @@ -347,6 +348,85 @@ def setup_llm(self): self.console.print("[yellow][WARNING][/yellow] Make sure Ollama is running and models are pulled") elif choice == "3": + self.console.print("[blue][INFO][/blue] OpenAI-Compatible custom endpoint selected") + self.console.print("This works with any provider that exposes an OpenAI-compatible API") + self.console.print("(e.g., Groq, Together AI, LM Studio, vLLM, etc.)") + self.console.print() + + # Prompt for base URL (required) + base_url = self.prompt_value( + "API Base URL (e.g., https://api.groq.com/openai/v1)", "" + ) + if not base_url: + self.console.print("[yellow][WARNING][/yellow] No base URL provided - skipping custom LLM setup") + else: + # Prompt for API key + api_key = self.prompt_with_existing_masked( + prompt_text="API Key (leave empty if not required)", + env_key="CUSTOM_LLM_API_KEY", + placeholders=['your_custom_llm_api_key_here'], + is_password=True, + default="" + ) + if api_key: + self.config["CUSTOM_LLM_API_KEY"] = api_key + + # Prompt for model name (required) + model_name = self.prompt_value( + "LLM Model name (e.g., llama-3.1-70b-versatile)", "" + ) + if not model_name: + self.console.print("[yellow][WARNING][/yellow] No model name provided - skipping custom LLM setup") + else: + # Create LLM model entry + llm_model = { + "name": "custom-llm", + "description": "Custom OpenAI-compatible LLM", + "model_type": "llm", + "model_provider": "openai", + "api_family": "openai", + "model_name": model_name, + "model_url": base_url, + "api_key": "${oc.env:CUSTOM_LLM_API_KEY,''}", + "model_params": { + "temperature": 0.2, + "max_tokens": 2000 + }, + "model_output": "json" + } + self.config_manager.add_or_update_model(llm_model) + + # Prompt for optional embedding model + embedding_model_name = self.prompt_value( + "Embedding model name (leave empty to use Ollama local-embed)", "" + ) + + if embedding_model_name: + embed_model = { + "name": "custom-embed", + "description": "Custom OpenAI-compatible embeddings", + "model_type": "embedding", + "model_provider": "openai", + "api_family": "openai", + "model_name": embedding_model_name, + "model_url": base_url, + "api_key": "${oc.env:CUSTOM_LLM_API_KEY,''}", + "embedding_dimensions": 1536, + "model_output": "vector" + } + self.config_manager.add_or_update_model(embed_model) + self.config_manager.update_config_defaults({"llm": "custom-llm", "embedding": "custom-embed"}) + self.console.print("[green][SUCCESS][/green] Custom LLM and embedding configured in config.yml") + self.console.print("[blue][INFO][/blue] Set defaults.llm: custom-llm") + self.console.print("[blue][INFO][/blue] Set defaults.embedding: custom-embed") + else: + self.config_manager.update_config_defaults({"llm": "custom-llm", "embedding": "local-embed"}) + self.console.print("[green][SUCCESS][/green] Custom LLM configured in config.yml") + self.console.print("[blue][INFO][/blue] Set defaults.llm: custom-llm") + self.console.print("[blue][INFO][/blue] Set defaults.embedding: local-embed (Ollama)") + self.console.print("[yellow][WARNING][/yellow] Make sure Ollama is running for embeddings") + + elif choice == "4": self.console.print("[blue][INFO][/blue] Skipping LLM setup - memory extraction disabled") # Disable memory extraction in config.yml self.config_manager.update_memory_config({"extraction": {"enabled": False}}) diff --git a/backends/advanced/tests/test_setup_llm_custom_provider.py b/backends/advanced/tests/test_setup_llm_custom_provider.py new file mode 100644 index 00000000..33014383 --- /dev/null +++ b/backends/advanced/tests/test_setup_llm_custom_provider.py @@ -0,0 +1,286 @@ +""" +Unit tests for OpenAI-Compatible custom LLM provider setup. + +Tests the wizard's choice "3" (OpenAI-Compatible) in setup_llm(), +including model creation in config.yml and defaults updates. +""" + +import os +import sys +import tempfile +import shutil +from pathlib import Path +from unittest.mock import patch, MagicMock + +import pytest +import yaml + +# Add repo root to path for imports +sys.path.insert(0, str(Path(__file__).resolve().parent.parent.parent.parent)) +from config_manager import ConfigManager + + +@pytest.fixture +def temp_config_dir(): + """Create a temporary directory with a minimal config.yml.""" + tmpdir = tempfile.mkdtemp() + config_dir = Path(tmpdir) / "config" + config_dir.mkdir() + + config = { + "defaults": { + "llm": "openai-llm", + "embedding": "openai-embed", + "stt": "stt-deepgram", + }, + "models": [ + { + "name": "openai-llm", + "description": "OpenAI GPT-4o-mini", + "model_type": "llm", + "model_provider": "openai", + "api_family": "openai", + "model_name": "gpt-4o-mini", + "model_url": "https://api.openai.com/v1", + "api_key": "${oc.env:OPENAI_API_KEY,''}", + "model_params": {"temperature": 0.2, "max_tokens": 2000}, + "model_output": "json", + }, + { + "name": "local-embed", + "description": "Local embeddings via Ollama", + "model_type": "embedding", + "model_provider": "ollama", + "api_family": "openai", + "model_name": "nomic-embed-text:latest", + "model_url": "http://localhost:11434/v1", + "api_key": "${oc.env:OPENAI_API_KEY,ollama}", + "embedding_dimensions": 768, + "model_output": "vector", + }, + ], + "memory": {"provider": "chronicle"}, + } + + config_path = config_dir / "config.yml" + with open(config_path, "w") as f: + yaml.dump(config, f, default_flow_style=False, sort_keys=False) + + yield tmpdir + + shutil.rmtree(tmpdir) + + +@pytest.fixture +def config_manager(temp_config_dir): + """Create a ConfigManager pointing to the temp config.""" + return ConfigManager(service_path=None, repo_root=Path(temp_config_dir)) + + +class TestAddOrUpdateModel: + """Tests for ConfigManager.add_or_update_model().""" + + def test_add_new_model(self, config_manager): + """add_or_update_model() should append a new model when name doesn't exist.""" + new_model = { + "name": "custom-llm", + "description": "Custom OpenAI-compatible LLM", + "model_type": "llm", + "model_provider": "openai", + "api_family": "openai", + "model_name": "llama-3.1-70b-versatile", + "model_url": "https://api.groq.com/openai/v1", + "api_key": "${oc.env:CUSTOM_LLM_API_KEY,''}", + "model_params": {"temperature": 0.2, "max_tokens": 2000}, + "model_output": "json", + } + + config_manager.add_or_update_model(new_model) + + config = config_manager.get_full_config() + model_names = [m["name"] for m in config["models"]] + assert "custom-llm" in model_names + + added = next(m for m in config["models"] if m["name"] == "custom-llm") + assert added["model_name"] == "llama-3.1-70b-versatile" + assert added["model_url"] == "https://api.groq.com/openai/v1" + assert added["model_type"] == "llm" + + def test_update_existing_model(self, config_manager): + """add_or_update_model() should replace an existing model with the same name.""" + # First add + model_v1 = { + "name": "custom-llm", + "model_type": "llm", + "model_name": "model-v1", + "model_url": "https://example.com/v1", + } + config_manager.add_or_update_model(model_v1) + + # Then update + model_v2 = { + "name": "custom-llm", + "model_type": "llm", + "model_name": "model-v2", + "model_url": "https://example.com/v2", + } + config_manager.add_or_update_model(model_v2) + + config = config_manager.get_full_config() + custom_models = [m for m in config["models"] if m["name"] == "custom-llm"] + assert len(custom_models) == 1 + assert custom_models[0]["model_name"] == "model-v2" + assert custom_models[0]["model_url"] == "https://example.com/v2" + + def test_add_model_to_empty_models_list(self, temp_config_dir): + """add_or_update_model() should create models list if it doesn't exist.""" + config_path = Path(temp_config_dir) / "config" / "config.yml" + with open(config_path, "w") as f: + yaml.dump({"defaults": {"llm": "openai-llm"}}, f) + + cm = ConfigManager(service_path=None, repo_root=Path(temp_config_dir)) + cm.add_or_update_model({"name": "test-model", "model_type": "llm"}) + + config = cm.get_full_config() + assert "models" in config + assert len(config["models"]) == 1 + assert config["models"][0]["name"] == "test-model" + + +class TestSetupLlmCustomProvider: + """Tests for the custom LLM provider flow in setup_llm().""" + + def _make_setup(self, temp_config_dir): + """Create a ChronicleSetup instance pointing at the temp config.""" + # We need to mock the ChronicleSetup constructor's checks + # Instead, we test the logic by calling config_manager directly, + # simulating what setup_llm() choice "3" does. + return ConfigManager(service_path=None, repo_root=Path(temp_config_dir)) + + def test_custom_llm_model_added_to_config(self, config_manager): + """Selecting custom provider should create correct model entry.""" + llm_model = { + "name": "custom-llm", + "description": "Custom OpenAI-compatible LLM", + "model_type": "llm", + "model_provider": "openai", + "api_family": "openai", + "model_name": "llama-3.1-70b-versatile", + "model_url": "https://api.groq.com/openai/v1", + "api_key": "${oc.env:CUSTOM_LLM_API_KEY,''}", + "model_params": {"temperature": 0.2, "max_tokens": 2000}, + "model_output": "json", + } + + config_manager.add_or_update_model(llm_model) + + config = config_manager.get_full_config() + model = next(m for m in config["models"] if m["name"] == "custom-llm") + assert model["model_provider"] == "openai" + assert model["api_family"] == "openai" + assert model["model_name"] == "llama-3.1-70b-versatile" + assert model["model_url"] == "https://api.groq.com/openai/v1" + assert model["api_key"] == "${oc.env:CUSTOM_LLM_API_KEY,''}" + assert model["model_params"]["temperature"] == 0.2 + assert model["model_output"] == "json" + + def test_custom_llm_and_embedding_model_added(self, config_manager): + """Both LLM and embedding models should be created when embedding model is provided.""" + llm_model = { + "name": "custom-llm", + "model_type": "llm", + "model_provider": "openai", + "api_family": "openai", + "model_name": "llama-3.1-70b-versatile", + "model_url": "https://api.groq.com/openai/v1", + "api_key": "${oc.env:CUSTOM_LLM_API_KEY,''}", + "model_params": {"temperature": 0.2, "max_tokens": 2000}, + "model_output": "json", + } + embed_model = { + "name": "custom-embed", + "description": "Custom OpenAI-compatible embeddings", + "model_type": "embedding", + "model_provider": "openai", + "api_family": "openai", + "model_name": "text-embedding-3-small", + "model_url": "https://api.groq.com/openai/v1", + "api_key": "${oc.env:CUSTOM_LLM_API_KEY,''}", + "embedding_dimensions": 1536, + "model_output": "vector", + } + + config_manager.add_or_update_model(llm_model) + config_manager.add_or_update_model(embed_model) + + config = config_manager.get_full_config() + model_names = [m["name"] for m in config["models"]] + assert "custom-llm" in model_names + assert "custom-embed" in model_names + + embed = next(m for m in config["models"] if m["name"] == "custom-embed") + assert embed["model_type"] == "embedding" + assert embed["model_name"] == "text-embedding-3-small" + assert embed["embedding_dimensions"] == 1536 + + def test_custom_llm_without_embedding_falls_back_to_local_embed(self, config_manager): + """defaults.embedding should be local-embed when no custom embedding is provided.""" + llm_model = { + "name": "custom-llm", + "model_type": "llm", + "model_name": "some-model", + "model_url": "https://api.example.com/v1", + } + config_manager.add_or_update_model(llm_model) + config_manager.update_config_defaults({"llm": "custom-llm", "embedding": "local-embed"}) + + defaults = config_manager.get_config_defaults() + assert defaults["llm"] == "custom-llm" + assert defaults["embedding"] == "local-embed" + + def test_custom_llm_updates_defaults_with_embedding(self, config_manager): + """defaults.llm and defaults.embedding should be updated correctly with custom embed.""" + config_manager.update_config_defaults({"llm": "custom-llm", "embedding": "custom-embed"}) + + defaults = config_manager.get_config_defaults() + assert defaults["llm"] == "custom-llm" + assert defaults["embedding"] == "custom-embed" + + def test_custom_llm_api_key_env_reference(self, config_manager): + """API key should use env var reference in config.yml model.""" + llm_model = { + "name": "custom-llm", + "model_type": "llm", + "model_name": "some-model", + "model_url": "https://api.example.com/v1", + "api_key": "${oc.env:CUSTOM_LLM_API_KEY,''}", + } + config_manager.add_or_update_model(llm_model) + + config = config_manager.get_full_config() + model = next(m for m in config["models"] if m["name"] == "custom-llm") + assert model["api_key"] == "${oc.env:CUSTOM_LLM_API_KEY,''}" + + def test_existing_models_preserved_after_adding_custom(self, config_manager): + """Adding a custom model should not remove existing models.""" + config_before = config_manager.get_full_config() + original_count = len(config_before["models"]) + + config_manager.add_or_update_model({ + "name": "custom-llm", + "model_type": "llm", + "model_name": "test-model", + "model_url": "https://example.com/v1", + }) + + config_after = config_manager.get_full_config() + assert len(config_after["models"]) == original_count + 1 + # Original models still present + model_names = [m["name"] for m in config_after["models"]] + assert "openai-llm" in model_names + assert "local-embed" in model_names + assert "custom-llm" in model_names + + +if __name__ == "__main__": + pytest.main([__file__, "-v"]) diff --git a/backends/advanced/tests/test_transcription_url_config.py b/backends/advanced/tests/test_transcription_url_config.py new file mode 100644 index 00000000..b9fcf9bc --- /dev/null +++ b/backends/advanced/tests/test_transcription_url_config.py @@ -0,0 +1,183 @@ +""" +Unit tests for transcription service URL configuration. + +Tests the fix for the double http:// prefix issue where environment variables +containing protocol prefixes were incorrectly combined with hardcoded prefixes +in config.yml. +""" + +import os +import pytest +from unittest.mock import patch, MagicMock +from omegaconf import OmegaConf + + +class TestTranscriptionURLConfiguration: + """Test transcription service URL configuration and parsing.""" + + def test_vibevoice_url_without_http_prefix(self): + """Test that VIBEVOICE_ASR_URL without http:// prefix works correctly.""" + # Simulate config.yml template: http://${oc.env:VIBEVOICE_ASR_URL} + config_template = {"model_url": "http://${oc.env:VIBEVOICE_ASR_URL,host.docker.internal:8767}"} + + with patch.dict(os.environ, {"VIBEVOICE_ASR_URL": "host.docker.internal:8767"}): + resolved = OmegaConf.create(config_template) + resolved = OmegaConf.to_container(resolved, resolve=True) + + assert resolved["model_url"] == "http://host.docker.internal:8767" + assert "http://http://" not in resolved["model_url"] + + def test_vibevoice_url_with_http_prefix_causes_double_prefix(self): + """Test that VIBEVOICE_ASR_URL WITH http:// causes double prefix (bug scenario).""" + config_template = {"model_url": "http://${oc.env:VIBEVOICE_ASR_URL,host.docker.internal:8767}"} + + # This is the BUG scenario - env var already has http:// + with patch.dict(os.environ, {"VIBEVOICE_ASR_URL": "http://host.docker.internal:8767"}): + resolved = OmegaConf.create(config_template) + resolved = OmegaConf.to_container(resolved, resolve=True) + + # This demonstrates the bug + assert resolved["model_url"] == "http://http://host.docker.internal:8767" + assert "http://http://" in resolved["model_url"] + + def test_vibevoice_url_default_fallback(self): + """Test that default fallback works when VIBEVOICE_ASR_URL is not set.""" + config_template = {"model_url": "http://${oc.env:VIBEVOICE_ASR_URL,host.docker.internal:8767}"} + + # No VIBEVOICE_ASR_URL set - should use default + with patch.dict(os.environ, {}, clear=True): + resolved = OmegaConf.create(config_template) + resolved = OmegaConf.to_container(resolved, resolve=True) + + assert resolved["model_url"] == "http://host.docker.internal:8767" + + def test_parakeet_url_configuration(self): + """Test that PARAKEET_ASR_URL follows same pattern.""" + config_template = {"model_url": "http://${oc.env:PARAKEET_ASR_URL,172.17.0.1:8767}"} + + # Correct format - without http:// prefix + with patch.dict(os.environ, {"PARAKEET_ASR_URL": "host.docker.internal:8767"}): + resolved = OmegaConf.create(config_template) + resolved = OmegaConf.to_container(resolved, resolve=True) + + assert resolved["model_url"] == "http://host.docker.internal:8767" + assert "http://http://" not in resolved["model_url"] + + def test_url_parsing_removes_double_slashes(self): + """Test that URL with double http:// causes connection failures.""" + from urllib.parse import urlparse + + # Valid URL + valid_url = "http://host.docker.internal:8767/transcribe" + parsed_valid = urlparse(valid_url) + assert parsed_valid.scheme == "http" + assert parsed_valid.netloc == "host.docker.internal:8767" + + # Invalid URL with double prefix + invalid_url = "http://http://host.docker.internal:8767/transcribe" + parsed_invalid = urlparse(invalid_url) + # urlparse treats "http:" as the netloc which causes DNS failures + assert parsed_invalid.scheme == "http" + assert parsed_invalid.netloc == "http:" # Invalid netloc causes "Name or service not known" + assert parsed_invalid.netloc != "host.docker.internal:8767" + + +class TestProviderSegmentsConfiguration: + """Test use_provider_segments configuration for different providers.""" + + def test_use_provider_segments_default_false(self): + """Test that use_provider_segments defaults to false.""" + config = OmegaConf.create({ + "backend": { + "transcription": {} + } + }) + + use_segments = config.backend.transcription.get("use_provider_segments", False) + assert use_segments is False + + def test_use_provider_segments_explicit_true(self): + """Test that use_provider_segments can be enabled.""" + config = OmegaConf.create({ + "backend": { + "transcription": { + "use_provider_segments": True + } + } + }) + + assert config.backend.transcription.use_provider_segments is True + + def test_vibevoice_should_use_provider_segments(self): + """ + Test that VibeVoice provider should have use_provider_segments=true + since it provides diarized segments. + """ + # VibeVoice provides segments with speaker diarization + vibevoice_capabilities = ["segments", "diarization"] + + # When provider has both capabilities, use_provider_segments should be true + has_diarization = "diarization" in vibevoice_capabilities + has_segments = "segments" in vibevoice_capabilities + + should_use_segments = has_diarization and has_segments + assert should_use_segments is True + + +class TestModelRegistryURLResolution: + """Test model registry URL resolution with environment variables.""" + + def test_model_url_resolution_with_env_var(self): + """Test that model URLs resolve correctly from environment.""" + config_template = """ + defaults: + stt: stt-vibevoice + models: + - name: stt-vibevoice + model_type: stt + model_provider: vibevoice + model_url: http://${oc.env:VIBEVOICE_ASR_URL,host.docker.internal:8767} + """ + + with patch.dict(os.environ, {"VIBEVOICE_ASR_URL": "host.docker.internal:8767"}): + config = OmegaConf.create(config_template) + resolved = OmegaConf.to_container(config, resolve=True) + + vibevoice_model = resolved["models"][0] + assert vibevoice_model["model_url"] == "http://host.docker.internal:8767" + + def test_multiple_asr_providers_url_resolution(self): + """Test that multiple ASR providers can use different URL patterns.""" + config_template = { + "models": [ + { + "name": "stt-vibevoice", + "model_url": "http://${oc.env:VIBEVOICE_ASR_URL,host.docker.internal:8767}" + }, + { + "name": "stt-parakeet", + "model_url": "http://${oc.env:PARAKEET_ASR_URL,172.17.0.1:8767}" + }, + { + "name": "stt-deepgram", + "model_url": "https://api.deepgram.com/v1" + } + ] + } + + env_vars = { + "VIBEVOICE_ASR_URL": "host.docker.internal:8767", + "PARAKEET_ASR_URL": "localhost:8080" + } + + with patch.dict(os.environ, env_vars): + config = OmegaConf.create(config_template) + resolved = OmegaConf.to_container(config, resolve=True) + + assert resolved["models"][0]["model_url"] == "http://host.docker.internal:8767" + assert resolved["models"][1]["model_url"] == "http://localhost:8080" + assert resolved["models"][2]["model_url"] == "https://api.deepgram.com/v1" + + +if __name__ == "__main__": + pytest.main([__file__, "-v"]) diff --git a/config_manager.py b/config_manager.py index 6f8a85a6..7919c5bc 100644 --- a/config_manager.py +++ b/config_manager.py @@ -325,6 +325,25 @@ def update_config_defaults(self, updates: Dict[str, str]): self._save_config_yml(config) + def add_or_update_model(self, model_def: Dict[str, Any]): + """ + Add or update a model in the models list by name. + + Args: + model_def: Model definition dict with at least a 'name' key. + """ + config = self._load_config_yml() + if "models" not in config: + config["models"] = [] + # Update existing or append + for i, m in enumerate(config["models"]): + if m.get("name") == model_def["name"]: + config["models"][i] = model_def + break + else: + config["models"].append(model_def) + self._save_config_yml(config) + def get_full_config(self) -> Dict[str, Any]: """ Get complete config.yml as dictionary. diff --git a/extras/asr-services/docker-compose.yml b/extras/asr-services/docker-compose.yml index 7e4b0aa5..d31ea7bf 100644 --- a/extras/asr-services/docker-compose.yml +++ b/extras/asr-services/docker-compose.yml @@ -90,6 +90,8 @@ services: build: context: . dockerfile: providers/vibevoice/Dockerfile + args: + PYTORCH_CUDA_VERSION: ${PYTORCH_CUDA_VERSION:-cu126} image: chronicle-asr-vibevoice:latest ports: - "${ASR_PORT:-8767}:8765" @@ -112,6 +114,9 @@ services: - DEVICE=${DEVICE:-cuda} - TORCH_DTYPE=${TORCH_DTYPE:-bfloat16} - MAX_NEW_TOKENS=${MAX_NEW_TOKENS:-8192} + dns: + - 8.8.8.8 + - 8.8.4.4 restart: unless-stopped # ============================================================================ diff --git a/extras/asr-services/providers/vibevoice/Dockerfile b/extras/asr-services/providers/vibevoice/Dockerfile index 218abb0c..a8d110e9 100644 --- a/extras/asr-services/providers/vibevoice/Dockerfile +++ b/extras/asr-services/providers/vibevoice/Dockerfile @@ -8,6 +8,8 @@ ######################### builder ################################# FROM ghcr.io/astral-sh/uv:python3.12-bookworm-slim AS builder +ARG PYTORCH_CUDA_VERSION=cu126 + WORKDIR /app # Install system dependencies for building @@ -17,7 +19,7 @@ RUN apt-get update && apt-get install -y --no-install-recommends \ # Dependency manifest first for cache-friendly installs COPY pyproject.toml uv.lock ./ -RUN uv sync --no-install-project --group vibevoice && \ +RUN uv sync --no-install-project --group vibevoice --extra ${PYTORCH_CUDA_VERSION} && \ uv cache clean ######################### runtime ################################# diff --git a/extras/asr-services/tests/test_cuda_version_config.py b/extras/asr-services/tests/test_cuda_version_config.py new file mode 100644 index 00000000..b4ded1c4 --- /dev/null +++ b/extras/asr-services/tests/test_cuda_version_config.py @@ -0,0 +1,220 @@ +""" +Unit tests for CUDA version configuration in ASR service Dockerfiles. + +Tests the configurable PYTORCH_CUDA_VERSION build arg that allows selecting +different CUDA versions (cu121, cu126, cu128) for different GPU architectures. +""" + +import os +import re +import pytest +from pathlib import Path + + +class TestDockerfileCUDASupport: + """Test that Dockerfiles support configurable CUDA versions.""" + + @pytest.fixture + def vibevoice_dockerfile_path(self): + """Path to VibeVoice Dockerfile.""" + return Path(__file__).parent.parent / "providers" / "vibevoice" / "Dockerfile" + + @pytest.fixture + def nemo_dockerfile_path(self): + """Path to NeMo Dockerfile.""" + return Path(__file__).parent.parent / "providers" / "nemo" / "Dockerfile" + + @pytest.fixture + def docker_compose_path(self): + """Path to docker-compose.yml.""" + return Path(__file__).parent.parent / "docker-compose.yml" + + def test_vibevoice_dockerfile_has_cuda_arg(self, vibevoice_dockerfile_path): + """Test that VibeVoice Dockerfile declares PYTORCH_CUDA_VERSION arg.""" + content = vibevoice_dockerfile_path.read_text() + + # Should have ARG declaration + assert re.search(r"ARG\s+PYTORCH_CUDA_VERSION", content), \ + "Dockerfile must declare PYTORCH_CUDA_VERSION build arg" + + # Should have default value + arg_match = re.search(r"ARG\s+PYTORCH_CUDA_VERSION=(\w+)", content) + assert arg_match, "PYTORCH_CUDA_VERSION should have default value" + default_version = arg_match.group(1) + assert default_version in ["cu121", "cu126", "cu128"], \ + f"Default CUDA version {default_version} should be cu121, cu126, or cu128" + + def test_vibevoice_dockerfile_uses_cuda_arg_in_uv_sync(self, vibevoice_dockerfile_path): + """Test that VibeVoice Dockerfile uses CUDA arg in uv sync command.""" + content = vibevoice_dockerfile_path.read_text() + + # Should use --extra ${PYTORCH_CUDA_VERSION} + assert re.search(r"uv\s+sync.*--extra\s+\$\{PYTORCH_CUDA_VERSION\}", content), \ + "uv sync command must include --extra ${PYTORCH_CUDA_VERSION}" + + def test_nemo_dockerfile_has_cuda_support(self, nemo_dockerfile_path): + """Test that NeMo Dockerfile (reference implementation) has CUDA support.""" + content = nemo_dockerfile_path.read_text() + + assert re.search(r"ARG\s+PYTORCH_CUDA_VERSION", content), \ + "NeMo Dockerfile should have PYTORCH_CUDA_VERSION arg" + + assert re.search(r"uv\s+sync.*--extra\s+\$\{PYTORCH_CUDA_VERSION\}", content), \ + "NeMo Dockerfile should use CUDA version in uv sync" + + def test_docker_compose_passes_cuda_arg_to_vibevoice(self, docker_compose_path): + """Test that docker-compose.yml passes PYTORCH_CUDA_VERSION to vibevoice service.""" + content = docker_compose_path.read_text() + + # Find vibevoice-asr service section + vibevoice_section = re.search( + r"vibevoice-asr:.*?(?=^\S|\Z)", + content, + re.MULTILINE | re.DOTALL + ) + assert vibevoice_section, "docker-compose.yml must have vibevoice-asr service" + + section_text = vibevoice_section.group(0) + + # Should have build args section + assert re.search(r"args:", section_text), \ + "vibevoice-asr service should have build args section" + + # Should pass PYTORCH_CUDA_VERSION + assert re.search( + r"PYTORCH_CUDA_VERSION:\s*\$\{PYTORCH_CUDA_VERSION:-cu126\}", + section_text + ), "vibevoice-asr should pass PYTORCH_CUDA_VERSION build arg with cu126 default" + + def test_docker_compose_cuda_arg_consistency(self, docker_compose_path): + """Test that all GPU-enabled services use consistent CUDA version pattern.""" + content = docker_compose_path.read_text() + + # Services that should have CUDA support + gpu_services = ["vibevoice-asr", "nemo-asr", "parakeet-asr"] + + for service_name in gpu_services: + service_match = re.search( + rf"{service_name}:.*?(?=^\S|\Z)", + content, + re.MULTILINE | re.DOTALL + ) + + if service_match: + service_text = service_match.group(0) + + # Check if service has GPU resources + if "devices:" in service_text and "nvidia" in service_text: + # Should have PYTORCH_CUDA_VERSION arg + assert re.search( + r"PYTORCH_CUDA_VERSION:\s*\$\{PYTORCH_CUDA_VERSION:-cu\d+\}", + service_text + ), f"{service_name} with GPU should have PYTORCH_CUDA_VERSION build arg" + + +class TestCUDAVersionEnvironmentVariable: + """Test CUDA version environment variable handling.""" + + def test_cuda_version_env_var_format(self): + """Test that CUDA version environment variables follow correct format.""" + valid_versions = ["cu121", "cu126", "cu128"] + + for version in valid_versions: + assert re.match(r"^cu\d{3}$", version), \ + f"{version} should match pattern cu### (e.g., cu121, cu126)" + + def test_cuda_version_from_env(self): + """Test reading CUDA version from environment.""" + test_version = "cu128" + + with pytest.MonkeyPatch.context() as mp: + mp.setenv("PYTORCH_CUDA_VERSION", test_version) + cuda_version = os.getenv("PYTORCH_CUDA_VERSION") + + assert cuda_version == test_version + assert cuda_version in ["cu121", "cu126", "cu128"] + + def test_cuda_version_default_fallback(self): + """Test that default CUDA version is used when env var not set.""" + with pytest.MonkeyPatch.context() as mp: + mp.delenv("PYTORCH_CUDA_VERSION", raising=False) + + # Simulate docker-compose default: ${PYTORCH_CUDA_VERSION:-cu126} + cuda_version = os.getenv("PYTORCH_CUDA_VERSION", "cu126") + + assert cuda_version == "cu126" + + +class TestGPUArchitectureCUDAMapping: + """Test that GPU architectures map to correct CUDA versions.""" + + def test_rtx_5090_requires_cu128(self): + """ + Test that RTX 5090 (sm_120) requires CUDA 12.8+. + + RTX 5090 has CUDA capability 12.0 (sm_120) which requires + PyTorch built with CUDA 12.8 or higher. + """ + gpu_arch = "sm_120" # RTX 5090 + required_cuda = "cu128" + + # Map GPU architecture to minimum CUDA version + arch_to_cuda = { + "sm_120": "cu128", # RTX 5090, RTX 50 series + "sm_90": "cu126", # RTX 4090, H100 + "sm_89": "cu121", # RTX 4090 + "sm_86": "cu121", # RTX 3090, A6000 + } + + assert arch_to_cuda.get(gpu_arch) == required_cuda, \ + f"GPU architecture {gpu_arch} requires CUDA version {required_cuda}" + + def test_older_gpus_work_with_cu121(self): + """Test that older GPUs (sm_86, sm_80) work with cu121.""" + older_archs = ["sm_86", "sm_80", "sm_75"] # RTX 3090, A100, RTX 2080 + + for arch in older_archs: + # cu121 supports these architectures + assert arch in ["sm_75", "sm_80", "sm_86"], \ + f"{arch} should be supported by CUDA 12.1" + + +class TestPyProjectCUDAExtras: + """Test that pyproject.toml defines CUDA version extras correctly.""" + + @pytest.fixture + def pyproject_path(self): + """Path to pyproject.toml.""" + return Path(__file__).parent.parent / "pyproject.toml" + + def test_pyproject_has_cuda_extras(self, pyproject_path): + """Test that pyproject.toml defines cu121, cu126, cu128 extras.""" + if not pyproject_path.exists(): + pytest.skip("pyproject.toml not found") + + content = pyproject_path.read_text() + + # Should have [project.optional-dependencies] or [tool.uv] with extras + cuda_versions = ["cu121", "cu126", "cu128"] + + for version in cuda_versions: + # Look for the CUDA version as an extra + assert re.search(rf'["\']?{version}["\']?\s*=', content), \ + f"pyproject.toml should define {version} extra" + + def test_pyproject_cuda_extras_have_pytorch(self, pyproject_path): + """Test that CUDA extras include torch/torchaudio dependencies.""" + if not pyproject_path.exists(): + pytest.skip("pyproject.toml not found") + + content = pyproject_path.read_text() + + # Each CUDA extra should reference torch with the appropriate index + # e.g., { extra = "cu128" } or { index = "pytorch-cu128" } + assert re.search(r'extra\s*=\s*["\']cu\d{3}["\']', content) or \ + re.search(r'index\s*=\s*["\']pytorch-cu\d{3}["\']', content), \ + "CUDA extras should reference PyTorch with CUDA version" + + +if __name__ == "__main__": + pytest.main([__file__, "-v"]) From e2b924ceeee8a68bb31562630b9df0c780d13750 Mon Sep 17 00:00:00 2001 From: 0xrushi <6279035+0xrushi@users.noreply.github.com> Date: Sat, 7 Feb 2026 00:02:22 -0500 Subject: [PATCH 02/11] Remove outdated unit tests for LLM custom provider and transcription URL configuration --- .../tests/test_setup_llm_custom_provider.py | 286 ------------------ .../tests/test_transcription_url_config.py | 183 ----------- .../test_llm_custom_provider.robot | 258 ++++++++++++++++ .../test_transcription_url.robot | 126 ++++++++ tests/libs/ConfigTestHelper.py | 73 +++++ tests/test-requirements.txt | 2 + 6 files changed, 459 insertions(+), 469 deletions(-) delete mode 100644 backends/advanced/tests/test_setup_llm_custom_provider.py delete mode 100644 backends/advanced/tests/test_transcription_url_config.py create mode 100644 tests/configuration/test_llm_custom_provider.robot create mode 100644 tests/configuration/test_transcription_url.robot create mode 100644 tests/libs/ConfigTestHelper.py diff --git a/backends/advanced/tests/test_setup_llm_custom_provider.py b/backends/advanced/tests/test_setup_llm_custom_provider.py deleted file mode 100644 index 33014383..00000000 --- a/backends/advanced/tests/test_setup_llm_custom_provider.py +++ /dev/null @@ -1,286 +0,0 @@ -""" -Unit tests for OpenAI-Compatible custom LLM provider setup. - -Tests the wizard's choice "3" (OpenAI-Compatible) in setup_llm(), -including model creation in config.yml and defaults updates. -""" - -import os -import sys -import tempfile -import shutil -from pathlib import Path -from unittest.mock import patch, MagicMock - -import pytest -import yaml - -# Add repo root to path for imports -sys.path.insert(0, str(Path(__file__).resolve().parent.parent.parent.parent)) -from config_manager import ConfigManager - - -@pytest.fixture -def temp_config_dir(): - """Create a temporary directory with a minimal config.yml.""" - tmpdir = tempfile.mkdtemp() - config_dir = Path(tmpdir) / "config" - config_dir.mkdir() - - config = { - "defaults": { - "llm": "openai-llm", - "embedding": "openai-embed", - "stt": "stt-deepgram", - }, - "models": [ - { - "name": "openai-llm", - "description": "OpenAI GPT-4o-mini", - "model_type": "llm", - "model_provider": "openai", - "api_family": "openai", - "model_name": "gpt-4o-mini", - "model_url": "https://api.openai.com/v1", - "api_key": "${oc.env:OPENAI_API_KEY,''}", - "model_params": {"temperature": 0.2, "max_tokens": 2000}, - "model_output": "json", - }, - { - "name": "local-embed", - "description": "Local embeddings via Ollama", - "model_type": "embedding", - "model_provider": "ollama", - "api_family": "openai", - "model_name": "nomic-embed-text:latest", - "model_url": "http://localhost:11434/v1", - "api_key": "${oc.env:OPENAI_API_KEY,ollama}", - "embedding_dimensions": 768, - "model_output": "vector", - }, - ], - "memory": {"provider": "chronicle"}, - } - - config_path = config_dir / "config.yml" - with open(config_path, "w") as f: - yaml.dump(config, f, default_flow_style=False, sort_keys=False) - - yield tmpdir - - shutil.rmtree(tmpdir) - - -@pytest.fixture -def config_manager(temp_config_dir): - """Create a ConfigManager pointing to the temp config.""" - return ConfigManager(service_path=None, repo_root=Path(temp_config_dir)) - - -class TestAddOrUpdateModel: - """Tests for ConfigManager.add_or_update_model().""" - - def test_add_new_model(self, config_manager): - """add_or_update_model() should append a new model when name doesn't exist.""" - new_model = { - "name": "custom-llm", - "description": "Custom OpenAI-compatible LLM", - "model_type": "llm", - "model_provider": "openai", - "api_family": "openai", - "model_name": "llama-3.1-70b-versatile", - "model_url": "https://api.groq.com/openai/v1", - "api_key": "${oc.env:CUSTOM_LLM_API_KEY,''}", - "model_params": {"temperature": 0.2, "max_tokens": 2000}, - "model_output": "json", - } - - config_manager.add_or_update_model(new_model) - - config = config_manager.get_full_config() - model_names = [m["name"] for m in config["models"]] - assert "custom-llm" in model_names - - added = next(m for m in config["models"] if m["name"] == "custom-llm") - assert added["model_name"] == "llama-3.1-70b-versatile" - assert added["model_url"] == "https://api.groq.com/openai/v1" - assert added["model_type"] == "llm" - - def test_update_existing_model(self, config_manager): - """add_or_update_model() should replace an existing model with the same name.""" - # First add - model_v1 = { - "name": "custom-llm", - "model_type": "llm", - "model_name": "model-v1", - "model_url": "https://example.com/v1", - } - config_manager.add_or_update_model(model_v1) - - # Then update - model_v2 = { - "name": "custom-llm", - "model_type": "llm", - "model_name": "model-v2", - "model_url": "https://example.com/v2", - } - config_manager.add_or_update_model(model_v2) - - config = config_manager.get_full_config() - custom_models = [m for m in config["models"] if m["name"] == "custom-llm"] - assert len(custom_models) == 1 - assert custom_models[0]["model_name"] == "model-v2" - assert custom_models[0]["model_url"] == "https://example.com/v2" - - def test_add_model_to_empty_models_list(self, temp_config_dir): - """add_or_update_model() should create models list if it doesn't exist.""" - config_path = Path(temp_config_dir) / "config" / "config.yml" - with open(config_path, "w") as f: - yaml.dump({"defaults": {"llm": "openai-llm"}}, f) - - cm = ConfigManager(service_path=None, repo_root=Path(temp_config_dir)) - cm.add_or_update_model({"name": "test-model", "model_type": "llm"}) - - config = cm.get_full_config() - assert "models" in config - assert len(config["models"]) == 1 - assert config["models"][0]["name"] == "test-model" - - -class TestSetupLlmCustomProvider: - """Tests for the custom LLM provider flow in setup_llm().""" - - def _make_setup(self, temp_config_dir): - """Create a ChronicleSetup instance pointing at the temp config.""" - # We need to mock the ChronicleSetup constructor's checks - # Instead, we test the logic by calling config_manager directly, - # simulating what setup_llm() choice "3" does. - return ConfigManager(service_path=None, repo_root=Path(temp_config_dir)) - - def test_custom_llm_model_added_to_config(self, config_manager): - """Selecting custom provider should create correct model entry.""" - llm_model = { - "name": "custom-llm", - "description": "Custom OpenAI-compatible LLM", - "model_type": "llm", - "model_provider": "openai", - "api_family": "openai", - "model_name": "llama-3.1-70b-versatile", - "model_url": "https://api.groq.com/openai/v1", - "api_key": "${oc.env:CUSTOM_LLM_API_KEY,''}", - "model_params": {"temperature": 0.2, "max_tokens": 2000}, - "model_output": "json", - } - - config_manager.add_or_update_model(llm_model) - - config = config_manager.get_full_config() - model = next(m for m in config["models"] if m["name"] == "custom-llm") - assert model["model_provider"] == "openai" - assert model["api_family"] == "openai" - assert model["model_name"] == "llama-3.1-70b-versatile" - assert model["model_url"] == "https://api.groq.com/openai/v1" - assert model["api_key"] == "${oc.env:CUSTOM_LLM_API_KEY,''}" - assert model["model_params"]["temperature"] == 0.2 - assert model["model_output"] == "json" - - def test_custom_llm_and_embedding_model_added(self, config_manager): - """Both LLM and embedding models should be created when embedding model is provided.""" - llm_model = { - "name": "custom-llm", - "model_type": "llm", - "model_provider": "openai", - "api_family": "openai", - "model_name": "llama-3.1-70b-versatile", - "model_url": "https://api.groq.com/openai/v1", - "api_key": "${oc.env:CUSTOM_LLM_API_KEY,''}", - "model_params": {"temperature": 0.2, "max_tokens": 2000}, - "model_output": "json", - } - embed_model = { - "name": "custom-embed", - "description": "Custom OpenAI-compatible embeddings", - "model_type": "embedding", - "model_provider": "openai", - "api_family": "openai", - "model_name": "text-embedding-3-small", - "model_url": "https://api.groq.com/openai/v1", - "api_key": "${oc.env:CUSTOM_LLM_API_KEY,''}", - "embedding_dimensions": 1536, - "model_output": "vector", - } - - config_manager.add_or_update_model(llm_model) - config_manager.add_or_update_model(embed_model) - - config = config_manager.get_full_config() - model_names = [m["name"] for m in config["models"]] - assert "custom-llm" in model_names - assert "custom-embed" in model_names - - embed = next(m for m in config["models"] if m["name"] == "custom-embed") - assert embed["model_type"] == "embedding" - assert embed["model_name"] == "text-embedding-3-small" - assert embed["embedding_dimensions"] == 1536 - - def test_custom_llm_without_embedding_falls_back_to_local_embed(self, config_manager): - """defaults.embedding should be local-embed when no custom embedding is provided.""" - llm_model = { - "name": "custom-llm", - "model_type": "llm", - "model_name": "some-model", - "model_url": "https://api.example.com/v1", - } - config_manager.add_or_update_model(llm_model) - config_manager.update_config_defaults({"llm": "custom-llm", "embedding": "local-embed"}) - - defaults = config_manager.get_config_defaults() - assert defaults["llm"] == "custom-llm" - assert defaults["embedding"] == "local-embed" - - def test_custom_llm_updates_defaults_with_embedding(self, config_manager): - """defaults.llm and defaults.embedding should be updated correctly with custom embed.""" - config_manager.update_config_defaults({"llm": "custom-llm", "embedding": "custom-embed"}) - - defaults = config_manager.get_config_defaults() - assert defaults["llm"] == "custom-llm" - assert defaults["embedding"] == "custom-embed" - - def test_custom_llm_api_key_env_reference(self, config_manager): - """API key should use env var reference in config.yml model.""" - llm_model = { - "name": "custom-llm", - "model_type": "llm", - "model_name": "some-model", - "model_url": "https://api.example.com/v1", - "api_key": "${oc.env:CUSTOM_LLM_API_KEY,''}", - } - config_manager.add_or_update_model(llm_model) - - config = config_manager.get_full_config() - model = next(m for m in config["models"] if m["name"] == "custom-llm") - assert model["api_key"] == "${oc.env:CUSTOM_LLM_API_KEY,''}" - - def test_existing_models_preserved_after_adding_custom(self, config_manager): - """Adding a custom model should not remove existing models.""" - config_before = config_manager.get_full_config() - original_count = len(config_before["models"]) - - config_manager.add_or_update_model({ - "name": "custom-llm", - "model_type": "llm", - "model_name": "test-model", - "model_url": "https://example.com/v1", - }) - - config_after = config_manager.get_full_config() - assert len(config_after["models"]) == original_count + 1 - # Original models still present - model_names = [m["name"] for m in config_after["models"]] - assert "openai-llm" in model_names - assert "local-embed" in model_names - assert "custom-llm" in model_names - - -if __name__ == "__main__": - pytest.main([__file__, "-v"]) diff --git a/backends/advanced/tests/test_transcription_url_config.py b/backends/advanced/tests/test_transcription_url_config.py deleted file mode 100644 index b9fcf9bc..00000000 --- a/backends/advanced/tests/test_transcription_url_config.py +++ /dev/null @@ -1,183 +0,0 @@ -""" -Unit tests for transcription service URL configuration. - -Tests the fix for the double http:// prefix issue where environment variables -containing protocol prefixes were incorrectly combined with hardcoded prefixes -in config.yml. -""" - -import os -import pytest -from unittest.mock import patch, MagicMock -from omegaconf import OmegaConf - - -class TestTranscriptionURLConfiguration: - """Test transcription service URL configuration and parsing.""" - - def test_vibevoice_url_without_http_prefix(self): - """Test that VIBEVOICE_ASR_URL without http:// prefix works correctly.""" - # Simulate config.yml template: http://${oc.env:VIBEVOICE_ASR_URL} - config_template = {"model_url": "http://${oc.env:VIBEVOICE_ASR_URL,host.docker.internal:8767}"} - - with patch.dict(os.environ, {"VIBEVOICE_ASR_URL": "host.docker.internal:8767"}): - resolved = OmegaConf.create(config_template) - resolved = OmegaConf.to_container(resolved, resolve=True) - - assert resolved["model_url"] == "http://host.docker.internal:8767" - assert "http://http://" not in resolved["model_url"] - - def test_vibevoice_url_with_http_prefix_causes_double_prefix(self): - """Test that VIBEVOICE_ASR_URL WITH http:// causes double prefix (bug scenario).""" - config_template = {"model_url": "http://${oc.env:VIBEVOICE_ASR_URL,host.docker.internal:8767}"} - - # This is the BUG scenario - env var already has http:// - with patch.dict(os.environ, {"VIBEVOICE_ASR_URL": "http://host.docker.internal:8767"}): - resolved = OmegaConf.create(config_template) - resolved = OmegaConf.to_container(resolved, resolve=True) - - # This demonstrates the bug - assert resolved["model_url"] == "http://http://host.docker.internal:8767" - assert "http://http://" in resolved["model_url"] - - def test_vibevoice_url_default_fallback(self): - """Test that default fallback works when VIBEVOICE_ASR_URL is not set.""" - config_template = {"model_url": "http://${oc.env:VIBEVOICE_ASR_URL,host.docker.internal:8767}"} - - # No VIBEVOICE_ASR_URL set - should use default - with patch.dict(os.environ, {}, clear=True): - resolved = OmegaConf.create(config_template) - resolved = OmegaConf.to_container(resolved, resolve=True) - - assert resolved["model_url"] == "http://host.docker.internal:8767" - - def test_parakeet_url_configuration(self): - """Test that PARAKEET_ASR_URL follows same pattern.""" - config_template = {"model_url": "http://${oc.env:PARAKEET_ASR_URL,172.17.0.1:8767}"} - - # Correct format - without http:// prefix - with patch.dict(os.environ, {"PARAKEET_ASR_URL": "host.docker.internal:8767"}): - resolved = OmegaConf.create(config_template) - resolved = OmegaConf.to_container(resolved, resolve=True) - - assert resolved["model_url"] == "http://host.docker.internal:8767" - assert "http://http://" not in resolved["model_url"] - - def test_url_parsing_removes_double_slashes(self): - """Test that URL with double http:// causes connection failures.""" - from urllib.parse import urlparse - - # Valid URL - valid_url = "http://host.docker.internal:8767/transcribe" - parsed_valid = urlparse(valid_url) - assert parsed_valid.scheme == "http" - assert parsed_valid.netloc == "host.docker.internal:8767" - - # Invalid URL with double prefix - invalid_url = "http://http://host.docker.internal:8767/transcribe" - parsed_invalid = urlparse(invalid_url) - # urlparse treats "http:" as the netloc which causes DNS failures - assert parsed_invalid.scheme == "http" - assert parsed_invalid.netloc == "http:" # Invalid netloc causes "Name or service not known" - assert parsed_invalid.netloc != "host.docker.internal:8767" - - -class TestProviderSegmentsConfiguration: - """Test use_provider_segments configuration for different providers.""" - - def test_use_provider_segments_default_false(self): - """Test that use_provider_segments defaults to false.""" - config = OmegaConf.create({ - "backend": { - "transcription": {} - } - }) - - use_segments = config.backend.transcription.get("use_provider_segments", False) - assert use_segments is False - - def test_use_provider_segments_explicit_true(self): - """Test that use_provider_segments can be enabled.""" - config = OmegaConf.create({ - "backend": { - "transcription": { - "use_provider_segments": True - } - } - }) - - assert config.backend.transcription.use_provider_segments is True - - def test_vibevoice_should_use_provider_segments(self): - """ - Test that VibeVoice provider should have use_provider_segments=true - since it provides diarized segments. - """ - # VibeVoice provides segments with speaker diarization - vibevoice_capabilities = ["segments", "diarization"] - - # When provider has both capabilities, use_provider_segments should be true - has_diarization = "diarization" in vibevoice_capabilities - has_segments = "segments" in vibevoice_capabilities - - should_use_segments = has_diarization and has_segments - assert should_use_segments is True - - -class TestModelRegistryURLResolution: - """Test model registry URL resolution with environment variables.""" - - def test_model_url_resolution_with_env_var(self): - """Test that model URLs resolve correctly from environment.""" - config_template = """ - defaults: - stt: stt-vibevoice - models: - - name: stt-vibevoice - model_type: stt - model_provider: vibevoice - model_url: http://${oc.env:VIBEVOICE_ASR_URL,host.docker.internal:8767} - """ - - with patch.dict(os.environ, {"VIBEVOICE_ASR_URL": "host.docker.internal:8767"}): - config = OmegaConf.create(config_template) - resolved = OmegaConf.to_container(config, resolve=True) - - vibevoice_model = resolved["models"][0] - assert vibevoice_model["model_url"] == "http://host.docker.internal:8767" - - def test_multiple_asr_providers_url_resolution(self): - """Test that multiple ASR providers can use different URL patterns.""" - config_template = { - "models": [ - { - "name": "stt-vibevoice", - "model_url": "http://${oc.env:VIBEVOICE_ASR_URL,host.docker.internal:8767}" - }, - { - "name": "stt-parakeet", - "model_url": "http://${oc.env:PARAKEET_ASR_URL,172.17.0.1:8767}" - }, - { - "name": "stt-deepgram", - "model_url": "https://api.deepgram.com/v1" - } - ] - } - - env_vars = { - "VIBEVOICE_ASR_URL": "host.docker.internal:8767", - "PARAKEET_ASR_URL": "localhost:8080" - } - - with patch.dict(os.environ, env_vars): - config = OmegaConf.create(config_template) - resolved = OmegaConf.to_container(config, resolve=True) - - assert resolved["models"][0]["model_url"] == "http://host.docker.internal:8767" - assert resolved["models"][1]["model_url"] == "http://localhost:8080" - assert resolved["models"][2]["model_url"] == "https://api.deepgram.com/v1" - - -if __name__ == "__main__": - pytest.main([__file__, "-v"]) diff --git a/tests/configuration/test_llm_custom_provider.robot b/tests/configuration/test_llm_custom_provider.robot new file mode 100644 index 00000000..fa9a09c3 --- /dev/null +++ b/tests/configuration/test_llm_custom_provider.robot @@ -0,0 +1,258 @@ +*** Settings *** +Documentation Tests for LLM Custom Provider Setup (ConfigManager) +Library OperatingSystem +Library Collections +Library String +Library ../libs/ConfigTestHelper.py + +*** Keywords *** +Setup Temp Config + [Documentation] Creates a temporary configuration environment + ${random_suffix}= Generate Random String 8 [NUMBERS] + ${temp_path}= Join Path ${OUTPUT DIR} temp_config_${random_suffix} + Create Directory ${temp_path} + + # Create initial default config content + ${defaults}= Create Dictionary llm=openai-llm embedding=openai-embed stt=stt-deepgram + ${model1_params}= Create Dictionary temperature=${0.2} max_tokens=${2000} + ${model1}= Create Dictionary + ... name=openai-llm + ... description=OpenAI GPT-4o-mini + ... model_type=llm + ... model_provider=openai + ... api_family=openai + ... model_name=gpt-4o-mini + ... model_url=https://api.openai.com/v1 + ... api_key=\${oc.env:OPENAI_API_KEY,''} + ... model_params=${model1_params} + ... model_output=json + + ${model2}= Create Dictionary + ... name=local-embed + ... description=Local embeddings via Ollama + ... model_type=embedding + ... model_provider=ollama + ... api_family=openai + ... model_name=nomic-embed-text:latest + ... model_url=http://localhost:11434/v1 + ... api_key=\${oc.env:OPENAI_API_KEY,ollama} + ... embedding_dimensions=${768} + ... model_output=vector + + ${models}= Create List ${model1} ${model2} + ${memory}= Create Dictionary provider=chronicle + ${config}= Create Dictionary defaults=${defaults} models=${models} memory=${memory} + + Create Temp Config Structure ${temp_path} ${config} + Set Test Variable ${TEMP_PATH} ${temp_path} + +Cleanup Temp Config + Remove Directory ${TEMP_PATH} recursive=True + +*** Test Cases *** +Add New Model To Config + [Documentation] add_or_update_model() should append a new model when name doesn't exist. + [Setup] Setup Temp Config + [Teardown] Cleanup Temp Config + + ${params}= Create Dictionary temperature=${0.2} max_tokens=${2000} + ${new_model}= Create Dictionary + ... name=custom-llm + ... description=Custom OpenAI-compatible LLM + ... model_type=llm + ... model_provider=openai + ... api_family=openai + ... model_name=llama-3.1-70b-versatile + ... model_url=https://api.groq.com/openai/v1 + ... api_key=\${oc.env:CUSTOM_LLM_API_KEY,''} + ... model_params=${params} + ... model_output=json + + ${cm}= Get Config Manager Instance ${TEMP_PATH} + Add Model To Config Manager ${cm} ${new_model} + + ${config}= Call Method ${cm} get_full_config + ${models}= Get From Dictionary ${config} models + + ${target_model}= Set Variable ${None} + FOR ${m} IN @{models} + Run Keyword If '${m["name"]}' == 'custom-llm' Set Test Variable ${target_model} ${m} + END + + Should Not Be Equal ${target_model} ${None} + Should Be Equal ${target_model["model_name"]} llama-3.1-70b-versatile + Should Be Equal ${target_model["model_url"]} https://api.groq.com/openai/v1 + Should Be Equal ${target_model["model_type"]} llm + +Update Existing Model + [Documentation] add_or_update_model() should replace an existing model with the same name. + [Setup] Setup Temp Config + [Teardown] Cleanup Temp Config + + ${cm}= Get Config Manager Instance ${TEMP_PATH} + + # First add + ${model_v1}= Create Dictionary name=custom-llm model_type=llm model_name=model-v1 model_url=https://example.com/v1 + Add Model To Config Manager ${cm} ${model_v1} + + # Then update + ${model_v2}= Create Dictionary name=custom-llm model_type=llm model_name=model-v2 model_url=https://example.com/v2 + Add Model To Config Manager ${cm} ${model_v2} + + ${config}= Call Method ${cm} get_full_config + ${models}= Get From Dictionary ${config} models + + ${count}= Set Variable 0 + ${target_model}= Set Variable ${None} + FOR ${m} IN @{models} + IF '${m["name"]}' == 'custom-llm' + Set Test Variable ${target_model} ${m} + ${count}= Evaluate ${count} + 1 + END + END + + Should Be Equal As Integers ${count} 1 + Should Be Equal ${target_model["model_name"]} model-v2 + Should Be Equal ${target_model["model_url"]} https://example.com/v2 + +Add Model To Empty Models List + [Documentation] add_or_update_model() should create models list if it doesn't exist. + [Setup] Setup Temp Config + [Teardown] Cleanup Temp Config + + # Overwrite config with empty models + ${defaults}= Create Dictionary llm=openai-llm + ${empty_config}= Create Dictionary defaults=${defaults} + Create Temp Config Structure ${TEMP_PATH} ${empty_config} + + ${cm}= Get Config Manager Instance ${TEMP_PATH} + ${test_model}= Create Dictionary name=test-model model_type=llm + Add Model To Config Manager ${cm} ${test_model} + + ${config}= Call Method ${cm} get_full_config + Dictionary Should Contain Key ${config} models + ${models}= Get From Dictionary ${config} models + Length Should Be ${models} 1 + Should Be Equal ${models[0]["name"]} test-model + +Custom LLM And Embedding Model Added + [Documentation] Both LLM and embedding models should be created when embedding model is provided. + [Setup] Setup Temp Config + [Teardown] Cleanup Temp Config + + ${cm}= Get Config Manager Instance ${TEMP_PATH} + + ${params}= Create Dictionary temperature=${0.2} max_tokens=${2000} + ${llm_model}= Create Dictionary + ... name=custom-llm + ... model_type=llm + ... model_provider=openai + ... api_family=openai + ... model_name=llama-3.1-70b-versatile + ... model_url=https://api.groq.com/openai/v1 + ... api_key=\${oc.env:CUSTOM_LLM_API_KEY,''} + ... model_params=${params} + ... model_output=json + + ${embed_model}= Create Dictionary + ... name=custom-embed + ... description=Custom OpenAI-compatible embeddings + ... model_type=embedding + ... model_provider=openai + ... api_family=openai + ... model_name=text-embedding-3-small + ... model_url=https://api.groq.com/openai/v1 + ... api_key=\${oc.env:CUSTOM_LLM_API_KEY,''} + ... embedding_dimensions=${1536} + ... model_output=vector + + Add Model To Config Manager ${cm} ${llm_model} + Add Model To Config Manager ${cm} ${embed_model} + + ${config}= Call Method ${cm} get_full_config + ${models}= Get From Dictionary ${config} models + ${model_names}= Create List + FOR ${m} IN @{models} + Append To List ${model_names} ${m["name"]} + END + + List Should Contain Value ${model_names} custom-llm + List Should Contain Value ${model_names} custom-embed + + ${target_embed}= Set Variable ${None} + FOR ${m} IN @{models} + Run Keyword If '${m["name"]}' == 'custom-embed' Set Test Variable ${target_embed} ${m} + END + + Should Be Equal ${target_embed["model_type"]} embedding + Should Be Equal ${target_embed["model_name"]} text-embedding-3-small + Should Be Equal As Integers ${target_embed["embedding_dimensions"]} 1536 + +Custom LLM Without Embedding Falls Back To Local + [Documentation] defaults.embedding should be local-embed when no custom embedding is provided. + [Setup] Setup Temp Config + [Teardown] Cleanup Temp Config + + ${cm}= Get Config Manager Instance ${TEMP_PATH} + + ${llm_model}= Create Dictionary + ... name=custom-llm + ... model_type=llm + ... model_name=some-model + ... model_url=https://api.example.com/v1 + + Add Model To Config Manager ${cm} ${llm_model} + ${defaults_update}= Create Dictionary llm=custom-llm embedding=local-embed + Update Defaults In Config Manager ${cm} ${defaults_update} + + ${defaults}= Call Method ${cm} get_config_defaults + Should Be Equal ${defaults["llm"]} custom-llm + Should Be Equal ${defaults["embedding"]} local-embed + +Custom LLM Updates Defaults With Embedding + [Documentation] defaults.llm and defaults.embedding should be updated correctly with custom embed. + [Setup] Setup Temp Config + [Teardown] Cleanup Temp Config + + ${cm}= Get Config Manager Instance ${TEMP_PATH} + + ${defaults_update}= Create Dictionary llm=custom-llm embedding=custom-embed + Update Defaults In Config Manager ${cm} ${defaults_update} + + ${defaults}= Call Method ${cm} get_config_defaults + Should Be Equal ${defaults["llm"]} custom-llm + Should Be Equal ${defaults["embedding"]} custom-embed + +Existing Models Preserved After Adding Custom + [Documentation] Adding a custom model should not remove existing models. + [Setup] Setup Temp Config + [Teardown] Cleanup Temp Config + + ${cm}= Get Config Manager Instance ${TEMP_PATH} + ${config_before}= Call Method ${cm} get_full_config + ${models_before}= Get From Dictionary ${config_before} models + ${original_count}= Get Length ${models_before} + + ${new_model}= Create Dictionary + ... name=custom-llm + ... model_type=llm + ... model_name=test-model + ... model_url=https://example.com/v1 + + Add Model To Config Manager ${cm} ${new_model} + + ${config_after}= Call Method ${cm} get_full_config + ${models_after}= Get From Dictionary ${config_after} models + ${new_count}= Get Length ${models_after} + ${expected_count}= Evaluate ${original_count} + 1 + + Should Be Equal As Integers ${new_count} ${expected_count} + + ${model_names}= Create List + FOR ${m} IN @{models_after} + Append To List ${model_names} ${m["name"]} + END + + List Should Contain Value ${model_names} openai-llm + List Should Contain Value ${model_names} local-embed + List Should Contain Value ${model_names} custom-llm \ No newline at end of file diff --git a/tests/configuration/test_transcription_url.robot b/tests/configuration/test_transcription_url.robot new file mode 100644 index 00000000..e0ba40e8 --- /dev/null +++ b/tests/configuration/test_transcription_url.robot @@ -0,0 +1,126 @@ +*** Settings *** +Documentation Tests for Transcription Service URL Configuration +Library Collections +Library ../libs/ConfigTestHelper.py + +*** Test Cases *** +Vibevoice Url Without Http Prefix + [Documentation] Test that VIBEVOICE_ASR_URL without http:// prefix works correctly. + ${config_template}= Create Dictionary model_url=http://\${oc.env:VIBEVOICE_ASR_URL,host.docker.internal:8767} + ${env_vars}= Create Dictionary VIBEVOICE_ASR_URL=host.docker.internal:8767 + + ${resolved}= Resolve Omega Config ${config_template} ${env_vars} + Should Be Equal ${resolved["model_url"]} http://host.docker.internal:8767 + Should Not Contain ${resolved["model_url"]} http://http:// + +Vibevoice Url With Http Prefix Causes Double Prefix + [Documentation] Test that VIBEVOICE_ASR_URL WITH http:// causes double prefix (bug scenario). + ${config_template}= Create Dictionary model_url=http://\${oc.env:VIBEVOICE_ASR_URL,host.docker.internal:8767} + ${env_vars}= Create Dictionary VIBEVOICE_ASR_URL=http://host.docker.internal:8767 + + ${resolved}= Resolve Omega Config ${config_template} ${env_vars} + Should Be Equal ${resolved["model_url"]} http://http://host.docker.internal:8767 + Should Contain ${resolved["model_url"]} http://http:// + +Vibevoice Url Default Fallback + [Documentation] Test that default fallback works when VIBEVOICE_ASR_URL is not set. + ${config_template}= Create Dictionary model_url=http://\${oc.env:VIBEVOICE_ASR_URL,host.docker.internal:8767} + ${env_vars}= Create Dictionary + + ${resolved}= Resolve Omega Config ${config_template} ${env_vars} + Should Be Equal ${resolved["model_url"]} http://host.docker.internal:8767 + +Parakeet Url Configuration + [Documentation] Test that PARAKEET_ASR_URL follows same pattern. + ${config_template}= Create Dictionary model_url=http://\${oc.env:PARAKEET_ASR_URL,172.17.0.1:8767} + ${env_vars}= Create Dictionary PARAKEET_ASR_URL=host.docker.internal:8767 + + ${resolved}= Resolve Omega Config ${config_template} ${env_vars} + Should Be Equal ${resolved["model_url"]} http://host.docker.internal:8767 + Should Not Contain ${resolved["model_url"]} http://http:// + +Url Parsing Removes Double Slashes + [Documentation] Test that URL with double http:// causes connection failures (simulated by parsing check). + + # Valid URL + ${valid_url}= Set Variable http://host.docker.internal:8767/transcribe + ${parsed_valid}= Check Url Parsing ${valid_url} + Should Be Equal ${parsed_valid["scheme"]} http + Should Be Equal ${parsed_valid["netloc"]} host.docker.internal:8767 + + # Invalid URL + ${invalid_url}= Set Variable http://http://host.docker.internal:8767/transcribe + ${parsed_invalid}= Check Url Parsing ${invalid_url} + Should Be Equal ${parsed_invalid["scheme"]} http + # In python urlparse, 'http:' becomes the netloc for 'http://http://...' + Should Be Equal ${parsed_invalid["netloc"]} http: + Should Not Be Equal ${parsed_invalid["netloc"]} host.docker.internal:8767 + +Use Provider Segments Default False + [Documentation] Test that use_provider_segments defaults to false. + ${transcription}= Create Dictionary + ${backend}= Create Dictionary transcription=${transcription} + ${config_template}= Create Dictionary backend=${backend} + ${env_vars}= Create Dictionary + + ${resolved}= Resolve Omega Config ${config_template} ${env_vars} + ${val}= Evaluate $resolved.get('backend', {}).get('transcription', {}).get('use_provider_segments', False) + Should Be Equal ${val} ${FALSE} + +Use Provider Segments Explicit True + [Documentation] Test that use_provider_segments can be enabled. + ${transcription}= Create Dictionary use_provider_segments=${TRUE} + ${backend}= Create Dictionary transcription=${transcription} + ${config_template}= Create Dictionary backend=${backend} + ${env_vars}= Create Dictionary + + ${resolved}= Resolve Omega Config ${config_template} ${env_vars} + ${val}= Evaluate $resolved['backend']['transcription']['use_provider_segments'] + Should Be Equal ${val} ${TRUE} + +Vibevoice Should Use Provider Segments + [Documentation] Test that VibeVoice provider should have use_provider_segments=true since it provides diarized segments. + # Logic simulation + ${vibevoice_capabilities}= Create List segments diarization + ${has_diarization}= Evaluate "diarization" in $vibevoice_capabilities + ${has_segments}= Evaluate "segments" in $vibevoice_capabilities + ${should_use_segments}= Evaluate $has_diarization and $has_segments + Should Be Equal ${should_use_segments} ${TRUE} + +Model Registry Url Resolution With Env Var + [Documentation] Test that model URLs resolve correctly from environment. + ${model_def}= Create Dictionary + ... name=stt-vibevoice + ... model_type=stt + ... model_provider=vibevoice + ... model_url=http://\${oc.env:VIBEVOICE_ASR_URL,host.docker.internal:8767} + + ${models}= Create List ${model_def} + ${defaults}= Create Dictionary stt=stt-vibevoice + ${config_template}= Create Dictionary defaults=${defaults} models=${models} + + ${env_vars}= Create Dictionary VIBEVOICE_ASR_URL=host.docker.internal:8767 + + ${resolved}= Resolve Omega Config ${config_template} ${env_vars} + ${resolved_models}= Get From Dictionary ${resolved} models + Should Be Equal ${resolved_models[0]["model_url"]} http://host.docker.internal:8767 + +Multiple Asr Providers Url Resolution + [Documentation] Test that multiple ASR providers can use different URL patterns. + ${m1}= Create Dictionary name=stt-vibevoice model_url=http://\${oc.env:VIBEVOICE_ASR_URL,host.docker.internal:8767} + ${m2}= Create Dictionary name=stt-parakeet model_url=http://\${oc.env:PARAKEET_ASR_URL,172.17.0.1:8767} + ${m3}= Create Dictionary name=stt-deepgram model_url=https://api.deepgram.com/v1 + + ${models}= Create List ${m1} ${m2} ${m3} + ${config_template}= Create Dictionary models=${models} + + ${env_vars}= Create Dictionary + ... VIBEVOICE_ASR_URL=host.docker.internal:8767 + ... PARAKEET_ASR_URL=localhost:8080 + + ${resolved}= Resolve Omega Config ${config_template} ${env_vars} + ${resolved_models}= Get From Dictionary ${resolved} models + + Should Be Equal ${resolved_models[0]["model_url"]} http://host.docker.internal:8767 + Should Be Equal ${resolved_models[1]["model_url"]} http://localhost:8080 + Should Be Equal ${resolved_models[2]["model_url"]} https://api.deepgram.com/v1 diff --git a/tests/libs/ConfigTestHelper.py b/tests/libs/ConfigTestHelper.py new file mode 100644 index 00000000..6fbdcab4 --- /dev/null +++ b/tests/libs/ConfigTestHelper.py @@ -0,0 +1,73 @@ +import os +import sys +import yaml +from pathlib import Path +from typing import Dict, Any, Optional, List +from omegaconf import OmegaConf +from unittest.mock import patch + +# Add repo root to path to import config_manager +sys.path.insert(0, str(Path(__file__).resolve().parent.parent.parent)) +from config_manager import ConfigManager + +class ConfigTestHelper: + """Helper library for testing configuration logic.""" + + def _to_dict(self, obj: Any) -> Any: + """Recursively converts Robot Framework DotDict to standard dict.""" + if isinstance(obj, dict): + return {k: self._to_dict(v) for k, v in obj.items()} + if isinstance(obj, list): + return [self._to_dict(v) for v in obj] + return obj + + def resolve_omega_config(self, config_template: Dict[str, Any], env_vars: Dict[str, str]) -> Dict[str, Any]: + """ + Resolves an OmegaConf configuration template with provided environment variables. + """ + config_template = self._to_dict(config_template) + # We need to ensure values are strings for os.environ + str_env_vars = {k: str(v) for k, v in env_vars.items()} + + with patch.dict(os.environ, str_env_vars): + conf = OmegaConf.create(config_template) + resolved = OmegaConf.to_container(conf, resolve=True) + return resolved + + def check_url_parsing(self, url: str) -> Dict[str, Any]: + """ + Parses a URL and returns its components to verify correct parsing. + """ + from urllib.parse import urlparse + parsed = urlparse(url) + return { + "scheme": parsed.scheme, + "netloc": parsed.netloc, + "path": parsed.path + } + + def create_temp_config_structure(self, base_path: str, content: Dict[str, Any]) -> str: + """ + Creates the config folder structure and config.yml within the given base path. + """ + content = self._to_dict(content) + path = Path(base_path) / "config" + path.mkdir(parents=True, exist_ok=True) + config_file = path / "config.yml" + with open(config_file, "w") as f: + yaml.dump(content, f, default_flow_style=False, sort_keys=False) + return str(base_path) + + def get_config_manager_instance(self, repo_root: str) -> ConfigManager: + """Returns a ConfigManager instance configured with the given repo_root.""" + return ConfigManager(service_path=None, repo_root=Path(repo_root)) + + def add_model_to_config_manager(self, cm: ConfigManager, model_def: Dict[str, Any]): + """Wrapper for add_or_update_model that converts arguments.""" + model_def = self._to_dict(model_def) + cm.add_or_update_model(model_def) + + def update_defaults_in_config_manager(self, cm: ConfigManager, updates: Dict[str, str]): + """Wrapper for update_config_defaults that converts arguments.""" + updates = self._to_dict(updates) + cm.update_config_defaults(updates) \ No newline at end of file diff --git a/tests/test-requirements.txt b/tests/test-requirements.txt index f32614e0..5cd8f020 100644 --- a/tests/test-requirements.txt +++ b/tests/test-requirements.txt @@ -6,4 +6,6 @@ robotframework-databaselibrary python-dotenv websockets pymongo +omegaconf +pyyaml \ No newline at end of file From 6b91a1e993b95eb5a867ffae488660380379ba06 Mon Sep 17 00:00:00 2001 From: Ankush Malaker <43288948+AnkushMalaker@users.noreply.github.com> Date: Thu, 19 Feb 2026 18:46:14 +0000 Subject: [PATCH 03/11] Enhance Docker Compose and service management for LangFuse integration - Updated `services.py` to include new options for service management, allowing for forced recreation of containers during startup. - Added LangFuse configuration options in the setup wizard, improving user experience for observability setup. - Introduced new API endpoints for retrieving observability configuration, enhancing integration with the frontend. - Enhanced error handling and logging for service startup processes, ensuring better visibility of configuration issues. - Updated documentation to reflect changes in service management and LangFuse integration. --- backends/advanced/init.py | 26 +++++++++++++ .../controllers/system_controller.py | 27 +++++++++++++ .../src/advanced_omi_backend/llm_client.py | 36 +++++++++++++++--- .../routers/modules/system_routes.py | 10 +++++ .../services/memory/base.py | 4 ++ .../services/memory/providers/chronicle.py | 7 +++- .../memory/providers/llm_providers.py | 38 ++++++++++++++----- .../utils/conversation_utils.py | 6 ++- .../workers/conversation_jobs.py | 6 ++- .../src/components/SpeakerNameDropdown.tsx | 22 ++++++++++- .../webui/src/pages/ConversationDetail.tsx | 27 ++++++++++++- backends/advanced/webui/src/services/api.ts | 3 ++ config/defaults.yml | 10 +++++ services.py | 29 ++++++++------ 14 files changed, 216 insertions(+), 35 deletions(-) diff --git a/backends/advanced/init.py b/backends/advanced/init.py index a1448876..4ea037b2 100644 --- a/backends/advanced/init.py +++ b/backends/advanced/init.py @@ -662,9 +662,15 @@ def setup_langfuse(self): self.config["LANGFUSE_PUBLIC_KEY"] = langfuse_pub self.config["LANGFUSE_SECRET_KEY"] = langfuse_sec self.config["LANGFUSE_BASE_URL"] = langfuse_host + + # Derive browser-accessible URL for deep-links + public_url = getattr(self.args, 'langfuse_public_url', None) or "http://localhost:3002" + self._save_langfuse_public_url(public_url) + source = "external" if "langfuse-web" not in langfuse_host else "local" self.console.print(f"[green][SUCCESS][/green] LangFuse auto-configured ({source})") self.console.print(f"[blue][INFO][/blue] Host: {langfuse_host}") + self.console.print(f"[blue][INFO][/blue] Public URL: {public_url}") self.console.print(f"[blue][INFO][/blue] Public key: {self.mask_api_key(langfuse_pub)}") return @@ -710,10 +716,28 @@ def setup_langfuse(self): if secret_key: self.config["LANGFUSE_SECRET_KEY"] = secret_key + # Browser-accessible URL for deep-links (stored in config.yml, not .env) + public_url = Prompt.ask( + "LangFuse browser URL (for dashboard links)", + default="http://localhost:3002", + ) + if public_url: + self._save_langfuse_public_url(public_url) + self.console.print("[green][SUCCESS][/green] LangFuse configured") else: self.console.print("[blue][INFO][/blue] LangFuse disabled") + def _save_langfuse_public_url(self, public_url: str): + """Save the Langfuse browser-accessible URL to config.yml.""" + full_config = self.config_manager.get_full_config() + if "observability" not in full_config: + full_config["observability"] = {} + if "langfuse" not in full_config["observability"]: + full_config["observability"]["langfuse"] = {} + full_config["observability"]["langfuse"]["public_url"] = public_url + self.config_manager.save_full_config(full_config) + def setup_network(self): """Configure network settings""" self.print_section("Network Configuration") @@ -1038,6 +1062,8 @@ def main(): help="LangFuse project secret key (from langfuse init or external)") parser.add_argument("--langfuse-host", help="LangFuse host URL (default: http://langfuse-web:3000 for local)") + parser.add_argument("--langfuse-public-url", + help="LangFuse browser-accessible URL for deep-links (default: http://localhost:3002)") parser.add_argument("--streaming-provider", choices=["deepgram", "smallest", "qwen3-asr"], help="Streaming provider when different from batch (enables batch re-transcription)") diff --git a/backends/advanced/src/advanced_omi_backend/controllers/system_controller.py b/backends/advanced/src/advanced_omi_backend/controllers/system_controller.py index 263b806f..bf3ce1b1 100644 --- a/backends/advanced/src/advanced_omi_backend/controllers/system_controller.py +++ b/backends/advanced/src/advanced_omi_backend/controllers/system_controller.py @@ -283,6 +283,33 @@ async def get_auth_config(): } +async def get_observability_config(): + """Get observability configuration for frontend (Langfuse deep-links). + + Returns non-secret data only (enabled status and browser URL). + """ + from advanced_omi_backend.openai_factory import is_langfuse_enabled + + enabled = is_langfuse_enabled() + session_base_url = None + + if enabled: + from advanced_omi_backend.config_loader import load_config + + cfg = load_config() + public_url = cfg.get("observability", {}).get("langfuse", {}).get("public_url", "") + if public_url: + # Strip trailing slash and build session URL + session_base_url = f"{public_url.rstrip('/')}/project/chronicle/sessions" + + return { + "langfuse": { + "enabled": enabled, + "session_base_url": session_base_url, + } + } + + # Audio file processing functions moved to audio_controller.py diff --git a/backends/advanced/src/advanced_omi_backend/llm_client.py b/backends/advanced/src/advanced_omi_backend/llm_client.py index ca640640..96ccc77b 100644 --- a/backends/advanced/src/advanced_omi_backend/llm_client.py +++ b/backends/advanced/src/advanced_omi_backend/llm_client.py @@ -11,7 +11,7 @@ from typing import Any, Dict, Optional from advanced_omi_backend.model_registry import get_models_registry -from advanced_omi_backend.openai_factory import create_openai_client +from advanced_omi_backend.openai_factory import create_openai_client, is_langfuse_enabled from advanced_omi_backend.services.memory.config import ( load_config_yml as _load_root_config, ) @@ -78,7 +78,8 @@ def __init__( raise def generate( - self, prompt: str, model: str | None = None, temperature: float | None = None + self, prompt: str, model: str | None = None, temperature: float | None = None, + **langfuse_kwargs, ) -> str: """Generate text completion using OpenAI-compatible API.""" try: @@ -90,6 +91,8 @@ def generate( "messages": [{"role": "user", "content": prompt}], "temperature": temp, } + if is_langfuse_enabled(): + params.update(langfuse_kwargs) response = self.client.chat.completions.create(**params) return response.choices[0].message.content.strip() @@ -98,7 +101,8 @@ def generate( raise def chat_with_tools( - self, messages: list, tools: list | None = None, model: str | None = None, temperature: float | None = None + self, messages: list, tools: list | None = None, model: str | None = None, + temperature: float | None = None, **langfuse_kwargs, ): """Chat completion with tool/function calling support. Returns raw response object.""" model_name = model or self.model @@ -109,6 +113,8 @@ def chat_with_tools( } if tools: params["tools"] = tools + if is_langfuse_enabled(): + params.update(langfuse_kwargs) return self.client.chat.completions.create(**params) def health_check(self) -> Dict: @@ -190,12 +196,20 @@ def reset_llm_client(): _llm_client = None +def _langfuse_metadata(session_id: str | None) -> dict: + """Return metadata dict with langfuse_session_id if Langfuse is enabled.""" + if session_id and is_langfuse_enabled(): + return {"langfuse_session_id": session_id} + return {} + + # Async wrapper for blocking LLM operations async def async_generate( prompt: str, model: str | None = None, temperature: float | None = None, operation: str | None = None, + langfuse_session_id: str | None = None, ) -> str: """Async wrapper for LLM text generation. @@ -203,6 +217,10 @@ async def async_generate( ``llm_operations`` config section via ``get_llm_operation()``. The resolved config determines model, temperature, max_tokens, etc. Explicit ``model``/``temperature`` kwargs still override the resolved values. + + When ``langfuse_session_id`` is provided and Langfuse is enabled, + the session ID is set on the current Langfuse trace to group all + LLM calls for a conversation. """ if operation: registry = get_models_registry() @@ -210,19 +228,21 @@ async def async_generate( op = registry.get_llm_operation(operation) client = op.get_client(is_async=True) api_params = op.to_api_params() - # Allow explicit overrides if temperature is not None: api_params["temperature"] = temperature if model is not None: api_params["model"] = model api_params["messages"] = [{"role": "user", "content": prompt}] + api_params["metadata"] = _langfuse_metadata(langfuse_session_id) response = await client.chat.completions.create(**api_params) return response.choices[0].message.content.strip() # Fallback: use singleton client client = get_llm_client() loop = asyncio.get_running_loop() - return await loop.run_in_executor(None, client.generate, prompt, model, temperature) + return await loop.run_in_executor( + None, lambda: client.generate(prompt, model, temperature) + ) async def async_chat_with_tools( @@ -231,6 +251,7 @@ async def async_chat_with_tools( model: str | None = None, temperature: float | None = None, operation: str | None = None, + langfuse_session_id: str | None = None, ): """Async wrapper for chat completion with tool calling. @@ -249,12 +270,15 @@ async def async_chat_with_tools( api_params["messages"] = messages if tools: api_params["tools"] = tools + api_params["metadata"] = _langfuse_metadata(langfuse_session_id) return await client.chat.completions.create(**api_params) # Fallback: use singleton client client = get_llm_client() loop = asyncio.get_running_loop() - return await loop.run_in_executor(None, client.chat_with_tools, messages, tools, model, temperature) + return await loop.run_in_executor( + None, lambda: client.chat_with_tools(messages, tools, model, temperature) + ) async def async_health_check() -> Dict: diff --git a/backends/advanced/src/advanced_omi_backend/routers/modules/system_routes.py b/backends/advanced/src/advanced_omi_backend/routers/modules/system_routes.py index 44d44c1a..277d7dc1 100644 --- a/backends/advanced/src/advanced_omi_backend/routers/modules/system_routes.py +++ b/backends/advanced/src/advanced_omi_backend/routers/modules/system_routes.py @@ -50,6 +50,16 @@ async def get_auth_config(): return await system_controller.get_auth_config() +@router.get("/observability") +async def get_observability_config(): + """Get observability configuration for frontend (Langfuse deep-links). + + Returns non-secret data: enabled status and browser-accessible session URL. + No authentication required. + """ + return await system_controller.get_observability_config() + + @router.get("/diarization-settings") async def get_diarization_settings(current_user: User = Depends(current_superuser)): """Get current diarization settings. Admin only.""" diff --git a/backends/advanced/src/advanced_omi_backend/services/memory/base.py b/backends/advanced/src/advanced_omi_backend/services/memory/base.py index ce0fe22b..bae18e56 100644 --- a/backends/advanced/src/advanced_omi_backend/services/memory/base.py +++ b/backends/advanced/src/advanced_omi_backend/services/memory/base.py @@ -342,6 +342,7 @@ class LLMProviderBase(ABC): @abstractmethod async def extract_memories( self, text: str, prompt: str, user_id: Optional[str] = None, + langfuse_session_id: Optional[str] = None, ) -> List[str]: """Extract meaningful fact memories from text using an LLM. @@ -349,6 +350,7 @@ async def extract_memories( text: Input text to extract memories from prompt: System prompt to guide the extraction process user_id: Optional user ID for per-user prompt override resolution + langfuse_session_id: Optional session ID for Langfuse trace grouping Returns: List of extracted fact memory strings @@ -373,6 +375,7 @@ async def propose_memory_actions( retrieved_old_memory: List[Dict[str, str]], new_facts: List[str], custom_prompt: Optional[str] = None, + langfuse_session_id: Optional[str] = None, ) -> Dict[str, Any]: """Propose memory management actions based on existing and new information. @@ -398,6 +401,7 @@ async def propose_reprocess_actions( diff_context: str, new_transcript: str, custom_prompt: Optional[str] = None, + langfuse_session_id: Optional[str] = None, ) -> Dict[str, Any]: """Propose memory updates after transcript reprocessing (e.g., speaker changes). diff --git a/backends/advanced/src/advanced_omi_backend/services/memory/providers/chronicle.py b/backends/advanced/src/advanced_omi_backend/services/memory/providers/chronicle.py index abfb7bb5..d1f51775 100644 --- a/backends/advanced/src/advanced_omi_backend/services/memory/providers/chronicle.py +++ b/backends/advanced/src/advanced_omi_backend/services/memory/providers/chronicle.py @@ -156,6 +156,7 @@ async def add_memory( fact_memories_text = await asyncio.wait_for( self.llm_provider.extract_memories( transcript, self.config.extraction_prompt, user_id=user_id, + langfuse_session_id=source_id, ), timeout=self.config.timeout_seconds, ) @@ -193,7 +194,8 @@ async def add_memory( if allow_update and fact_memories_text: memory_logger.info(f"šŸ” Allowing update for {source_id}") created_ids = await self._process_memory_updates( - fact_memories_text, embeddings, user_id, client_id, source_id, user_email + fact_memories_text, embeddings, user_id, client_id, source_id, user_email, + langfuse_session_id=source_id, ) else: memory_logger.info(f"šŸ” Not allowing update for {source_id}") @@ -578,6 +580,7 @@ async def reprocess_memory( existing_memories=existing_memory_dicts, diff_context=diff_text, new_transcript=transcript, + langfuse_session_id=source_id, ) memory_logger.info( f"šŸ”„ Reprocess LLM returned actions: {actions_obj}" @@ -786,6 +789,7 @@ async def _process_memory_updates( client_id: str, source_id: str, user_email: str, + langfuse_session_id: Optional[str] = None, ) -> List[str]: """Process memory updates using LLM-driven action proposals. @@ -848,6 +852,7 @@ async def _process_memory_updates( retrieved_old_memory=retrieved_old_memory, new_facts=memories_text, custom_prompt=None, + langfuse_session_id=langfuse_session_id, ) memory_logger.info(f"šŸ“ UpdateMemory LLM returned: {type(actions_obj)} - {actions_obj}") except Exception as e_actions: diff --git a/backends/advanced/src/advanced_omi_backend/services/memory/providers/llm_providers.py b/backends/advanced/src/advanced_omi_backend/services/memory/providers/llm_providers.py index 25afc38a..2d83d24c 100644 --- a/backends/advanced/src/advanced_omi_backend/services/memory/providers/llm_providers.py +++ b/backends/advanced/src/advanced_omi_backend/services/memory/providers/llm_providers.py @@ -15,7 +15,7 @@ from typing import Any, Dict, List, Optional from advanced_omi_backend.model_registry import ModelDef, get_models_registry -from advanced_omi_backend.openai_factory import create_openai_client +from advanced_omi_backend.openai_factory import create_openai_client, is_langfuse_enabled from advanced_omi_backend.prompt_registry import get_prompt_registry from ..base import LLMProviderBase @@ -38,6 +38,13 @@ memory_logger = logging.getLogger("memory_service") +def _langfuse_metadata(session_id: str | None) -> dict: + """Return metadata dict with langfuse_session_id if Langfuse is enabled.""" + if session_id and is_langfuse_enabled(): + return {"langfuse_session_id": session_id} + return {} + + def _get_openai_client(api_key: str, base_url: str, is_async: bool = False): """Get OpenAI client with optional Langfuse tracing. @@ -186,6 +193,7 @@ def __init__(self, config: Dict[str, Any]): async def extract_memories( self, text: str, prompt: str, user_id: Optional[str] = None, + langfuse_session_id: Optional[str] = None, ) -> List[str]: """Extract memories using OpenAI API with the enhanced fact retrieval prompt. @@ -193,6 +201,7 @@ async def extract_memories( text: Input text to extract memories from prompt: System prompt to guide extraction (uses default if empty) user_id: Optional user ID for per-user prompt override resolution + langfuse_session_id: Optional session ID for Langfuse trace grouping Returns: List of extracted memory strings @@ -209,12 +218,12 @@ async def extract_memories( user_id, current_date=datetime.now().strftime("%Y-%m-%d"), ) - + # local models can only handle small chunks of input text text_chunks = chunk_text_with_spacy(text) - + # Process all chunks in sequence, not concurrently - results = [await self._process_chunk(system_prompt, chunk, i) for i, chunk in enumerate(text_chunks)] + results = [await self._process_chunk(system_prompt, chunk, i, langfuse_session_id=langfuse_session_id) for i, chunk in enumerate(text_chunks)] # Spread list of list of facts into a single list of facts cleaned_facts = [] @@ -228,23 +237,26 @@ async def extract_memories( memory_logger.error(f"OpenAI memory extraction failed: {e}") return [] - async def _process_chunk(self, system_prompt: str, chunk: str, index: int) -> List[str]: + async def _process_chunk( + self, system_prompt: str, chunk: str, index: int, + langfuse_session_id: Optional[str] = None, + ) -> List[str]: """Process a single text chunk to extract memories using OpenAI API. - + This private method handles the LLM interaction for a single chunk of text, sending it to OpenAI's chat completion API with the specified system prompt to extract structured memory facts. - + Args: - client: OpenAI async client instance for API communication system_prompt: System prompt that guides the memory extraction behavior chunk: Individual text chunk to process for memory extraction index: Index of the chunk for logging and error tracking purposes - + langfuse_session_id: Optional session ID for Langfuse trace grouping + Returns: List of extracted memory fact strings from the chunk. Returns empty list if no facts are found or if an error occurs during processing. - + Note: Errors are logged but don't propagate to avoid failing the entire memory extraction process. @@ -258,6 +270,7 @@ async def _process_chunk(self, system_prompt: str, chunk: str, index: int) -> Li {"role": "system", "content": system_prompt}, {"role": "user", "content": chunk}, ], + metadata=_langfuse_metadata(langfuse_session_id), ) facts = (response.choices[0].message.content or "").strip() if not facts: @@ -314,6 +327,7 @@ async def propose_memory_actions( retrieved_old_memory: List[Dict[str, str]] | List[str], new_facts: List[str], custom_prompt: Optional[str] = None, + langfuse_session_id: Optional[str] = None, ) -> Dict[str, Any]: """Use OpenAI chat completion with enhanced prompt to propose memory actions. @@ -321,6 +335,7 @@ async def propose_memory_actions( retrieved_old_memory: List of existing memories for context new_facts: List of new facts to process custom_prompt: Optional custom prompt to override default + langfuse_session_id: Optional session ID for Langfuse trace grouping Returns: Dictionary containing proposed memory actions @@ -340,6 +355,7 @@ async def propose_memory_actions( response = await client.chat.completions.create( **op.to_api_params(), messages=update_memory_messages, + metadata=_langfuse_metadata(langfuse_session_id), ) content = (response.choices[0].message.content or "").strip() if not content: @@ -365,6 +381,7 @@ async def propose_reprocess_actions( diff_context: str, new_transcript: str, custom_prompt: Optional[str] = None, + langfuse_session_id: Optional[str] = None, ) -> Dict[str, Any]: """Propose memory updates after speaker re-identification. @@ -427,6 +444,7 @@ async def propose_reprocess_actions( response = await client.chat.completions.create( **op.to_api_params(), messages=messages, + metadata=_langfuse_metadata(langfuse_session_id), ) content = (response.choices[0].message.content or "").strip() diff --git a/backends/advanced/src/advanced_omi_backend/utils/conversation_utils.py b/backends/advanced/src/advanced_omi_backend/utils/conversation_utils.py index a199d5fa..63036ce1 100644 --- a/backends/advanced/src/advanced_omi_backend/utils/conversation_utils.py +++ b/backends/advanced/src/advanced_omi_backend/utils/conversation_utils.py @@ -164,6 +164,7 @@ async def generate_title_and_summary( text: str, segments: Optional[list] = None, user_id: Optional[str] = None, + langfuse_session_id: Optional[str] = None, ) -> tuple[str, str]: """ Generate title and short summary in a single LLM call using full conversation context. @@ -221,7 +222,7 @@ async def generate_title_and_summary( "{conversation_text}" """ - response = await async_generate(prompt, operation="title_summary") + response = await async_generate(prompt, operation="title_summary", langfuse_session_id=langfuse_session_id) # Parse response for Title: and Summary: lines title = None @@ -253,6 +254,7 @@ async def generate_detailed_summary( text: str, segments: Optional[list] = None, memory_context: Optional[str] = None, + langfuse_session_id: Optional[str] = None, ) -> str: """ Generate a comprehensive, detailed summary of the conversation. @@ -328,7 +330,7 @@ async def generate_detailed_summary( "{conversation_text}" """ - summary = await async_generate(prompt, operation="detailed_summary") + summary = await async_generate(prompt, operation="detailed_summary", langfuse_session_id=langfuse_session_id) return summary.strip().strip('"').strip("'") or "No meaningful content to summarize" except Exception as e: diff --git a/backends/advanced/src/advanced_omi_backend/workers/conversation_jobs.py b/backends/advanced/src/advanced_omi_backend/workers/conversation_jobs.py index 021bc9da..b1bbb8cd 100644 --- a/backends/advanced/src/advanced_omi_backend/workers/conversation_jobs.py +++ b/backends/advanced/src/advanced_omi_backend/workers/conversation_jobs.py @@ -1025,10 +1025,12 @@ async def generate_title_summary_job(conversation_id: str, *, redis_client=None) (title, short_summary), detailed_summary = await asyncio.gather( generate_title_and_summary( - transcript_text, segments=segments, user_id=conversation.user_id + transcript_text, segments=segments, user_id=conversation.user_id, + langfuse_session_id=conversation_id, ), generate_detailed_summary( - transcript_text, segments=segments, memory_context=memory_context + transcript_text, segments=segments, memory_context=memory_context, + langfuse_session_id=conversation_id, ), ) diff --git a/backends/advanced/webui/src/components/SpeakerNameDropdown.tsx b/backends/advanced/webui/src/components/SpeakerNameDropdown.tsx index 45b0ff07..b0553fc7 100644 --- a/backends/advanced/webui/src/components/SpeakerNameDropdown.tsx +++ b/backends/advanced/webui/src/components/SpeakerNameDropdown.tsx @@ -1,7 +1,9 @@ import { useState, useRef, useEffect } from 'react' -import { Check, Plus } from 'lucide-react' +import { Check, Plus, UserX } from 'lucide-react' import { useSortedSpeakers } from '../hooks/useSortedSpeakers' +const UNKNOWN_SPEAKER = 'Unknown Speaker' + interface SpeakerNameDropdownProps { currentSpeaker: string enrolledSpeakers: Array<{ speaker_id: string; name: string }> @@ -112,6 +114,24 @@ export default function SpeakerNameDropdown({ /> + {/* Unknown Speaker option */} + {(!searchQuery || UNKNOWN_SPEAKER.toLowerCase().includes(searchQuery.toLowerCase())) && ( +
+ +
+ )} + {/* Speaker list */}
{hasResults ? ( diff --git a/backends/advanced/webui/src/pages/ConversationDetail.tsx b/backends/advanced/webui/src/pages/ConversationDetail.tsx index bfbe0022..a7ec7fa4 100644 --- a/backends/advanced/webui/src/pages/ConversationDetail.tsx +++ b/backends/advanced/webui/src/pages/ConversationDetail.tsx @@ -4,9 +4,9 @@ import { useQueryClient } from '@tanstack/react-query' import { ArrowLeft, Calendar, User, Trash2, RefreshCw, MoreVertical, RotateCcw, Zap, Play, Pause, - Save, X, Pencil, Brain, Clock, Database, Layers, Star + Save, X, Pencil, Brain, Clock, Database, Layers, Star, BarChart3 } from 'lucide-react' -import { annotationsApi, speakerApi, BACKEND_URL } from '../services/api' +import { annotationsApi, speakerApi, systemApi, BACKEND_URL } from '../services/api' import { useConversationDetail, useConversationMemories, useDeleteConversation, useReprocessTranscript, useReprocessMemory, useReprocessSpeakers, useToggleStar @@ -88,6 +88,17 @@ export default function ConversationDetail() { // Dropdown menu state const [openDropdown, setOpenDropdown] = useState(false) + // Langfuse observability link + const [langfuseSessionUrl, setLangfuseSessionUrl] = useState(null) + useEffect(() => { + systemApi.getObservabilityConfig().then(res => { + const cfg = res.data?.langfuse + if (cfg?.enabled && cfg?.session_base_url) { + setLangfuseSessionUrl(cfg.session_base_url) + } + }).catch(() => {}) + }, []) + // Reprocessing state const [reprocessingTranscript, setReprocessingTranscript] = useState(false) const [reprocessingMemory, setReprocessingMemory] = useState(false) @@ -455,6 +466,7 @@ export default function ConversationDetail() { }) } setEnrolledSpeakers(prev => { + if (newSpeaker === 'Unknown Speaker') return prev if (prev.some(s => s.name === newSpeaker)) return prev return [...prev, { speaker_id: `temp_${Date.now()}_${newSpeaker}`, name: newSpeaker }] }) @@ -589,6 +601,17 @@ export default function ConversationDetail() {
+ {langfuseSessionUrl && conversation.conversation_id && ( + + + + )}
@@ -2160,21 +2178,32 @@ const Queue: React.FC = () => { } - {evt.plugins_executed.length === 0 ? ( - no plugins ran - ) : allSuccess ? ( - - - OK - - ) : anyFailure ? ( - !p.success).map(p => `${p.plugin_id}: ${p.message}`).join('; ')}> - - Error - - ) : ( - partial - )} +
+ {evt.plugins_executed.length === 0 ? ( + no plugins ran + ) : allSuccess ? ( + + + OK + + ) : anyFailure ? ( + + + Error + + ) : ( + partial + )} + {evt.plugins_executed.length > 0 && ( + + )} +
); @@ -2631,6 +2660,84 @@ const Queue: React.FC = () => {
)} + {/* Event Detail Modal */} + {selectedEvent && ( +
+
+
+

Event Details

+ +
+ +
+
+
+ +

{new Date(selectedEvent.timestamp * 1000).toLocaleString()}

+
+
+ + + {selectedEvent.event} + +
+
+ +

{selectedEvent.user_id}

+
+ {selectedEvent.metadata?.client_id && ( +
+ +

{selectedEvent.metadata.client_id}

+
+ )} +
+ +
+ +
+ {selectedEvent.plugins_executed.map((p, i) => ( +
+
+ {p.success + ? + : + } + {p.plugin_id} + + {p.success ? 'OK' : 'Error'} + +
+ {p.message && ( +

+ {p.message} +

+ )} +
+ ))} +
+
+ + {Object.keys(selectedEvent.metadata).length > 0 && ( +
+ + Raw Metadata + + 
+                    {JSON.stringify(selectedEvent.metadata, null, 2)}
+
+
+ )} +
+
+
+ )} + {/* Flush Jobs Modal */} {showFlushModal && (
diff --git a/backends/advanced/webui/src/pages/System.tsx b/backends/advanced/webui/src/pages/System.tsx index 3345134c..d06602de 100644 --- a/backends/advanced/webui/src/pages/System.tsx +++ b/backends/advanced/webui/src/pages/System.tsx @@ -22,7 +22,7 @@ export default function System() { // UI state const [menuOpen, setMenuOpen] = useState(false) - const [confirmModal, setConfirmModal] = useState<'workers' | 'backend' | null>(null) + const [confirmModal, setConfirmModal] = useState<'workers' | 'backend' | 'both' | null>(null) const [restartingBackend, setRestartingBackend] = useState(false) const [workerBanner, setWorkerBanner] = useState(false) const menuRef = useRef(null) @@ -108,6 +108,19 @@ export default function System() { }) } + const handleRestartBoth = () => { + setConfirmModal(null) + restartWorkersMutation.mutate(undefined, { + onSuccess: () => { + restartBackendMutation.mutate(undefined, { + onSuccess: () => { + pollHealth() + }, + }) + }, + }) + } + const getStatusIcon = (healthy: boolean) => { return healthy ? @@ -233,7 +246,6 @@ export default function System() { Restart Workers -
+
+
)}
@@ -285,7 +305,7 @@ export default function System() {
- ) : ( + ) : confirmModal === 'backend' ? ( <>
@@ -318,6 +338,39 @@ export default function System() {
+ ) : ( + <> +
+
+ +
+

+ Restart Both +

+
+

+ This will restart workers and then the backend. The service will be briefly unavailable. +

+
+

+ Active WebSocket connections and streaming sessions will be dropped. +

+
+
+ + +
+ )}
diff --git a/extras/asr-services/common/base_service.py b/extras/asr-services/common/base_service.py index 4b5107d1..2b81df3f 100644 --- a/extras/asr-services/common/base_service.py +++ b/extras/asr-services/common/base_service.py @@ -4,10 +4,12 @@ Provides a common interface and FastAPI app setup for all ASR providers. """ +import json import logging import os import tempfile import time +import wave from abc import ABC, abstractmethod from typing import Optional @@ -17,7 +19,7 @@ TranscriptionResult, ) from fastapi import FastAPI, File, Form, HTTPException, UploadFile -from fastapi.responses import JSONResponse +from fastapi.responses import JSONResponse, StreamingResponse logger = logging.getLogger(__name__) @@ -97,12 +99,47 @@ def get_supported_languages(self) -> Optional[list[str]]: """ return None + def supports_batch_progress(self, audio_duration: float) -> bool: + """Return True if this provider reports progress for long audio. + + Providers that batch long audio into windows can override this to + return True when the audio exceeds their batching threshold. The + ``/transcribe`` endpoint uses this to decide whether to return an + NDJSON streaming response with progress counters. + + Default implementation returns False (no progress reporting). + """ + return False + + def transcribe_with_progress(self, audio_file_path: str, context_info=None): + """Generator that yields progress counters then a final result. + + Only called when ``supports_batch_progress()`` returns True. + Subclasses that support batch progress must override this. + + Yields: + {"type": "progress", "current": i, "total": n} + {"type": "result", ...} (TranscriptionResult.to_dict()) + """ + raise NotImplementedError( + f"{self.__class__.__name__} does not implement transcribe_with_progress" + ) + @property def is_ready(self) -> bool: """Return True if service is ready to handle requests.""" return self._is_ready +def _get_audio_duration(file_path: str) -> Optional[float]: + """Return audio duration in seconds, or None if unreadable.""" + try: + with wave.open(file_path, "rb") as wf: + return wf.getnframes() / wf.getframerate() + except Exception: + return None + + def create_asr_app(service: BaseASRService) -> FastAPI: """ Create a FastAPI application with standard ASR endpoints. @@ -165,6 +202,7 @@ async def transcribe( logger.info(f"Transcription request started") tmp_filename = None + streaming_response = False try: # Read uploaded file file_read_start = time.time() @@ -186,7 +224,33 @@ async def transcribe( tmp_file.write(audio_content) tmp_filename = tmp_file.name - # Transcribe + # Check if provider supports batch progress for this audio + audio_duration = _get_audio_duration(tmp_filename) + if audio_duration and service.supports_batch_progress(audio_duration): + logger.info( + f"Audio is {audio_duration:.1f}s, using batch progress reporting" + ) + streaming_response = True + + def _ndjson_generator(): + """Wrap sync generator as NDJSON lines, clean up temp file when done.""" + try: + for event in service.transcribe_with_progress( + tmp_filename, context_info=context_info, + ): + yield json.dumps(event) + "\n" + finally: + try: + os.unlink(tmp_filename) + except Exception as e: + logger.warning(f"Failed to delete temp file {tmp_filename}: {e}") + + return StreamingResponse( + _ndjson_generator(), + media_type="application/x-ndjson", + ) + + # Normal path: single JSON response transcribe_start = time.time() result = await service.transcribe( tmp_filename, @@ -208,8 +272,9 @@ async def transcribe( raise HTTPException(status_code=500, detail=f"Transcription failed: {e}") finally: - # Cleanup temporary file - if tmp_filename: + # Streaming path owns its own cleanup via the generator's finally block. + # Only clean up here for the normal (non-streaming) path. + if tmp_filename and not streaming_response: try: os.unlink(tmp_filename) except Exception as e: diff --git a/extras/asr-services/providers/vibevoice/service.py b/extras/asr-services/providers/vibevoice/service.py index 69bc0cd1..085af599 100644 --- a/extras/asr-services/providers/vibevoice/service.py +++ b/extras/asr-services/providers/vibevoice/service.py @@ -102,6 +102,24 @@ def get_capabilities(self) -> list[str]: "long_form", ] + def supports_batch_progress(self, audio_duration: float) -> bool: + """Return True if audio is long enough for batched transcription with progress.""" + if self.transcriber is None: + return False + return self.transcriber.supports_batch_progress(audio_duration) + + def transcribe_with_progress(self, audio_file_path: str, context_info=None): + """Yield progress counters then final result for long audio. + + Delegates to the transcriber's _transcribe_batched_with_progress() generator. + Runs synchronously (called via run_in_executor by the endpoint). + """ + if self.transcriber is None: + raise RuntimeError("Service not initialized") + yield from self.transcriber._transcribe_batched_with_progress( + audio_file_path, hotwords=context_info, + ) + def _run_lora_training( data_dir: str, diff --git a/extras/asr-services/providers/vibevoice/transcriber.py b/extras/asr-services/providers/vibevoice/transcriber.py index 884b1965..55947d31 100644 --- a/extras/asr-services/providers/vibevoice/transcriber.py +++ b/extras/asr-services/providers/vibevoice/transcriber.py @@ -37,6 +37,7 @@ import torch from common.audio_utils import STANDARD_SAMPLE_RATE, load_audio_file +from omegaconf import OmegaConf from common.batching import ( extract_context_tail, split_audio_file, @@ -48,29 +49,28 @@ def load_vibevoice_config() -> dict: - """Load asr_services.vibevoice config from config.yml/defaults.yml.""" - try: - from omegaconf import OmegaConf - - config_dir = Path(os.getenv("CONFIG_DIR", "/app/config")) - defaults_path = config_dir / "defaults.yml" - config_path = config_dir / "config.yml" - - if not defaults_path.exists() and not config_path.exists(): - return {} - - defaults = OmegaConf.load(defaults_path) if defaults_path.exists() else {} - user_config = OmegaConf.load(config_path) if config_path.exists() else {} - merged = OmegaConf.merge(defaults, user_config) - - asr_config = merged.get("asr_services", {}).get("vibevoice", {}) - resolved = OmegaConf.to_container(asr_config, resolve=True) - logger.info(f"Loaded vibevoice config: {resolved}") - return resolved - except Exception as e: - logger.warning(f"Failed to load config: {e}, using env/defaults") + """Load asr_services.vibevoice config from config.yml/defaults.yml. + + Returns an empty dict only when neither config file exists. + Raises on any load/parse error so misconfigurations are caught early. + """ + config_dir = Path(os.getenv("CONFIG_DIR", "/app/config")) + defaults_path = config_dir / "defaults.yml" + config_path = config_dir / "config.yml" + + if not defaults_path.exists() and not config_path.exists(): + logger.info("No config files found in %s, using env/defaults", config_dir) return {} + defaults = OmegaConf.load(defaults_path) if defaults_path.exists() else {} + user_config = OmegaConf.load(config_path) if config_path.exists() else {} + merged = OmegaConf.merge(defaults, user_config) + + asr_config = merged.get("asr_services", {}).get("vibevoice", {}) + resolved = OmegaConf.to_container(asr_config, resolve=True) + logger.info(f"Loaded vibevoice config: {resolved}") + return resolved + class VibeVoiceTranscriber: """ @@ -462,7 +462,6 @@ def _transcribe_batched( ) batch_results = [] - prev_context = None for i, (temp_path, start_time, end_time) in enumerate(windows): try: @@ -470,9 +469,12 @@ def _transcribe_batched( f"Batch {i+1}/{len(windows)}: [{start_time:.0f}s - {end_time:.0f}s]" ) - result = self._transcribe_single(temp_path, context=prev_context, context_info=hotwords) + # NOTE: We intentionally do NOT pass prev_context between windows. + # Passing transcript tails as context can trigger degenerate + # repetition loops in model.generate(). The 30s audio overlap + + # midpoint stitching already handles boundary continuity. + result = self._transcribe_single(temp_path, context_info=hotwords) batch_results.append((result, start_time, end_time)) - prev_context = extract_context_tail(result, max_chars=500) logger.info( f"Batch {i+1} done: {len(result.segments)} segments, " f"{len(result.text)} chars" @@ -483,6 +485,55 @@ def _transcribe_batched( return stitch_transcription_results(batch_results, overlap_seconds=self.batch_overlap) + def _transcribe_batched_with_progress( + self, + audio_file_path: str, + hotwords: Optional[str] = None, + ): + """ + Transcribe a long audio file with progress reporting. + + Same logic as _transcribe_batched() but yields progress counters + between windows so callers can report how far along the batch is. + + Yields: + {"type": "progress", "current": i, "total": n} after each window + {"type": "result", ...} as the final item (TranscriptionResult.to_dict()) + """ + windows = split_audio_file( + audio_file_path, + batch_duration=self.batch_duration, + overlap=self.batch_overlap, + ) + + batch_results = [] + + for i, (temp_path, start_time, end_time) in enumerate(windows): + try: + logger.info( + f"Batch {i+1}/{len(windows)}: [{start_time:.0f}s - {end_time:.0f}s]" + ) + + # No inter-window context ā€” see note in _transcribe_batched() + result = self._transcribe_single(temp_path, context_info=hotwords) + batch_results.append((result, start_time, end_time)) + logger.info( + f"Batch {i+1} done: {len(result.segments)} segments, " + f"{len(result.text)} chars" + ) + + finally: + os.unlink(temp_path) + + yield {"type": "progress", "current": i + 1, "total": len(windows)} + + final = stitch_transcription_results(batch_results, overlap_seconds=self.batch_overlap) + yield {"type": "result", **final.to_dict()} + + def supports_batch_progress(self, audio_duration: float) -> bool: + """Return True if this audio is long enough to use batched transcription with progress.""" + return audio_duration > self.batch_threshold + def _parse_vibevoice_output(self, raw_output: str) -> dict: """ Parse VibeVoice raw output to extract segments with speaker info. diff --git a/extras/asr-services/pyproject.toml b/extras/asr-services/pyproject.toml index 5dd8a43f..fa01b2cf 100644 --- a/extras/asr-services/pyproject.toml +++ b/extras/asr-services/pyproject.toml @@ -118,6 +118,7 @@ vibevoice = [ "librosa>=0.10.0", "bitsandbytes>=0.43.0", "peft>=0.17.0", + "omegaconf>=2.3.0", ] qwen3-asr = [ diff --git a/extras/asr-services/uv.lock b/extras/asr-services/uv.lock index 716f1f2e..752a5c90 100644 --- a/extras/asr-services/uv.lock +++ b/extras/asr-services/uv.lock @@ -305,326 +305,21 @@ wheels = [ { url = "https://files.pythonhosted.org/packages/87/04/9d75e1d3bb4ab8ec67ff10919476ccdee06c098bcfcf3a352da5f985171d/absl_py-2.3.0-py3-none-any.whl", hash = "sha256:9824a48b654a306168f63e0d97714665f8490b8d89ec7bf2efc24bf67cf579b3", size = 135657, upload_time = "2025-05-27T09:15:48.742Z" }, ] -[[package]] -name = "accelerate" -version = "1.7.0" -source = { registry = "https://pypi.org/simple" } -resolution-markers = [ - "python_full_version >= '3.13' and platform_machine != 'aarch64' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra == 'group-12-asr-services-vibevoice'", - "python_full_version >= '3.13' and sys_platform != 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra == 'group-12-asr-services-vibevoice'", - "python_full_version == '3.12.*' and platform_machine != 'aarch64' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra == 'group-12-asr-services-vibevoice'", - "python_full_version == '3.12.*' and sys_platform != 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra == 'group-12-asr-services-vibevoice'", - "python_full_version >= '3.13' and platform_machine == 'aarch64' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra == 'group-12-asr-services-vibevoice'", - "python_full_version == '3.12.*' and platform_machine == 'aarch64' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra == 'group-12-asr-services-vibevoice'", - "python_full_version == '3.11.*' and platform_machine != 'aarch64' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra == 'group-12-asr-services-vibevoice'", - "python_full_version == '3.11.*' and sys_platform != 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra == 'group-12-asr-services-vibevoice'", - "python_full_version < '3.11' and platform_machine != 'aarch64' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra == 'group-12-asr-services-vibevoice'", - "python_full_version < '3.11' and sys_platform != 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra == 'group-12-asr-services-vibevoice'", - "python_full_version == '3.11.*' and platform_machine == 'aarch64' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra == 'group-12-asr-services-vibevoice'", - "python_full_version < '3.11' and platform_machine == 'aarch64' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra == 'group-12-asr-services-vibevoice'", - "python_full_version >= '3.13' and platform_machine != 'aarch64' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra == 'group-12-asr-services-vibevoice'", - "python_full_version >= '3.13' and sys_platform != 'linux' and extra != 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra == 'group-12-asr-services-vibevoice'", - "python_full_version == '3.12.*' and platform_machine != 'aarch64' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra == 'group-12-asr-services-vibevoice'", - "python_full_version == '3.12.*' and sys_platform != 'linux' and extra != 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra == 'group-12-asr-services-vibevoice'", - "python_full_version >= '3.13' and platform_machine == 'aarch64' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra == 'group-12-asr-services-vibevoice'", - "python_full_version == '3.12.*' and platform_machine == 'aarch64' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra == 'group-12-asr-services-vibevoice'", - "python_full_version == '3.11.*' and platform_machine != 'aarch64' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra == 'group-12-asr-services-vibevoice'", - "python_full_version == '3.11.*' and sys_platform != 'linux' and extra != 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra == 'group-12-asr-services-vibevoice'", - "python_full_version < '3.11' and platform_machine != 'aarch64' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra == 'group-12-asr-services-vibevoice'", - "python_full_version < '3.11' and sys_platform != 'linux' and extra != 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra == 'group-12-asr-services-vibevoice'", - "python_full_version == '3.11.*' and platform_machine == 'aarch64' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra == 'group-12-asr-services-vibevoice'", - "python_full_version < '3.11' and platform_machine == 'aarch64' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra == 'group-12-asr-services-vibevoice'", - "python_full_version >= '3.13' and sys_platform == 'linux' and extra == 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra == 'group-12-asr-services-vibevoice'", - "python_full_version >= '3.13' and sys_platform != 'linux' and extra == 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra == 'group-12-asr-services-vibevoice'", - "python_full_version == '3.12.*' and sys_platform == 'linux' and extra == 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra == 'group-12-asr-services-vibevoice'", - "python_full_version == '3.12.*' and sys_platform != 'linux' and extra == 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra == 'group-12-asr-services-vibevoice'", - "python_full_version == '3.11.*' and sys_platform == 'linux' and extra == 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra == 'group-12-asr-services-vibevoice'", - "python_full_version == '3.11.*' and sys_platform != 'linux' and extra == 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra == 'group-12-asr-services-vibevoice'", - "python_full_version < '3.11' and sys_platform == 'linux' and extra == 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra == 'group-12-asr-services-vibevoice'", - "python_full_version < '3.11' and sys_platform != 'linux' and extra == 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra == 'group-12-asr-services-vibevoice'", - "python_full_version >= '3.13' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra == 'group-12-asr-services-vibevoice'", - "python_full_version >= '3.13' and sys_platform != 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra == 'group-12-asr-services-vibevoice'", - "python_full_version == '3.12.*' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra == 'group-12-asr-services-vibevoice'", - "python_full_version == '3.12.*' and sys_platform != 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra == 'group-12-asr-services-vibevoice'", - "python_full_version == '3.11.*' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra == 'group-12-asr-services-vibevoice'", - "python_full_version == '3.11.*' and sys_platform != 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra == 'group-12-asr-services-vibevoice'", - "python_full_version < '3.11' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra == 'group-12-asr-services-vibevoice'", - "python_full_version < '3.11' and sys_platform != 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra == 'group-12-asr-services-vibevoice'", - "python_full_version >= '3.13' and platform_machine != 'aarch64' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version >= '3.13' and sys_platform != 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version == '3.12.*' and platform_machine != 'aarch64' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version == '3.12.*' and sys_platform != 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version >= '3.13' and platform_machine == 'aarch64' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version == '3.12.*' and platform_machine == 'aarch64' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version == '3.11.*' and platform_machine != 'aarch64' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version == '3.11.*' and sys_platform != 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version < '3.11' and platform_machine != 'aarch64' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version < '3.11' and sys_platform != 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version == '3.11.*' and platform_machine == 'aarch64' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version < '3.11' and platform_machine == 'aarch64' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version >= '3.13' and platform_machine != 'aarch64' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version >= '3.13' and sys_platform != 'linux' and extra != 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version == '3.12.*' and platform_machine != 'aarch64' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version == '3.12.*' and sys_platform != 'linux' and extra != 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version >= '3.13' and platform_machine == 'aarch64' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version == '3.12.*' and platform_machine == 'aarch64' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version == '3.11.*' and platform_machine != 'aarch64' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version == '3.11.*' and sys_platform != 'linux' and extra != 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version < '3.11' and platform_machine != 'aarch64' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version < '3.11' and sys_platform != 'linux' and extra != 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version == '3.11.*' and platform_machine == 'aarch64' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version < '3.11' and platform_machine == 'aarch64' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version >= '3.13' and sys_platform == 'linux' and extra == 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version >= '3.13' and sys_platform != 'linux' and extra == 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version == '3.12.*' and sys_platform == 'linux' and extra == 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version == '3.12.*' and sys_platform != 'linux' and extra == 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version == '3.11.*' and sys_platform == 'linux' and extra == 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version == '3.11.*' and sys_platform != 'linux' and extra == 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version < '3.11' and sys_platform == 'linux' and extra == 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version < '3.11' and sys_platform != 'linux' and extra == 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version >= '3.13' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version >= '3.13' and sys_platform != 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version == '3.12.*' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version == '3.12.*' and sys_platform != 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version == '3.11.*' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version == '3.11.*' and sys_platform != 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version < '3.11' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version < '3.11' and sys_platform != 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", -] -dependencies = [ - { name = "huggingface-hub", marker = "(extra == 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126') or (extra == 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu128') or (extra == 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-nemo') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-nemo') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-vibevoice')" }, - { name = "numpy", marker = "(extra == 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126') or (extra == 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu128') or (extra == 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-nemo') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-nemo') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-vibevoice')" }, - { name = "packaging", marker = "(extra == 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126') or (extra == 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu128') or (extra == 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-nemo') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-nemo') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-vibevoice')" }, - { name = "psutil", marker = "(extra == 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126') or (extra == 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu128') or (extra == 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-nemo') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-nemo') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-vibevoice')" }, - { name = "pyyaml", marker = "(extra == 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126') or (extra == 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu128') or (extra == 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-nemo') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-nemo') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-vibevoice')" }, - { name = "safetensors", marker = "(extra == 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126') or (extra == 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu128') or (extra == 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-nemo') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-nemo') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-vibevoice')" }, - { name = "torch", version = "2.5.1+cu121", source = { registry = "https://download.pytorch.org/whl/cu121" }, marker = "(extra == 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126') or (extra == 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu128') or (extra == 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-nemo') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-nemo') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-vibevoice') or (extra == 'extra-12-asr-services-cu121' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-transformers') or (extra == 'extra-12-asr-services-cu121' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-vibevoice') or (extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-transformers') or (extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-vibevoice') or (extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-transformers' and extra == 'group-12-asr-services-vibevoice')" }, - { name = "torch", version = "2.7.1", source = { registry = "https://pypi.org/simple" }, marker = "(extra == 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126') or (extra == 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu128') or (extra == 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128') or (extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-transformers') or (extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-vibevoice') or (extra == 'group-12-asr-services-transformers' and extra == 'group-12-asr-services-vibevoice') or (extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-nemo') or (extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr') or (extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-transformers') or (extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-vibevoice') or (extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr') or (extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-transformers') or (extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-vibevoice') or (extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-transformers') or (extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-vibevoice')" }, - { name = "torch", version = "2.9.1+cu126", source = { registry = "https://download.pytorch.org/whl/cu126" }, marker = "(extra == 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126') or (extra == 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu128') or (extra == 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-nemo') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-transformers' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-nemo') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-vibevoice') or (extra == 'extra-12-asr-services-cu126' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-transformers') or (extra == 'extra-12-asr-services-cu126' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu121' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu121' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu121' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-transformers' and extra == 'group-12-asr-services-vibevoice')" }, - { name = "torch", version = "2.9.1+cu128", source = { registry = "https://download.pytorch.org/whl/cu128" }, marker = "(extra == 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126') or (extra == 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu128') or (extra == 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128') or (extra == 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-transformers') or (extra == 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-vibevoice') or (extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-nemo') or (extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr') or (extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-transformers') or (extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-vibevoice') or (extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr') or (extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-transformers') or (extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-vibevoice') or (extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-transformers') or (extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-vibevoice') or (extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-transformers') or (extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-vibevoice') or (extra == 'group-12-asr-services-transformers' and extra == 'group-12-asr-services-vibevoice')" }, -] -sdist = { url = "https://files.pythonhosted.org/packages/97/33/47bbd507e3a851d33d19ce7b2141c5ea3689bfae91ba168044d7db24b0e9/accelerate-1.7.0.tar.gz", hash = "sha256:e8a2a5503d6237b9eee73cc8d36cf543f9c2d8dd2c6713450b322f5e6d53a610", size = 376026, upload_time = "2025-05-15T10:00:52.117Z" } -wheels = [ - { url = "https://files.pythonhosted.org/packages/f8/bb/be8146c196ad6e4dec78385d91e92591f8a433576c4e04c342a636fcd811/accelerate-1.7.0-py3-none-any.whl", hash = "sha256:cf57165cca28769c6cf2650812371c81b18e05743dfa3c748524b1bb4f2b272f", size = 362095, upload_time = "2025-05-15T10:00:49.914Z" }, -] - [[package]] name = "accelerate" version = "1.12.0" source = { registry = "https://pypi.org/simple" } -resolution-markers = [ - "python_full_version >= '3.13' and platform_machine != 'aarch64' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version >= '3.13' and sys_platform != 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version == '3.12.*' and platform_machine != 'aarch64' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version == '3.12.*' and sys_platform != 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version >= '3.13' and platform_machine == 'aarch64' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version == '3.12.*' and platform_machine == 'aarch64' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version == '3.11.*' and platform_machine != 'aarch64' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version == '3.11.*' and sys_platform != 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version < '3.11' and platform_machine != 'aarch64' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version < '3.11' and sys_platform != 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version == '3.11.*' and platform_machine == 'aarch64' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version < '3.11' and platform_machine == 'aarch64' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version >= '3.13' and platform_machine != 'aarch64' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version >= '3.13' and sys_platform != 'linux' and extra != 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version == '3.12.*' and platform_machine != 'aarch64' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version == '3.12.*' and sys_platform != 'linux' and extra != 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version >= '3.13' and platform_machine == 'aarch64' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version == '3.12.*' and platform_machine == 'aarch64' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version == '3.11.*' and platform_machine != 'aarch64' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version == '3.11.*' and sys_platform != 'linux' and extra != 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version < '3.11' and platform_machine != 'aarch64' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version < '3.11' and sys_platform != 'linux' and extra != 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version == '3.11.*' and platform_machine == 'aarch64' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version < '3.11' and platform_machine == 'aarch64' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version >= '3.13' and sys_platform == 'linux' and extra == 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version >= '3.13' and sys_platform != 'linux' and extra == 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version == '3.12.*' and sys_platform == 'linux' and extra == 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version == '3.12.*' and sys_platform != 'linux' and extra == 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version == '3.11.*' and sys_platform == 'linux' and extra == 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version == '3.11.*' and sys_platform != 'linux' and extra == 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version < '3.11' and sys_platform == 'linux' and extra == 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version < '3.11' and sys_platform != 'linux' and extra == 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version >= '3.13' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version >= '3.13' and sys_platform != 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version == '3.12.*' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version == '3.12.*' and sys_platform != 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version == '3.11.*' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version == '3.11.*' and sys_platform != 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version < '3.11' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version < '3.11' and sys_platform != 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version >= '3.13' and platform_machine != 'aarch64' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version >= '3.13' and sys_platform != 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version == '3.12.*' and platform_machine != 'aarch64' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version == '3.12.*' and sys_platform != 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version >= '3.13' and platform_machine == 'aarch64' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version == '3.12.*' and platform_machine == 'aarch64' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version == '3.11.*' and platform_machine != 'aarch64' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version == '3.11.*' and sys_platform != 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version < '3.11' and platform_machine != 'aarch64' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version < '3.11' and sys_platform != 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version == '3.11.*' and platform_machine == 'aarch64' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version < '3.11' and platform_machine == 'aarch64' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version >= '3.13' and platform_machine != 'aarch64' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version >= '3.13' and sys_platform != 'linux' and extra != 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version == '3.12.*' and platform_machine != 'aarch64' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version == '3.12.*' and sys_platform != 'linux' and extra != 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version >= '3.13' and platform_machine == 'aarch64' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version == '3.12.*' and platform_machine == 'aarch64' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version == '3.11.*' and platform_machine != 'aarch64' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version == '3.11.*' and sys_platform != 'linux' and extra != 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version < '3.11' and platform_machine != 'aarch64' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version < '3.11' and sys_platform != 'linux' and extra != 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version == '3.11.*' and platform_machine == 'aarch64' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version < '3.11' and platform_machine == 'aarch64' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version >= '3.13' and sys_platform == 'linux' and extra == 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version >= '3.13' and sys_platform != 'linux' and extra == 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version == '3.12.*' and sys_platform == 'linux' and extra == 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version == '3.12.*' and sys_platform != 'linux' and extra == 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version == '3.11.*' and sys_platform == 'linux' and extra == 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version == '3.11.*' and sys_platform != 'linux' and extra == 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version < '3.11' and sys_platform == 'linux' and extra == 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version < '3.11' and sys_platform != 'linux' and extra == 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version >= '3.13' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version >= '3.13' and sys_platform != 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version == '3.12.*' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version == '3.12.*' and sys_platform != 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version == '3.11.*' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version == '3.11.*' and sys_platform != 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version < '3.11' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version < '3.11' and sys_platform != 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version >= '3.13' and platform_machine != 'aarch64' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version >= '3.13' and sys_platform != 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version == '3.12.*' and platform_machine != 'aarch64' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version == '3.12.*' and sys_platform != 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version >= '3.13' and platform_machine == 'aarch64' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version == '3.12.*' and platform_machine == 'aarch64' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version == '3.11.*' and platform_machine != 'aarch64' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version == '3.11.*' and sys_platform != 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version < '3.11' and platform_machine != 'aarch64' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version < '3.11' and sys_platform != 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version == '3.11.*' and platform_machine == 'aarch64' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version < '3.11' and platform_machine == 'aarch64' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version >= '3.13' and platform_machine != 'aarch64' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version >= '3.13' and sys_platform != 'linux' and extra != 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version == '3.12.*' and platform_machine != 'aarch64' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version == '3.12.*' and sys_platform != 'linux' and extra != 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version >= '3.13' and platform_machine == 'aarch64' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version == '3.12.*' and platform_machine == 'aarch64' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version == '3.11.*' and platform_machine != 'aarch64' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version == '3.11.*' and sys_platform != 'linux' and extra != 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version < '3.11' and platform_machine != 'aarch64' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version < '3.11' and sys_platform != 'linux' and extra != 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version == '3.11.*' and platform_machine == 'aarch64' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version < '3.11' and platform_machine == 'aarch64' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version >= '3.13' and sys_platform == 'linux' and extra == 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version >= '3.13' and sys_platform != 'linux' and extra == 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version == '3.12.*' and sys_platform == 'linux' and extra == 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version == '3.12.*' and sys_platform != 'linux' and extra == 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version == '3.11.*' and sys_platform == 'linux' and extra == 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version == '3.11.*' and sys_platform != 'linux' and extra == 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version < '3.11' and sys_platform == 'linux' and extra == 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version < '3.11' and sys_platform != 'linux' and extra == 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version >= '3.13' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version >= '3.13' and sys_platform != 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version == '3.12.*' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version == '3.12.*' and sys_platform != 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version == '3.11.*' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version == '3.11.*' and sys_platform != 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version < '3.11' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version < '3.11' and sys_platform != 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version >= '3.13' and platform_machine != 'aarch64' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version >= '3.13' and sys_platform != 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version == '3.12.*' and platform_machine != 'aarch64' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version == '3.12.*' and sys_platform != 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version >= '3.13' and platform_machine == 'aarch64' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version == '3.12.*' and platform_machine == 'aarch64' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version == '3.11.*' and platform_machine != 'aarch64' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version == '3.11.*' and sys_platform != 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version < '3.11' and platform_machine != 'aarch64' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version < '3.11' and sys_platform != 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version == '3.11.*' and platform_machine == 'aarch64' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version < '3.11' and platform_machine == 'aarch64' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version >= '3.13' and platform_machine != 'aarch64' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version >= '3.13' and sys_platform != 'linux' and extra != 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version == '3.12.*' and platform_machine != 'aarch64' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version == '3.12.*' and sys_platform != 'linux' and extra != 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version >= '3.13' and platform_machine == 'aarch64' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version == '3.12.*' and platform_machine == 'aarch64' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version == '3.11.*' and platform_machine != 'aarch64' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version == '3.11.*' and sys_platform != 'linux' and extra != 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version < '3.11' and platform_machine != 'aarch64' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version < '3.11' and sys_platform != 'linux' and extra != 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version == '3.11.*' and platform_machine == 'aarch64' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version < '3.11' and platform_machine == 'aarch64' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version >= '3.13' and sys_platform == 'linux' and extra == 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version >= '3.13' and sys_platform != 'linux' and extra == 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version == '3.12.*' and sys_platform == 'linux' and extra == 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version == '3.12.*' and sys_platform != 'linux' and extra == 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version == '3.11.*' and sys_platform == 'linux' and extra == 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version == '3.11.*' and sys_platform != 'linux' and extra == 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version < '3.11' and sys_platform == 'linux' and extra == 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version < '3.11' and sys_platform != 'linux' and extra == 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version >= '3.13' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version >= '3.13' and sys_platform != 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version == '3.12.*' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version == '3.12.*' and sys_platform != 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version == '3.11.*' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version == '3.11.*' and sys_platform != 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version < '3.11' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version < '3.11' and sys_platform != 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version >= '3.13' and platform_machine != 'aarch64' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version >= '3.13' and sys_platform != 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version == '3.12.*' and platform_machine != 'aarch64' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version == '3.12.*' and sys_platform != 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version >= '3.13' and platform_machine == 'aarch64' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version == '3.12.*' and platform_machine == 'aarch64' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version == '3.11.*' and platform_machine != 'aarch64' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version == '3.11.*' and sys_platform != 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version < '3.11' and platform_machine != 'aarch64' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version < '3.11' and sys_platform != 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version == '3.11.*' and platform_machine == 'aarch64' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version < '3.11' and platform_machine == 'aarch64' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version >= '3.13' and platform_machine != 'aarch64' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version >= '3.13' and sys_platform != 'linux' and extra != 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version == '3.12.*' and platform_machine != 'aarch64' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version == '3.12.*' and sys_platform != 'linux' and extra != 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version >= '3.13' and platform_machine == 'aarch64' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version == '3.12.*' and platform_machine == 'aarch64' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version == '3.11.*' and platform_machine != 'aarch64' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version == '3.11.*' and sys_platform != 'linux' and extra != 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version < '3.11' and platform_machine != 'aarch64' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version < '3.11' and sys_platform != 'linux' and extra != 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version == '3.11.*' and platform_machine == 'aarch64' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version < '3.11' and platform_machine == 'aarch64' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version >= '3.13' and sys_platform == 'linux' and extra == 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version >= '3.13' and sys_platform != 'linux' and extra == 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version == '3.12.*' and sys_platform == 'linux' and extra == 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version == '3.12.*' and sys_platform != 'linux' and extra == 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version == '3.11.*' and sys_platform == 'linux' and extra == 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version == '3.11.*' and sys_platform != 'linux' and extra == 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version < '3.11' and sys_platform == 'linux' and extra == 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version < '3.11' and sys_platform != 'linux' and extra == 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version >= '3.13' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version >= '3.13' and sys_platform != 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version == '3.12.*' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version == '3.12.*' and sys_platform != 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version == '3.11.*' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version == '3.11.*' and sys_platform != 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version < '3.11' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", - "python_full_version < '3.11' and sys_platform != 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", -] dependencies = [ - { name = "huggingface-hub", marker = "(extra == 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126') or (extra == 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu128') or (extra == 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-nemo') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-qwen3-asr') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-transformers' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-faster-whisper') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-nemo') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-qwen3-asr') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-transformers' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice')" }, - { name = "numpy", marker = "(extra == 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126') or (extra == 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu128') or (extra == 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-nemo') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-qwen3-asr') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-transformers' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-faster-whisper') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-nemo') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-qwen3-asr') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-transformers' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice')" }, - { name = "packaging", marker = "(extra == 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126') or (extra == 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu128') or (extra == 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-nemo') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-qwen3-asr') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-transformers' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-faster-whisper') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-nemo') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-qwen3-asr') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-transformers' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice')" }, - { name = "psutil", marker = "(extra == 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126') or (extra == 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu128') or (extra == 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-nemo') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-qwen3-asr') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-transformers' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-faster-whisper') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-nemo') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-qwen3-asr') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-transformers' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice')" }, - { name = "pyyaml", marker = "(extra == 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126') or (extra == 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu128') or (extra == 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-nemo') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-qwen3-asr') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-transformers' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-faster-whisper') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-nemo') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-qwen3-asr') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-transformers' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice')" }, - { name = "safetensors", marker = "(extra == 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126') or (extra == 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu128') or (extra == 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-nemo') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-qwen3-asr') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-transformers' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-faster-whisper') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-nemo') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-qwen3-asr') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-transformers' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice')" }, - { name = "torch", version = "2.5.1+cu121", source = { registry = "https://download.pytorch.org/whl/cu121" }, marker = "(extra == 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126') or (extra == 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu128') or (extra == 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-faster-whisper') or (extra == 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-nemo') or (extra == 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-qwen3-asr') or (extra == 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra == 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128') or (extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-nemo') or (extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr') or (extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-transformers') or (extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-vibevoice') or (extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr') or (extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-transformers') or (extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-vibevoice') or (extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-transformers') or (extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-vibevoice') or (extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-transformers') or (extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-vibevoice') or (extra == 'extra-12-asr-services-cu121' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-transformers' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-transformers' and extra == 'group-12-asr-services-vibevoice')" }, - { name = "torch", version = "2.7.1", source = { registry = "https://pypi.org/simple" }, marker = "(extra == 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126') or (extra == 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu128') or (extra == 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128') or (extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-nemo') or (extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr') or (extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-transformers') or (extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-vibevoice') or (extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr') or (extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-transformers') or (extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-vibevoice') or (extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-transformers') or (extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-vibevoice') or (extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-transformers') or (extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-vibevoice') or (extra == 'group-12-asr-services-transformers' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-nemo') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-qwen3-asr') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice')" }, - { name = "torch", version = "2.9.1+cu126", source = { registry = "https://download.pytorch.org/whl/cu126" }, marker = "(extra == 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126') or (extra == 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu128') or (extra == 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128') or (extra == 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-faster-whisper') or (extra == 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-nemo') or (extra == 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-qwen3-asr') or (extra == 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra == 'extra-12-asr-services-cu126' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-nemo') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-nemo') or (extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr') or (extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr') or (extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-transformers' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-transformers' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-transformers' and extra == 'group-12-asr-services-vibevoice')" }, - { name = "torch", version = "2.9.1+cu128", source = { registry = "https://download.pytorch.org/whl/cu128" }, marker = "(extra == 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126') or (extra == 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu128') or (extra == 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128') or (extra == 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper') or (extra == 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-nemo') or (extra == 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-qwen3-asr') or (extra == 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-nemo') or (extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr') or (extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-transformers') or (extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-vibevoice') or (extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr') or (extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-transformers') or (extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-vibevoice') or (extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-transformers') or (extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-vibevoice') or (extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-transformers') or (extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-vibevoice') or (extra == 'group-12-asr-services-transformers' and extra == 'group-12-asr-services-vibevoice') or (extra == 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice')" }, + { name = "huggingface-hub" }, + { name = "numpy" }, + { name = "packaging" }, + { name = "psutil" }, + { name = "pyyaml" }, + { name = "safetensors" }, + { name = "torch", version = "2.5.1+cu121", source = { registry = "https://download.pytorch.org/whl/cu121" }, marker = "extra == 'extra-12-asr-services-cu121' or (extra == 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128') or (extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-nemo') or (extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr') or (extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-transformers') or (extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-vibevoice') or (extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr') or (extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-transformers') or (extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-vibevoice') or (extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-transformers') or (extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-vibevoice') or (extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-transformers') or (extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-vibevoice') or (extra == 'group-12-asr-services-transformers' and extra == 'group-12-asr-services-vibevoice')" }, + { name = "torch", version = "2.7.1", source = { registry = "https://pypi.org/simple" }, marker = "(extra == 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126') or (extra == 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu128') or (extra == 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128') or (extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-nemo') or (extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr') or (extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-transformers') or (extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-vibevoice') or (extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr') or (extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-transformers') or (extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-vibevoice') or (extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-transformers') or (extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-vibevoice') or (extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-transformers') or (extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-vibevoice') or (extra == 'group-12-asr-services-transformers' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128')" }, + { name = "torch", version = "2.9.1+cu126", source = { registry = "https://download.pytorch.org/whl/cu126" }, marker = "extra == 'extra-12-asr-services-cu126' or (extra == 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu128') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-nemo') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-transformers' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-nemo') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-transformers' and extra == 'group-12-asr-services-vibevoice')" }, + { name = "torch", version = "2.9.1+cu128", source = { registry = "https://download.pytorch.org/whl/cu128" }, marker = "(extra == 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126') or (extra != 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128') or (extra != 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu128') or (extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-nemo') or (extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr') or (extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr') or (extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-transformers' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-nemo') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-transformers' and extra == 'group-12-asr-services-vibevoice')" }, ] sdist = { url = "https://files.pythonhosted.org/packages/4a/8e/ac2a9566747a93f8be36ee08532eb0160558b07630a081a6056a9f89bf1d/accelerate-1.12.0.tar.gz", hash = "sha256:70988c352feb481887077d2ab845125024b2a137a5090d6d7a32b57d03a45df6", size = 398399, upload_time = "2025-11-21T11:27:46.973Z" } wheels = [ @@ -903,7 +598,7 @@ qwen3-asr-full = [ { name = "websockets" }, ] transformers = [ - { name = "accelerate", version = "1.7.0", source = { registry = "https://pypi.org/simple" } }, + { name = "accelerate" }, { name = "diffusers" }, { name = "torch", version = "2.5.1+cu121", source = { registry = "https://download.pytorch.org/whl/cu121" }, marker = "(extra == 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126') or (extra == 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu128') or (extra == 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-nemo') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-nemo') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-vibevoice') or (extra == 'extra-12-asr-services-cu121' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-transformers') or (extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-transformers' and extra == 'group-12-asr-services-vibevoice')" }, { name = "torch", version = "2.7.1", source = { registry = "https://pypi.org/simple" }, marker = "(extra == 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126') or (extra == 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu128') or (extra == 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128') or (extra == 'group-12-asr-services-transformers' and extra == 'group-12-asr-services-vibevoice') or (extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-nemo') or (extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr') or (extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-transformers') or (extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-vibevoice') or (extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr') or (extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-transformers') or (extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-vibevoice') or (extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-transformers') or (extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-vibevoice') or (extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-transformers') or (extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-transformers')" }, @@ -918,10 +613,11 @@ transformers = [ { name = "transformers", version = "4.53.3", source = { registry = "https://pypi.org/simple" } }, ] vibevoice = [ - { name = "accelerate", version = "1.7.0", source = { registry = "https://pypi.org/simple" } }, + { name = "accelerate" }, { name = "bitsandbytes" }, { name = "diffusers" }, { name = "librosa" }, + { name = "omegaconf" }, { name = "peft" }, { name = "torch", version = "2.5.1+cu121", source = { registry = "https://download.pytorch.org/whl/cu121" }, marker = "(extra == 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126') or (extra == 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu128') or (extra == 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-nemo') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-nemo') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-vibevoice') or (extra == 'extra-12-asr-services-cu121' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-vibevoice') or (extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-transformers' and extra == 'group-12-asr-services-vibevoice')" }, { name = "torch", version = "2.7.1", source = { registry = "https://pypi.org/simple" }, marker = "(extra == 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126') or (extra == 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu128') or (extra == 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128') or (extra == 'group-12-asr-services-transformers' and extra == 'group-12-asr-services-vibevoice') or (extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-nemo') or (extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr') or (extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-transformers') or (extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-vibevoice') or (extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr') or (extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-transformers') or (extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-vibevoice') or (extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-transformers') or (extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-vibevoice') or (extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-transformers') or (extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-vibevoice')" }, @@ -1007,6 +703,7 @@ vibevoice = [ { name = "bitsandbytes", specifier = ">=0.43.0" }, { name = "diffusers", specifier = ">=0.30.0" }, { name = "librosa", specifier = ">=0.10.0" }, + { name = "omegaconf", specifier = ">=2.3.0" }, { name = "peft", specifier = ">=0.17.0" }, { name = "torch", specifier = ">=2.3" }, { name = "torchaudio", specifier = ">=2.3" }, @@ -1912,7 +1609,7 @@ name = "exceptiongroup" version = "1.3.0" source = { registry = "https://pypi.org/simple" } dependencies = [ - { name = "typing-extensions", marker = "python_full_version < '3.12' or (python_full_version == '3.12.*' and extra == 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126') or (python_full_version == '3.12.*' and extra == 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu128') or (python_full_version == '3.12.*' and extra == 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128') or (python_full_version >= '3.13' and extra == 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126') or (python_full_version >= '3.13' and extra == 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu128') or (python_full_version >= '3.13' and extra == 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128') or (python_full_version == '3.12.*' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128')" }, + { name = "typing-extensions", marker = "python_full_version < '3.11' or (extra == 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126') or (extra == 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu128') or (extra == 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128')" }, ] sdist = { url = "https://files.pythonhosted.org/packages/0b/9f/a65090624ecf468cdca03533906e7c69ed7588582240cfe7cc9e770b50eb/exceptiongroup-1.3.0.tar.gz", hash = "sha256:b241f5885f560bc56a59ee63ca4c6a8bfa46ae4ad651af316d4e81817bb9fd88", size = 29749, upload_time = "2025-05-10T17:42:51.123Z" } wheels = [ @@ -5447,7 +5144,7 @@ resolution-markers = [ "python_full_version < '3.11' and sys_platform == 'linux'", ] dependencies = [ - { name = "nvidia-cublas-cu12", version = "12.1.3.1", source = { registry = "https://pypi.org/simple" }, marker = "extra == 'extra-12-asr-services-cu121' or (extra == 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128') or (extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-nemo') or (extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr') or (extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-transformers') or (extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-vibevoice') or (extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr') or (extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-transformers') or (extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-vibevoice') or (extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-transformers') or (extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-vibevoice') or (extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-transformers') or (extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-vibevoice') or (extra == 'group-12-asr-services-transformers' and extra == 'group-12-asr-services-vibevoice')" }, + { name = "nvidia-cublas-cu12", version = "12.1.3.1", source = { registry = "https://pypi.org/simple" }, marker = "(sys_platform == 'linux' and extra == 'extra-12-asr-services-cu121') or (extra == 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126') or (extra == 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu128') or (extra == 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128') or (extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-nemo') or (extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr') or (extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-transformers') or (extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-vibevoice') or (extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr') or (extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-transformers') or (extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-vibevoice') or (extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-transformers') or (extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-vibevoice') or (extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-transformers') or (extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-vibevoice') or (extra == 'group-12-asr-services-transformers' and extra == 'group-12-asr-services-vibevoice')" }, ] wheels = [ { url = "https://files.pythonhosted.org/packages/9f/fd/713452cd72343f682b1c7b9321e23829f00b842ceaedcda96e742ea0b0b3/nvidia_cudnn_cu12-9.1.0.70-py3-none-manylinux2014_x86_64.whl", hash = "sha256:165764f44ef8c61fcdfdfdbe769d687e06374059fbb388b6c89ecb0e28793a6f", size = 664752741, upload_time = "2024-04-22T15:24:15.253Z" }, @@ -5465,7 +5162,7 @@ resolution-markers = [ "python_full_version < '3.11' and sys_platform == 'linux'", ] dependencies = [ - { name = "nvidia-cublas-cu12", version = "12.6.4.1", source = { registry = "https://pypi.org/simple" }, marker = "(extra == 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126') or (extra == 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu128') or (extra == 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128')" }, + { name = "nvidia-cublas-cu12", version = "12.6.4.1", source = { registry = "https://pypi.org/simple" }, marker = "(sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128') or (extra == 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126') or (extra == 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu128') or (extra == 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128')" }, ] wheels = [ { url = "https://files.pythonhosted.org/packages/99/93/a201a12d3ec1caa8c6ac34c1c2f9eeb696b886f0c36ff23c638b46603bd0/nvidia_cudnn_cu12-9.5.1.17-py3-none-manylinux_2_28_aarch64.whl", hash = "sha256:9fd4584468533c61873e5fda8ca41bac3a38bcb2d12350830c69b0a96a7e4def", size = 570523509, upload_time = "2024-10-25T19:53:03.148Z" }, @@ -5488,8 +5185,8 @@ resolution-markers = [ "python_full_version < '3.11' and platform_machine == 'aarch64' and sys_platform == 'linux'", ] dependencies = [ - { name = "nvidia-cublas-cu12", version = "12.6.4.1", source = { registry = "https://pypi.org/simple" }, marker = "extra == 'extra-12-asr-services-cu126' or (extra == 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu128') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-nemo') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-transformers' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-nemo') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-transformers' and extra == 'group-12-asr-services-vibevoice')" }, - { name = "nvidia-cublas-cu12", version = "12.8.4.1", source = { registry = "https://pypi.org/simple" }, marker = "(extra == 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126') or (extra != 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128') or (extra != 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu128') or (extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-nemo') or (extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr') or (extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr') or (extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-transformers' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-nemo') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-transformers' and extra == 'group-12-asr-services-vibevoice')" }, + { name = "nvidia-cublas-cu12", version = "12.6.4.1", source = { registry = "https://pypi.org/simple" }, marker = "(sys_platform == 'linux' and extra == 'extra-12-asr-services-cu126') or (extra == 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128') or (extra == 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126') or (extra == 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu128') or (extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-nemo') or (extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr') or (extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-transformers') or (extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-vibevoice') or (extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr') or (extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-transformers') or (extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-vibevoice') or (extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-transformers') or (extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-vibevoice') or (extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-transformers') or (extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-vibevoice') or (extra == 'group-12-asr-services-transformers' and extra == 'group-12-asr-services-vibevoice')" }, + { name = "nvidia-cublas-cu12", version = "12.8.4.1", source = { registry = "https://pypi.org/simple" }, marker = "(sys_platform == 'linux' and extra == 'extra-12-asr-services-cu128') or (sys_platform != 'linux' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-nemo') or (sys_platform != 'linux' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr') or (sys_platform != 'linux' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr-full') or (sys_platform != 'linux' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-transformers') or (sys_platform != 'linux' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-vibevoice') or (sys_platform != 'linux' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr') or (sys_platform != 'linux' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr-full') or (sys_platform != 'linux' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-transformers') or (sys_platform != 'linux' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-vibevoice') or (sys_platform != 'linux' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full') or (sys_platform != 'linux' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-transformers') or (sys_platform != 'linux' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-vibevoice') or (sys_platform != 'linux' and extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-transformers') or (sys_platform != 'linux' and extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-vibevoice') or (sys_platform != 'linux' and extra == 'group-12-asr-services-transformers' and extra == 'group-12-asr-services-vibevoice') or (extra == 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128') or (extra == 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126') or (extra == 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu128') or (extra == 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-nemo') or (extra == 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr') or (extra == 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra == 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-transformers') or (extra == 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-vibevoice') or (extra == 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr') or (extra == 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra == 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-transformers') or (extra == 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-vibevoice') or (extra == 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra == 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-transformers') or (extra == 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-vibevoice') or (extra == 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-transformers') or (extra == 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-vibevoice') or (extra == 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-transformers' and extra == 'group-12-asr-services-vibevoice') or (extra == 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-nemo') or (extra == 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr') or (extra == 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra == 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-transformers') or (extra == 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-vibevoice') or (extra == 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr') or (extra == 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra == 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-transformers') or (extra == 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-vibevoice') or (extra == 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra == 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-transformers') or (extra == 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-vibevoice') or (extra == 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-transformers') or (extra == 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-vibevoice') or (extra == 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-transformers' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-nemo') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-transformers' and extra == 'group-12-asr-services-vibevoice')" }, ] wheels = [ { url = "https://files.pythonhosted.org/packages/fa/41/e79269ce215c857c935fd86bcfe91a451a584dfc27f1e068f568b9ad1ab7/nvidia_cudnn_cu12-9.10.2.21-py3-none-manylinux_2_27_aarch64.whl", hash = "sha256:c9132cc3f8958447b4910a1720036d9eff5928cc3179b0a51fb6d167c6cc87d8", size = 705026878, upload_time = "2025-06-06T21:52:51.348Z" }, @@ -5603,7 +5300,7 @@ resolution-markers = [ "python_full_version < '3.11' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", ] dependencies = [ - { name = "nvidia-nvjitlink-cu12", version = "12.6.85", source = { registry = "https://pypi.org/simple" } }, + { name = "nvidia-nvjitlink-cu12", version = "12.6.85", source = { registry = "https://pypi.org/simple" }, marker = "sys_platform == 'linux'" }, ] wheels = [ { url = "https://files.pythonhosted.org/packages/1f/37/c50d2b2f2c07e146776389e3080f4faf70bcc4fa6e19d65bb54ca174ebc3/nvidia_cufft_cu12-11.3.0.4-py3-none-manylinux2014_aarch64.manylinux_2_17_aarch64.whl", hash = "sha256:d16079550df460376455cba121db6564089176d9bac9e4f360493ca4741b22a6", size = 200164144, upload_time = "2024-11-20T17:40:58.288Z" }, @@ -5628,7 +5325,7 @@ resolution-markers = [ "python_full_version < '3.11' and platform_machine == 'aarch64' and sys_platform == 'linux'", ] dependencies = [ - { name = "nvidia-nvjitlink-cu12", version = "12.8.93", source = { registry = "https://pypi.org/simple" }, marker = "(extra == 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126') or (extra != 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128') or (extra != 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu128') or (extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-nemo') or (extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr') or (extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr') or (extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-transformers' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-nemo') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-transformers' and extra == 'group-12-asr-services-vibevoice')" }, + { name = "nvidia-nvjitlink-cu12", version = "12.8.93", source = { registry = "https://pypi.org/simple" }, marker = "(sys_platform == 'linux' and extra == 'extra-12-asr-services-cu128') or (sys_platform != 'linux' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-nemo') or (sys_platform != 'linux' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr') or (sys_platform != 'linux' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr-full') or (sys_platform != 'linux' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-transformers') or (sys_platform != 'linux' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-vibevoice') or (sys_platform != 'linux' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr') or (sys_platform != 'linux' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr-full') or (sys_platform != 'linux' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-transformers') or (sys_platform != 'linux' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-vibevoice') or (sys_platform != 'linux' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full') or (sys_platform != 'linux' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-transformers') or (sys_platform != 'linux' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-vibevoice') or (sys_platform != 'linux' and extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-transformers') or (sys_platform != 'linux' and extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-vibevoice') or (sys_platform != 'linux' and extra == 'group-12-asr-services-transformers' and extra == 'group-12-asr-services-vibevoice') or (extra == 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128') or (extra == 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126') or (extra == 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu128') or (extra == 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-nemo') or (extra == 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr') or (extra == 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra == 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-transformers') or (extra == 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-vibevoice') or (extra == 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr') or (extra == 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra == 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-transformers') or (extra == 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-vibevoice') or (extra == 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra == 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-transformers') or (extra == 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-vibevoice') or (extra == 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-transformers') or (extra == 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-vibevoice') or (extra == 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-transformers' and extra == 'group-12-asr-services-vibevoice') or (extra == 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-nemo') or (extra == 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr') or (extra == 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra == 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-transformers') or (extra == 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-vibevoice') or (extra == 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr') or (extra == 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra == 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-transformers') or (extra == 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-vibevoice') or (extra == 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra == 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-transformers') or (extra == 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-vibevoice') or (extra == 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-transformers') or (extra == 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-vibevoice') or (extra == 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-transformers' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-nemo') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-transformers' and extra == 'group-12-asr-services-vibevoice')" }, ] wheels = [ { url = "https://files.pythonhosted.org/packages/60/bc/7771846d3a0272026c416fbb7e5f4c1f146d6d80704534d0b187dd6f4800/nvidia_cufft_cu12-11.3.3.83-py3-none-manylinux2014_aarch64.manylinux_2_17_aarch64.whl", hash = "sha256:848ef7224d6305cdb2a4df928759dca7b1201874787083b6e7550dd6765ce69a", size = 193109211, upload_time = "2025-03-07T01:44:56.873Z" }, @@ -5894,9 +5591,9 @@ resolution-markers = [ "python_full_version < '3.11' and sys_platform == 'linux'", ] dependencies = [ - { name = "nvidia-cublas-cu12", version = "12.1.3.1", source = { registry = "https://pypi.org/simple" }, marker = "extra == 'extra-12-asr-services-cu121' or (extra == 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128') or (extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-nemo') or (extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr') or (extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-transformers') or (extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-vibevoice') or (extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr') or (extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-transformers') or (extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-vibevoice') or (extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-transformers') or (extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-vibevoice') or (extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-transformers') or (extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-vibevoice') or (extra == 'group-12-asr-services-transformers' and extra == 'group-12-asr-services-vibevoice')" }, - { name = "nvidia-cusparse-cu12", version = "12.1.0.106", source = { registry = "https://pypi.org/simple" }, marker = "extra == 'extra-12-asr-services-cu121' or (extra == 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128') or (extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-nemo') or (extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr') or (extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-transformers') or (extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-vibevoice') or (extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr') or (extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-transformers') or (extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-vibevoice') or (extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-transformers') or (extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-vibevoice') or (extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-transformers') or (extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-vibevoice') or (extra == 'group-12-asr-services-transformers' and extra == 'group-12-asr-services-vibevoice')" }, - { name = "nvidia-nvjitlink-cu12", version = "12.8.93", source = { registry = "https://pypi.org/simple" }, marker = "extra == 'extra-12-asr-services-cu121' or (extra == 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128') or (extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-nemo') or (extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr') or (extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-transformers') or (extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-vibevoice') or (extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr') or (extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-transformers') or (extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-vibevoice') or (extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-transformers') or (extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-vibevoice') or (extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-transformers') or (extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-vibevoice') or (extra == 'group-12-asr-services-transformers' and extra == 'group-12-asr-services-vibevoice')" }, + { name = "nvidia-cublas-cu12", version = "12.1.3.1", source = { registry = "https://pypi.org/simple" }, marker = "(sys_platform == 'linux' and extra == 'extra-12-asr-services-cu121') or (extra == 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126') or (extra == 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu128') or (extra == 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128') or (extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-nemo') or (extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr') or (extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-transformers') or (extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-vibevoice') or (extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr') or (extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-transformers') or (extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-vibevoice') or (extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-transformers') or (extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-vibevoice') or (extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-transformers') or (extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-vibevoice') or (extra == 'group-12-asr-services-transformers' and extra == 'group-12-asr-services-vibevoice')" }, + { name = "nvidia-cusparse-cu12", version = "12.1.0.106", source = { registry = "https://pypi.org/simple" }, marker = "(sys_platform == 'linux' and extra == 'extra-12-asr-services-cu121') or (extra == 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126') or (extra == 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu128') or (extra == 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128') or (extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-nemo') or (extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr') or (extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-transformers') or (extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-vibevoice') or (extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr') or (extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-transformers') or (extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-vibevoice') or (extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-transformers') or (extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-vibevoice') or (extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-transformers') or (extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-vibevoice') or (extra == 'group-12-asr-services-transformers' and extra == 'group-12-asr-services-vibevoice')" }, + { name = "nvidia-nvjitlink-cu12", version = "12.8.93", source = { registry = "https://pypi.org/simple" }, marker = "(sys_platform == 'linux' and extra == 'extra-12-asr-services-cu121') or (extra == 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126') or (extra == 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu128') or (extra == 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128') or (extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-nemo') or (extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr') or (extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-transformers') or (extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-vibevoice') or (extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr') or (extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-transformers') or (extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-vibevoice') or (extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-transformers') or (extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-vibevoice') or (extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-transformers') or (extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-vibevoice') or (extra == 'group-12-asr-services-transformers' and extra == 'group-12-asr-services-vibevoice')" }, ] wheels = [ { url = "https://files.pythonhosted.org/packages/bc/1d/8de1e5c67099015c834315e333911273a8c6aaba78923dd1d1e25fc5f217/nvidia_cusolver_cu12-11.4.5.107-py3-none-manylinux1_x86_64.whl", hash = "sha256:8a7ec542f0412294b15072fa7dab71d31334014a69f953004ea7a118206fe0dd", size = 124161928, upload_time = "2023-04-19T15:51:25.781Z" }, @@ -5994,9 +5691,9 @@ resolution-markers = [ "python_full_version < '3.11' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", ] dependencies = [ - { name = "nvidia-cublas-cu12", version = "12.6.4.1", source = { registry = "https://pypi.org/simple" } }, - { name = "nvidia-cusparse-cu12", version = "12.5.4.2", source = { registry = "https://pypi.org/simple" } }, - { name = "nvidia-nvjitlink-cu12", version = "12.6.85", source = { registry = "https://pypi.org/simple" } }, + { name = "nvidia-cublas-cu12", version = "12.6.4.1", source = { registry = "https://pypi.org/simple" }, marker = "sys_platform == 'linux'" }, + { name = "nvidia-cusparse-cu12", version = "12.5.4.2", source = { registry = "https://pypi.org/simple" }, marker = "sys_platform == 'linux'" }, + { name = "nvidia-nvjitlink-cu12", version = "12.6.85", source = { registry = "https://pypi.org/simple" }, marker = "sys_platform == 'linux'" }, ] wheels = [ { url = "https://files.pythonhosted.org/packages/93/17/dbe1aa865e4fdc7b6d4d0dd308fdd5aaab60f939abfc0ea1954eac4fb113/nvidia_cusolver_cu12-11.7.1.2-py3-none-manylinux2014_aarch64.whl", hash = "sha256:0ce237ef60acde1efc457335a2ddadfd7610b892d94efee7b776c64bb1cac9e0", size = 157833628, upload_time = "2024-10-01T17:05:05.591Z" }, @@ -6021,9 +5718,9 @@ resolution-markers = [ "python_full_version < '3.11' and platform_machine == 'aarch64' and sys_platform == 'linux'", ] dependencies = [ - { name = "nvidia-cublas-cu12", version = "12.8.4.1", source = { registry = "https://pypi.org/simple" }, marker = "(extra == 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126') or (extra != 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128') or (extra != 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu128') or (extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-nemo') or (extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr') or (extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr') or (extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-transformers' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-nemo') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-transformers' and extra == 'group-12-asr-services-vibevoice')" }, - { name = "nvidia-cusparse-cu12", version = "12.5.8.93", source = { registry = "https://pypi.org/simple" }, marker = "(extra == 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126') or (extra != 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128') or (extra != 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu128') or (extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-nemo') or (extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr') or (extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr') or (extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-transformers' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-nemo') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-transformers' and extra == 'group-12-asr-services-vibevoice')" }, - { name = "nvidia-nvjitlink-cu12", version = "12.8.93", source = { registry = "https://pypi.org/simple" }, marker = "(extra == 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126') or (extra != 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128') or (extra != 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu128') or (extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-nemo') or (extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr') or (extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr') or (extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-transformers' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-nemo') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-transformers' and extra == 'group-12-asr-services-vibevoice')" }, + { name = "nvidia-cublas-cu12", version = "12.8.4.1", source = { registry = "https://pypi.org/simple" }, marker = "(sys_platform == 'linux' and extra == 'extra-12-asr-services-cu128') or (sys_platform != 'linux' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-nemo') or (sys_platform != 'linux' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr') or (sys_platform != 'linux' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr-full') or (sys_platform != 'linux' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-transformers') or (sys_platform != 'linux' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-vibevoice') or (sys_platform != 'linux' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr') or (sys_platform != 'linux' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr-full') or (sys_platform != 'linux' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-transformers') or (sys_platform != 'linux' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-vibevoice') or (sys_platform != 'linux' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full') or (sys_platform != 'linux' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-transformers') or (sys_platform != 'linux' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-vibevoice') or (sys_platform != 'linux' and extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-transformers') or (sys_platform != 'linux' and extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-vibevoice') or (sys_platform != 'linux' and extra == 'group-12-asr-services-transformers' and extra == 'group-12-asr-services-vibevoice') or (extra == 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128') or (extra == 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126') or (extra == 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu128') or (extra == 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-nemo') or (extra == 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr') or (extra == 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra == 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-transformers') or (extra == 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-vibevoice') or (extra == 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr') or (extra == 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra == 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-transformers') or (extra == 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-vibevoice') or (extra == 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra == 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-transformers') or (extra == 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-vibevoice') or (extra == 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-transformers') or (extra == 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-vibevoice') or (extra == 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-transformers' and extra == 'group-12-asr-services-vibevoice') or (extra == 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-nemo') or (extra == 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr') or (extra == 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra == 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-transformers') or (extra == 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-vibevoice') or (extra == 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr') or (extra == 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra == 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-transformers') or (extra == 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-vibevoice') or (extra == 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra == 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-transformers') or (extra == 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-vibevoice') or (extra == 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-transformers') or (extra == 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-vibevoice') or (extra == 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-transformers' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-nemo') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-transformers' and extra == 'group-12-asr-services-vibevoice')" }, + { name = "nvidia-cusparse-cu12", version = "12.5.8.93", source = { registry = "https://pypi.org/simple" }, marker = "(sys_platform == 'linux' and extra == 'extra-12-asr-services-cu128') or (sys_platform != 'linux' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-nemo') or (sys_platform != 'linux' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr') or (sys_platform != 'linux' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr-full') or (sys_platform != 'linux' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-transformers') or (sys_platform != 'linux' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-vibevoice') or (sys_platform != 'linux' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr') or (sys_platform != 'linux' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr-full') or (sys_platform != 'linux' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-transformers') or (sys_platform != 'linux' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-vibevoice') or (sys_platform != 'linux' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full') or (sys_platform != 'linux' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-transformers') or (sys_platform != 'linux' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-vibevoice') or (sys_platform != 'linux' and extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-transformers') or (sys_platform != 'linux' and extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-vibevoice') or (sys_platform != 'linux' and extra == 'group-12-asr-services-transformers' and extra == 'group-12-asr-services-vibevoice') or (extra == 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128') or (extra == 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126') or (extra == 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu128') or (extra == 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-nemo') or (extra == 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr') or (extra == 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra == 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-transformers') or (extra == 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-vibevoice') or (extra == 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr') or (extra == 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra == 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-transformers') or (extra == 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-vibevoice') or (extra == 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra == 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-transformers') or (extra == 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-vibevoice') or (extra == 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-transformers') or (extra == 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-vibevoice') or (extra == 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-transformers' and extra == 'group-12-asr-services-vibevoice') or (extra == 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-nemo') or (extra == 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr') or (extra == 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra == 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-transformers') or (extra == 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-vibevoice') or (extra == 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr') or (extra == 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra == 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-transformers') or (extra == 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-vibevoice') or (extra == 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra == 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-transformers') or (extra == 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-vibevoice') or (extra == 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-transformers') or (extra == 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-vibevoice') or (extra == 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-transformers' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-nemo') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-transformers' and extra == 'group-12-asr-services-vibevoice')" }, + { name = "nvidia-nvjitlink-cu12", version = "12.8.93", source = { registry = "https://pypi.org/simple" }, marker = "(sys_platform == 'linux' and extra == 'extra-12-asr-services-cu128') or (sys_platform != 'linux' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-nemo') or (sys_platform != 'linux' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr') or (sys_platform != 'linux' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr-full') or (sys_platform != 'linux' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-transformers') or (sys_platform != 'linux' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-vibevoice') or (sys_platform != 'linux' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr') or (sys_platform != 'linux' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr-full') or (sys_platform != 'linux' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-transformers') or (sys_platform != 'linux' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-vibevoice') or (sys_platform != 'linux' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full') or (sys_platform != 'linux' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-transformers') or (sys_platform != 'linux' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-vibevoice') or (sys_platform != 'linux' and extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-transformers') or (sys_platform != 'linux' and extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-vibevoice') or (sys_platform != 'linux' and extra == 'group-12-asr-services-transformers' and extra == 'group-12-asr-services-vibevoice') or (extra == 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128') or (extra == 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126') or (extra == 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu128') or (extra == 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-nemo') or (extra == 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr') or (extra == 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra == 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-transformers') or (extra == 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-vibevoice') or (extra == 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr') or (extra == 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra == 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-transformers') or (extra == 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-vibevoice') or (extra == 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra == 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-transformers') or (extra == 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-vibevoice') or (extra == 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-transformers') or (extra == 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-vibevoice') or (extra == 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-transformers' and extra == 'group-12-asr-services-vibevoice') or (extra == 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-nemo') or (extra == 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr') or (extra == 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra == 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-transformers') or (extra == 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-vibevoice') or (extra == 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr') or (extra == 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra == 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-transformers') or (extra == 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-vibevoice') or (extra == 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra == 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-transformers') or (extra == 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-vibevoice') or (extra == 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-transformers') or (extra == 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-vibevoice') or (extra == 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-transformers' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-nemo') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-transformers' and extra == 'group-12-asr-services-vibevoice')" }, ] wheels = [ { url = "https://files.pythonhosted.org/packages/c8/32/f7cd6ce8a7690544d084ea21c26e910a97e077c9b7f07bf5de623ee19981/nvidia_cusolver_cu12-11.7.3.90-py3-none-manylinux_2_27_aarch64.whl", hash = "sha256:db9ed69dbef9715071232caa9b69c52ac7de3a95773c2db65bdba85916e4e5c0", size = 267229841, upload_time = "2025-03-07T01:46:54.356Z" }, @@ -6042,7 +5739,7 @@ resolution-markers = [ "python_full_version < '3.11' and sys_platform == 'linux'", ] dependencies = [ - { name = "nvidia-nvjitlink-cu12", version = "12.8.93", source = { registry = "https://pypi.org/simple" }, marker = "extra == 'extra-12-asr-services-cu121' or (extra == 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128') or (extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-nemo') or (extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr') or (extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-transformers') or (extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-vibevoice') or (extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr') or (extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-transformers') or (extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-vibevoice') or (extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-transformers') or (extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-vibevoice') or (extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-transformers') or (extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-vibevoice') or (extra == 'group-12-asr-services-transformers' and extra == 'group-12-asr-services-vibevoice')" }, + { name = "nvidia-nvjitlink-cu12", version = "12.8.93", source = { registry = "https://pypi.org/simple" }, marker = "(sys_platform == 'linux' and extra == 'extra-12-asr-services-cu121') or (extra == 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126') or (extra == 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu128') or (extra == 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128') or (extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-nemo') or (extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr') or (extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-transformers') or (extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-vibevoice') or (extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr') or (extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-transformers') or (extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-vibevoice') or (extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-transformers') or (extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-vibevoice') or (extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-transformers') or (extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-vibevoice') or (extra == 'group-12-asr-services-transformers' and extra == 'group-12-asr-services-vibevoice')" }, ] wheels = [ { url = "https://files.pythonhosted.org/packages/65/5b/cfaeebf25cd9fdec14338ccb16f6b2c4c7fa9163aefcf057d86b9cc248bb/nvidia_cusparse_cu12-12.1.0.106-py3-none-manylinux1_x86_64.whl", hash = "sha256:f3b50f42cf363f86ab21f720998517a659a48131e8d538dc02f8768237bd884c", size = 195958278, upload_time = "2023-04-19T15:51:49.939Z" }, @@ -6140,7 +5837,7 @@ resolution-markers = [ "python_full_version < '3.11' and sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra != 'group-12-asr-services-qwen3-asr-full' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice'", ] dependencies = [ - { name = "nvidia-nvjitlink-cu12", version = "12.6.85", source = { registry = "https://pypi.org/simple" } }, + { name = "nvidia-nvjitlink-cu12", version = "12.6.85", source = { registry = "https://pypi.org/simple" }, marker = "sys_platform == 'linux'" }, ] wheels = [ { url = "https://files.pythonhosted.org/packages/eb/eb/6681efd0aa7df96b4f8067b3ce7246833dd36830bb4cec8896182773db7d/nvidia_cusparse_cu12-12.5.4.2-py3-none-manylinux2014_aarch64.manylinux_2_17_aarch64.whl", hash = "sha256:d25b62fb18751758fe3c93a4a08eff08effedfe4edf1c6bb5afd0890fe88f887", size = 216451147, upload_time = "2024-11-20T17:44:18.055Z" }, @@ -6165,7 +5862,7 @@ resolution-markers = [ "python_full_version < '3.11' and platform_machine == 'aarch64' and sys_platform == 'linux'", ] dependencies = [ - { name = "nvidia-nvjitlink-cu12", version = "12.8.93", source = { registry = "https://pypi.org/simple" }, marker = "(extra == 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126') or (extra != 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128') or (extra != 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu128') or (extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-nemo') or (extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr') or (extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr') or (extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-transformers' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-nemo') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-transformers' and extra == 'group-12-asr-services-vibevoice')" }, + { name = "nvidia-nvjitlink-cu12", version = "12.8.93", source = { registry = "https://pypi.org/simple" }, marker = "(sys_platform == 'linux' and extra == 'extra-12-asr-services-cu128') or (sys_platform != 'linux' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-nemo') or (sys_platform != 'linux' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr') or (sys_platform != 'linux' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr-full') or (sys_platform != 'linux' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-transformers') or (sys_platform != 'linux' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-vibevoice') or (sys_platform != 'linux' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr') or (sys_platform != 'linux' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr-full') or (sys_platform != 'linux' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-transformers') or (sys_platform != 'linux' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-vibevoice') or (sys_platform != 'linux' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full') or (sys_platform != 'linux' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-transformers') or (sys_platform != 'linux' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-vibevoice') or (sys_platform != 'linux' and extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-transformers') or (sys_platform != 'linux' and extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-vibevoice') or (sys_platform != 'linux' and extra == 'group-12-asr-services-transformers' and extra == 'group-12-asr-services-vibevoice') or (extra == 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128') or (extra == 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126') or (extra == 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu128') or (extra == 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-nemo') or (extra == 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr') or (extra == 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra == 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-transformers') or (extra == 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-vibevoice') or (extra == 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr') or (extra == 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra == 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-transformers') or (extra == 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-vibevoice') or (extra == 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra == 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-transformers') or (extra == 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-vibevoice') or (extra == 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-transformers') or (extra == 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-vibevoice') or (extra == 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-transformers' and extra == 'group-12-asr-services-vibevoice') or (extra == 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-nemo') or (extra == 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr') or (extra == 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra == 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-transformers') or (extra == 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-vibevoice') or (extra == 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr') or (extra == 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra == 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-transformers') or (extra == 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-vibevoice') or (extra == 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra == 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-transformers') or (extra == 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-vibevoice') or (extra == 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-transformers') or (extra == 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-vibevoice') or (extra == 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-transformers' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-nemo') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-transformers' and extra == 'group-12-asr-services-vibevoice')" }, ] wheels = [ { url = "https://files.pythonhosted.org/packages/bc/f7/cd777c4109681367721b00a106f491e0d0d15cfa1fd59672ce580ce42a97/nvidia_cusparse_cu12-12.5.8.93-py3-none-manylinux2014_aarch64.manylinux_2_17_aarch64.whl", hash = "sha256:9b6c161cb130be1a07a27ea6923df8141f3c295852f4b260c65f18f3e0a091dc", size = 288117129, upload_time = "2025-03-07T01:47:40.407Z" }, @@ -6846,8 +6543,7 @@ name = "peft" version = "0.18.1" source = { registry = "https://pypi.org/simple" } dependencies = [ - { name = "accelerate", version = "1.7.0", source = { registry = "https://pypi.org/simple" }, marker = "(extra == 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126') or (extra == 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu128') or (extra == 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-nemo') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-nemo') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'group-12-asr-services-faster-whisper' and extra != 'group-12-asr-services-nemo' and extra != 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-vibevoice')" }, - { name = "accelerate", version = "1.12.0", source = { registry = "https://pypi.org/simple" }, marker = "(extra == 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126') or (extra == 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu128') or (extra == 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-nemo') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-qwen3-asr') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-transformers' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu128' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-faster-whisper') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-nemo') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-qwen3-asr') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra == 'group-12-asr-services-transformers' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'group-12-asr-services-transformers' and extra != 'group-12-asr-services-vibevoice')" }, + { name = "accelerate" }, { name = "huggingface-hub" }, { name = "numpy" }, { name = "packaging" }, @@ -7942,7 +7638,7 @@ name = "qwen-asr" version = "0.0.6" source = { registry = "https://pypi.org/simple" } dependencies = [ - { name = "accelerate", version = "1.12.0", source = { registry = "https://pypi.org/simple" } }, + { name = "accelerate" }, { name = "flask" }, { name = "gradio" }, { name = "librosa" }, @@ -10337,7 +10033,7 @@ resolution-markers = [ "python_full_version < '3.11' and sys_platform == 'linux'", ] dependencies = [ - { name = "filelock", marker = "extra == 'extra-12-asr-services-cu121' or (extra == 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128') or (extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-nemo') or (extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr') or (extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-transformers') or (extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-vibevoice') or (extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr') or (extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-transformers') or (extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-vibevoice') or (extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-transformers') or (extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-vibevoice') or (extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-transformers') or (extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-vibevoice') or (extra == 'group-12-asr-services-transformers' and extra == 'group-12-asr-services-vibevoice')" }, + { name = "filelock", marker = "(python_full_version < '3.13' and sys_platform == 'linux' and extra == 'extra-12-asr-services-cu121') or (python_full_version >= '3.13' and extra == 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126') or (python_full_version >= '3.13' and extra == 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu128') or (python_full_version >= '3.13' and extra == 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-nemo') or (python_full_version >= '3.13' and extra == 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr') or (python_full_version >= '3.13' and extra == 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr-full') or (python_full_version >= '3.13' and extra == 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-transformers') or (python_full_version >= '3.13' and extra == 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-vibevoice') or (python_full_version >= '3.13' and extra == 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr') or (python_full_version >= '3.13' and extra == 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr-full') or (python_full_version >= '3.13' and extra == 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-transformers') or (python_full_version >= '3.13' and extra == 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-vibevoice') or (python_full_version >= '3.13' and extra == 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full') or (python_full_version >= '3.13' and extra == 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-transformers') or (python_full_version >= '3.13' and extra == 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-vibevoice') or (python_full_version >= '3.13' and extra == 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-transformers') or (python_full_version >= '3.13' and extra == 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-vibevoice') or (python_full_version >= '3.13' and extra == 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-transformers' and extra == 'group-12-asr-services-vibevoice') or (sys_platform != 'linux' and extra == 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126') or (sys_platform != 'linux' and extra == 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu128') or (sys_platform != 'linux' and extra == 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128') or (sys_platform != 'linux' and extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-nemo') or (sys_platform != 'linux' and extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr') or (sys_platform != 'linux' and extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr-full') or (sys_platform != 'linux' and extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-transformers') or (sys_platform != 'linux' and extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-vibevoice') or (sys_platform != 'linux' and extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr') or (sys_platform != 'linux' and extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr-full') or (sys_platform != 'linux' and extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-transformers') or (sys_platform != 'linux' and extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-vibevoice') or (sys_platform != 'linux' and extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full') or (sys_platform != 'linux' and extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-transformers') or (sys_platform != 'linux' and extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-vibevoice') or (sys_platform != 'linux' and extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-transformers') or (sys_platform != 'linux' and extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-vibevoice') or (sys_platform != 'linux' and extra != 'extra-12-asr-services-cu128' and extra == 'group-12-asr-services-transformers' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-nemo') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-faster-whisper' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-nemo' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-qwen3-asr-full') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-qwen3-asr' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-transformers') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-qwen3-asr-full' and extra == 'group-12-asr-services-vibevoice') or (extra != 'extra-12-asr-services-cu121' and extra == 'group-12-asr-services-transformers' and extra == 'group-12-asr-services-vibevoice')" }, ] wheels = [ { url = "https://files.pythonhosted.org/packages/98/29/69aa56dc0b2eb2602b553881e34243475ea2afd9699be042316842788ff5/triton-3.1.0-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:6b0dd10a925263abbe9fa37dcde67a5e9b2383fc269fdf59f5657cac38c5d1d8", size = 209460013, upload_time = "2024-10-14T16:05:32.106Z" }, @@ -10356,7 +10052,7 @@ resolution-markers = [ "python_full_version < '3.11' and sys_platform == 'linux'", ] dependencies = [ - { name = "setuptools", marker = "(extra == 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126') or (extra == 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu128') or (extra == 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128') or (extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128')" }, + { name = "setuptools", marker = "(sys_platform == 'linux' and extra != 'extra-12-asr-services-cu121' and extra != 'extra-12-asr-services-cu126' and extra != 'extra-12-asr-services-cu128') or (extra == 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu126') or (extra == 'extra-12-asr-services-cu121' and extra == 'extra-12-asr-services-cu128') or (extra == 'extra-12-asr-services-cu126' and extra == 'extra-12-asr-services-cu128')" }, ] wheels = [ { url = "https://files.pythonhosted.org/packages/8d/a9/549e51e9b1b2c9b854fd761a1d23df0ba2fbc60bd0c13b489ffa518cfcb7/triton-3.3.1-cp310-cp310-manylinux_2_27_x86_64.manylinux_2_28_x86_64.whl", hash = "sha256:b74db445b1c562844d3cfad6e9679c72e93fdfb1a90a24052b03bb5c49d1242e", size = 155600257, upload_time = "2025-05-29T23:39:36.085Z" }, From 1398f78d622fac81b3502bef98bfecab086d121d Mon Sep 17 00:00:00 2001 From: Ankush Malaker <43288948+AnkushMalaker@users.noreply.github.com> Date: Sun, 22 Feb 2026 16:59:53 +0530 Subject: [PATCH 05/11] Refactor pre-commit configuration and enhance development scripts (#309) * Refactor pre-commit configuration and enhance development scripts - Removed local hooks for Robot Framework tests and cleanup from `.pre-commit-config.yaml`, streamlining the pre-commit setup. - Updated `Makefile` to install pre-commit using the `uv` tool, improving dependency management. - Enhanced `restart.sh`, `start.sh`, `status.sh`, and `stop.sh` scripts to source a new `check_uv.sh` script for better environment validation. - Added new environment variables for Galileo observability in `.env.template`, improving observability setup. - Introduced OpenTelemetry initialization in `app_factory.py` for enhanced observability during application runtime. * Update button event handling and plugin architecture - Changed button state terminology from `SINGLE_TAP` and `DOUBLE_TAP` to `SINGLE_PRESS` and `DOUBLE_PRESS` across various files, including documentation and code implementations. - Enhanced the `send_button_event` method to reflect the updated button state values, ensuring consistency in event handling. - Introduced new methods for managing button events in the plugin architecture, improving the overall interaction with device buttons. - Updated tests to align with the new button state definitions, ensuring robust coverage for the updated functionality. * Add unit tests for Qwen3-ASR output parsing and repetition detection - Introduced a new test file `test_qwen3_asr_parsing.py` to validate the functionality of the `_parse_qwen3_output` and `detect_and_fix_repetitions` methods. - Implemented various test cases covering standard and edge cases for ASR output parsing, including language detection, handling of empty inputs, and unexpected text. - Added tests for repetition detection to ensure proper functionality based on specified thresholds. - Enhanced the `Makefile` to include a new target for running specific tests by name, tag, or file, improving test execution flexibility. - Created a shared prerequisite check script `check_uv.sh` to ensure the `uv` package manager is installed before running scripts, enhancing setup reliability. * Add unit tests for Qwen3-ASR output parsing and repetition detection - Introduced a new test file `test_qwen3_asr_parsing.py` to validate the functionality of the `_parse_qwen3_output` and `detect_and_fix_repetitions` methods. - Implemented various test cases covering standard and edge cases for ASR output parsing, including language detection, handling of empty inputs, and unexpected text. - Added tests for repetition detection to ensure proper functionality based on specified thresholds. * Refactor Redis session handling and enhance error management - Updated session retrieval logic in `queue_routes.py` to ensure proper closure of Redis connections using `await redis_client.aclose()`, improving resource management. - Enhanced error handling during session data retrieval, providing clearer logging for issues encountered while fetching session information. - Streamlined the session key scanning process, maintaining existing functionality while improving code readability and maintainability. - Added optional parameters to the `transcribe` method in `mock_provider.py` for better flexibility in handling context information and progress callbacks during transcription tasks. --- .pre-commit-config.yaml | 30 +- Makefile | 3 +- backends/advanced/.env.template | 6 + .../advanced/Docs/plugin-development-guide.md | 4 +- .../src/advanced_omi_backend/app_factory.py | 128 ++- .../clients/audio_stream_client.py | 66 +- .../src/advanced_omi_backend/config_loader.py | 40 +- .../controllers/conversation_controller.py | 555 ++++++---- .../controllers/websocket_controller.py | 987 +++++++++--------- .../observability/__init__.py | 0 .../observability/otel_setup.py | 129 +++ .../src/advanced_omi_backend/plugins/base.py | 27 +- .../advanced_omi_backend/plugins/events.py | 38 +- .../advanced_omi_backend/plugins/router.py | 203 +++- .../advanced_omi_backend/plugins/services.py | 43 +- .../advanced_omi_backend/prompt_defaults.py | 2 +- .../advanced_omi_backend/prompt_registry.py | 25 +- .../routers/modules/queue_routes.py | 593 ++++++----- .../services/audio_stream/producer.py | 140 ++- .../services/transcription/__init__.py | 236 ++++- .../services/transcription/mock_provider.py | 28 +- .../workers/conversation_jobs.py | 190 +++- .../workers/memory_jobs.py | 84 +- .../workers/rq_worker_entry.py | 24 +- .../workers/transcription_jobs.py | 189 +++- config/config.yml.template | 8 + config/defaults.yml | 4 + config/plugins.yml.template | 7 +- .../tests/test_qwen3_asr_parsing.py | 125 +++ extras/friend-lite-sdk/friend_lite/button.py | 4 +- plugins/homeassistant/command_parser.py | 22 +- plugins/homeassistant/plugin.py | 254 ++++- restart.sh | 1 + scripts/check_uv.sh | 15 + start.sh | 2 + status.sh | 1 + stop.sh | 2 + tests/Makefile | 15 +- tests/config/plugins.test.yml | 11 + .../audio_streaming_integration_tests.robot | 20 +- .../websocket_streaming_tests.robot | 41 +- .../websocket_transcription_e2e_test.robot | 23 +- tests/libs/audio_stream_library.py | 18 +- tests/resources/websocket_keywords.robot | 6 + wizard.py | 782 +++++++++----- wizard.sh | 2 + 46 files changed, 3442 insertions(+), 1691 deletions(-) create mode 100644 backends/advanced/src/advanced_omi_backend/observability/__init__.py create mode 100644 backends/advanced/src/advanced_omi_backend/observability/otel_setup.py create mode 100644 extras/asr-services/tests/test_qwen3_asr_parsing.py create mode 100755 scripts/check_uv.sh diff --git a/.pre-commit-config.yaml b/.pre-commit-config.yaml index 6ebb6573..859eca3c 100644 --- a/.pre-commit-config.yaml +++ b/.pre-commit-config.yaml @@ -1,42 +1,22 @@ repos: - # Local hooks (project-specific checks) - - repo: local - hooks: - # Run Robot Framework endpoint tests before push - - id: robot-framework-tests - name: Robot Framework Tests (Endpoints) - entry: bash -c 'cd tests && make endpoints OUTPUTDIR=.pre-commit-results' - language: system - pass_filenames: false - stages: [push] - verbose: true - - # Clean up test results after hook runs - - id: cleanup-test-results - name: Cleanup Test Results - entry: bash -c 'cd tests && rm -rf .pre-commit-results' - language: system - pass_filenames: false - stages: [push] - always_run: true - # Code formatting - repo: https://github.com/psf/black rev: 24.4.2 hooks: - id: black - files: ^backends/advanced-backend/src/.*\.py$ + exclude: \.venv/ - repo: https://github.com/PyCQA/isort rev: 5.13.2 hooks: - id: isort - files: ^backends/advanced-backend/src/.*\.py$ + args: ["--profile", "black"] + exclude: \.venv/ # File hygiene - repo: https://github.com/pre-commit/pre-commit-hooks rev: v4.5.0 hooks: - id: trailing-whitespace - files: ^backends/advanced-backend/src/.* + exclude: \.venv/ - id: end-of-file-fixer - files: ^backends/advanced-backend/src/.* \ No newline at end of file + exclude: \.venv/ diff --git a/Makefile b/Makefile index d821819e..ae9e784d 100644 --- a/Makefile +++ b/Makefile @@ -122,8 +122,9 @@ help: ## Show detailed help for all targets setup-dev: ## Setup development environment (git hooks, pre-commit) @echo "šŸ› ļø Setting up development environment..." @echo "" + @bash scripts/check_uv.sh @echo "šŸ“¦ Installing pre-commit..." - @pip install pre-commit 2>/dev/null || pip3 install pre-commit + @uv tool install pre-commit @echo "" @echo "šŸ”§ Installing git hooks..." @pre-commit install --hook-type pre-push diff --git a/backends/advanced/.env.template b/backends/advanced/.env.template index 6de583fd..293b299f 100644 --- a/backends/advanced/.env.template +++ b/backends/advanced/.env.template @@ -62,6 +62,12 @@ LANGFUSE_PUBLIC_KEY= LANGFUSE_SECRET_KEY= LANGFUSE_BASE_URL=http://langfuse-web:3000 +# Galileo (OTEL-based LLM observability) +GALILEO_API_KEY= +GALILEO_PROJECT=chronicle +GALILEO_LOG_STREAM=default +# GALILEO_CONSOLE_URL=https://app.galileo.ai # Default; override for self-hosted + # Qwen3-ASR (offline ASR via vLLM) # QWEN3_ASR_URL=host.docker.internal:8767 # QWEN3_ASR_STREAM_URL=host.docker.internal:8769 diff --git a/backends/advanced/Docs/plugin-development-guide.md b/backends/advanced/Docs/plugin-development-guide.md index d5ddf3fa..32fa700d 100644 --- a/backends/advanced/Docs/plugin-development-guide.md +++ b/backends/advanced/Docs/plugin-development-guide.md @@ -206,7 +206,7 @@ async def on_memory_processed(self, context: PluginContext): **When**: OMI device button is pressed **Context Data**: -- `state` (str): Button state (`SINGLE_TAP`, `DOUBLE_TAP`) +- `state` (str): Button state (`SINGLE_PRESS`, `DOUBLE_PRESS`) - `timestamp` (float): Unix timestamp of the event - `audio_uuid` (str): Current audio session UUID (may be None) - `session_id` (str): Streaming session ID (for conversation close) @@ -222,7 +222,7 @@ friend-lite-sdk (extras/friend-lite-sdk/) ā†’ parse_button_event() converts payload ā†’ ButtonState IntEnum ā†“ BLE Client (extras/local-wearable-client/ or mobile app) - ā†’ Formats as Wyoming protocol: {"type": "button-event", "data": {"state": "SINGLE_TAP"}} + ā†’ Formats as Wyoming protocol: {"type": "button-event", "data": {"state": "SINGLE_PRESS"}} ā†’ Sends over WebSocket ā†“ Backend (websocket_controller.py) diff --git a/backends/advanced/src/advanced_omi_backend/app_factory.py b/backends/advanced/src/advanced_omi_backend/app_factory.py index 74cddd49..6083de97 100644 --- a/backends/advanced/src/advanced_omi_backend/app_factory.py +++ b/backends/advanced/src/advanced_omi_backend/app_factory.py @@ -71,26 +71,48 @@ async def initialize_openmemory_user() -> None: # Get configured user_id and server_url openmemory_config = memory_provider_config.openmemory_config - user_id = openmemory_config.get("user_id", "openmemory") if openmemory_config else "openmemory" - server_url = openmemory_config.get("server_url", "http://host.docker.internal:8765") if openmemory_config else "http://host.docker.internal:8765" - client_name = openmemory_config.get("client_name", "chronicle") if openmemory_config else "chronicle" + user_id = ( + openmemory_config.get("user_id", "openmemory") + if openmemory_config + else "openmemory" + ) + server_url = ( + openmemory_config.get("server_url", "http://host.docker.internal:8765") + if openmemory_config + else "http://host.docker.internal:8765" + ) + client_name = ( + openmemory_config.get("client_name", "chronicle") + if openmemory_config + else "chronicle" + ) - application_logger.info(f"Registering OpenMemory user: {user_id} at {server_url}") + application_logger.info( + f"Registering OpenMemory user: {user_id} at {server_url}" + ) # Make a lightweight registration call (create and delete dummy memory) - async with MCPClient(server_url=server_url, client_name=client_name, user_id=user_id) as client: + async with MCPClient( + server_url=server_url, client_name=client_name, user_id=user_id + ) as client: # Test connection first is_connected = await client.test_connection() if is_connected: # Create and immediately delete a dummy memory to trigger user creation - memory_ids = await client.add_memories("Chronicle initialization - user registration test") + memory_ids = await client.add_memories( + "Chronicle initialization - user registration test" + ) if memory_ids: # Delete the test memory await client.delete_memory(memory_ids[0]) application_logger.info(f"āœ… Registered OpenMemory user: {user_id}") else: - application_logger.warning(f"āš ļø OpenMemory MCP not reachable at {server_url}") - application_logger.info("User will be auto-created on first memory operation") + application_logger.warning( + f"āš ļø OpenMemory MCP not reachable at {server_url}" + ) + application_logger.info( + "User will be auto-created on first memory operation" + ) except Exception as e: application_logger.warning(f"āš ļø Could not register OpenMemory user: {e}") application_logger.info("User will be auto-created on first memory operation") @@ -116,7 +138,13 @@ async def lifespan(app: FastAPI): await init_beanie( database=config.db, - document_models=[User, Conversation, AudioChunkDocument, WaveformData, Annotation], + document_models=[ + User, + Conversation, + AudioChunkDocument, + WaveformData, + Annotation, + ], ) application_logger.info("Beanie initialized for all document models") except Exception as e: @@ -133,12 +161,17 @@ async def lifespan(app: FastAPI): # Initialize Redis connection for RQ try: from advanced_omi_backend.controllers.queue_controller import redis_conn + redis_conn.ping() application_logger.info("Redis connection established for RQ") - application_logger.info("RQ workers can be started with: rq worker transcription memory default") + application_logger.info( + "RQ workers can be started with: rq worker transcription memory default" + ) except Exception as e: application_logger.error(f"Failed to connect to Redis for RQ: {e}") - application_logger.warning("RQ queue system will not be available - check Redis connection") + application_logger.warning( + "RQ queue system will not be available - check Redis connection" + ) # Initialize BackgroundTaskManager (must happen before any code path uses it) try: @@ -153,6 +186,14 @@ async def lifespan(app: FastAPI): get_client_manager() application_logger.info("ClientManager initialized") + # Initialize OTEL/Galileo if configured (before LLM client so instrumentor patches OpenAI first) + try: + from advanced_omi_backend.observability.otel_setup import init_otel + + init_otel() + except Exception as e: + application_logger.warning(f"OTEL initialization skipped: {e}") + # Initialize prompt registry with defaults; seed into LangFuse in background try: from advanced_omi_backend.prompt_defaults import register_all_defaults @@ -176,6 +217,7 @@ async def _deferred_seed(): # Initialize LLM client eagerly (catch config errors at startup, not on first request) try: from advanced_omi_backend.llm_client import get_llm_client + get_llm_client() application_logger.info("LLM client initialized from config.yml") except Exception as e: @@ -186,35 +228,47 @@ async def _deferred_seed(): audio_service = get_audio_stream_service() await audio_service.connect() application_logger.info("Audio stream service connected to Redis Streams") - application_logger.info("Audio stream workers can be started with: python -m advanced_omi_backend.workers.audio_stream_worker") + application_logger.info( + "Audio stream workers can be started with: python -m advanced_omi_backend.workers.audio_stream_worker" + ) except Exception as e: application_logger.error(f"Failed to connect audio stream service: {e}") - application_logger.warning("Redis Streams audio processing will not be available") + application_logger.warning( + "Redis Streams audio processing will not be available" + ) # Initialize Redis client for audio streaming producer (used by WebSocket handlers) try: app.state.redis_audio_stream = await redis.from_url( - config.redis_url, - encoding="utf-8", - decode_responses=False + config.redis_url, encoding="utf-8", decode_responses=False ) from advanced_omi_backend.services.audio_stream import AudioStreamProducer - app.state.audio_stream_producer = AudioStreamProducer(app.state.redis_audio_stream) - application_logger.info("āœ… Redis client for audio streaming producer initialized") + + app.state.audio_stream_producer = AudioStreamProducer( + app.state.redis_audio_stream + ) + application_logger.info( + "āœ… Redis client for audio streaming producer initialized" + ) # Initialize ClientManager Redis for cross-container clientā†’user mapping from advanced_omi_backend.client_manager import ( initialize_redis_for_client_manager, ) + initialize_redis_for_client_manager(config.redis_url) except Exception as e: - application_logger.error(f"Failed to initialize Redis client for audio streaming: {e}", exc_info=True) + application_logger.error( + f"Failed to initialize Redis client for audio streaming: {e}", exc_info=True + ) application_logger.warning("Audio streaming producer will not be available") # Skip memory service pre-initialization to avoid blocking FastAPI startup # Memory service will be lazily initialized when first used - application_logger.info("Memory service will be initialized on first use (lazy loading)") + application_logger.info( + "Memory service will be initialized on first use (lazy loading)" + ) # Register OpenMemory user if using openmemory_mcp provider await initialize_openmemory_user() @@ -264,12 +318,15 @@ async def _deferred_seed(): application_logger.info(f"āœ… Plugin '{plugin_id}' initialized") except Exception as e: plugin_router.mark_plugin_failed(plugin_id, str(e)) - application_logger.error(f"Failed to initialize plugin '{plugin_id}': {e}", exc_info=True) + application_logger.error( + f"Failed to initialize plugin '{plugin_id}': {e}", + exc_info=True, + ) health = plugin_router.get_health_summary() application_logger.info( f"Plugins initialized: {health['initialized']}/{health['total']} active" - + (f", {health['failed']} failed" if health['failed'] else "") + + (f", {health['failed']} failed" if health["failed"] else "") ) # Store in app state for API access @@ -281,10 +338,14 @@ async def _deferred_seed(): app.state.plugin_router = None except Exception as e: - application_logger.error(f"Failed to initialize plugin system: {e}", exc_info=True) + application_logger.error( + f"Failed to initialize plugin system: {e}", exc_info=True + ) app.state.plugin_router = None - application_logger.info("Application ready - using application-level processing architecture.") + application_logger.info( + "Application ready - using application-level processing architecture." + ) logger.info("App ready") try: @@ -300,6 +361,7 @@ async def _deferred_seed(): from advanced_omi_backend.controllers.websocket_controller import ( cleanup_client_state, ) + await cleanup_client_state(client_id) except Exception as e: application_logger.error(f"Error cleaning up client {client_id}: {e}") @@ -327,9 +389,14 @@ async def _deferred_seed(): # Close Redis client for audio streaming producer try: - if hasattr(app.state, 'redis_audio_stream') and app.state.redis_audio_stream: + if ( + hasattr(app.state, "redis_audio_stream") + and app.state.redis_audio_stream + ): await app.state.redis_audio_stream.close() - application_logger.info("Redis client for audio streaming producer closed") + application_logger.info( + "Redis client for audio streaming producer closed" + ) except Exception as e: application_logger.error(f"Error closing Redis audio streaming client: {e}") @@ -341,6 +408,7 @@ async def _deferred_seed(): from advanced_omi_backend.services.plugin_service import ( cleanup_plugin_router, ) + await cleanup_plugin_router() application_logger.info("Plugins shut down") except Exception as e: @@ -380,7 +448,7 @@ def create_app() -> FastAPI: # Add WebSocket router at root level (not under /api prefix) app.include_router(websocket_router) - # Add authentication routers + # Add authentication routers app.include_router( fastapi_users.get_auth_router(cookie_backend), prefix="/auth/cookie", @@ -403,6 +471,8 @@ def create_app() -> FastAPI: CHUNK_DIR = Path("/app/audio_chunks") app.mount("/audio", StaticFiles(directory=CHUNK_DIR), name="audio") - logger.info("FastAPI application created with all routers and middleware configured") + logger.info( + "FastAPI application created with all routers and middleware configured" + ) - return app \ No newline at end of file + return app diff --git a/backends/advanced/src/advanced_omi_backend/clients/audio_stream_client.py b/backends/advanced/src/advanced_omi_backend/clients/audio_stream_client.py index 0595f3a4..b374d8fe 100644 --- a/backends/advanced/src/advanced_omi_backend/clients/audio_stream_client.py +++ b/backends/advanced/src/advanced_omi_backend/clients/audio_stream_client.py @@ -114,7 +114,11 @@ async def connect(self, wait_for_ready: bool = True) -> WebSocketClientProtocol: if wait_for_ready and "codec=pcm" in self.endpoint: # PCM codec sends "ready" message after auth (line 261-268 in websocket_controller.py) ready_msg = await self.ws.recv() - ready = json.loads(ready_msg.strip() if isinstance(ready_msg, str) else ready_msg.decode().strip()) + ready = json.loads( + ready_msg.strip() + if isinstance(ready_msg, str) + else ready_msg.decode().strip() + ) if ready.get("type") != "ready": raise RuntimeError(f"Expected 'ready' message, got: {ready}") logger.info("Received ready message from server") @@ -154,10 +158,8 @@ async def send_audio_start( }, "payload_length": None, } - print(f"šŸ”µ CLIENT: Sending audio-start message: {header}") logger.info(f"šŸ”µ CLIENT: Sending audio-start message: {header}") await self.ws.send(json.dumps(header) + "\n") - print(f"āœ… CLIENT: Sent audio-start with mode={recording_mode}") logger.info(f"āœ… CLIENT: Sent audio-start with mode={recording_mode}") async def send_audio_chunk_wyoming( @@ -197,7 +199,9 @@ async def send_audio_chunk_wyoming( self.total_bytes += len(audio_data) if self.chunk_count <= 3 or self.chunk_count % 100 == 0: - logger.debug(f"Sent audio chunk #{self.chunk_count}: {len(audio_data)} bytes") + logger.debug( + f"Sent audio chunk #{self.chunk_count}: {len(audio_data)} bytes" + ) async def send_audio_chunk_raw(self, audio_data: bytes) -> None: """Send raw binary audio without Wyoming header (legacy mode). @@ -222,7 +226,22 @@ async def send_audio_stop(self) -> None: header = {"type": "audio-stop"} await self.ws.send(json.dumps(header) + "\n") - logger.info(f"Sent audio-stop (total: {self.chunk_count} chunks, {self.total_bytes} bytes)") + logger.info( + f"Sent audio-stop (total: {self.chunk_count} chunks, {self.total_bytes} bytes)" + ) + + async def send_button_event(self, button_state: str = "SINGLE_PRESS") -> None: + """Send button event via Wyoming protocol. + + Args: + button_state: Button state ("SINGLE_PRESS" or "DOUBLE_PRESS") + """ + if not self.ws: + raise RuntimeError("Not connected. Call connect() first.") + + header = {"type": "button-event", "data": {"state": button_state}} + await self.ws.send(json.dumps(header) + "\n") + logger.info(f"Sent button event: {button_state}") async def send_ping(self) -> None: """Send keepalive ping.""" @@ -305,7 +324,9 @@ async def stream_wav_file( # Send audio-stop await self.send_audio_stop() - logger.info(f"Finished streaming: {self.chunk_count} chunks, {self.total_bytes} bytes") + logger.info( + f"Finished streaming: {self.chunk_count} chunks, {self.total_bytes} bytes" + ) return self.chunk_count async def close(self) -> None: @@ -318,7 +339,7 @@ async def close(self) -> None: except asyncio.TimeoutError: logger.warning("WebSocket close timed out after 2s, forcing close") # Force close without waiting for handshake - if hasattr(self.ws, 'transport') and self.ws.transport: + if hasattr(self.ws, "transport") and self.ws.transport: self.ws.transport.close() except Exception as e: logger.error(f"Error during WebSocket close: {e}") @@ -448,10 +469,14 @@ def run_loop(): # Connect and send audio-start async def _connect_and_start(): try: - logger.info(f"šŸ”µ CLIENT: Stream {stream_id} connecting for {device_name}...") + logger.info( + f"šŸ”µ CLIENT: Stream {stream_id} connecting for {device_name}..." + ) await client.connect() session.connected = True - logger.info(f"āœ… CLIENT: Stream {stream_id} connected, sending audio-start...") + logger.info( + f"āœ… CLIENT: Stream {stream_id} connected, sending audio-start..." + ) await client.send_audio_start(recording_mode=recording_mode) session.audio_started = True logger.info(f"āœ… CLIENT: Stream {stream_id} started for {device_name}") @@ -595,9 +620,30 @@ async def _close_abruptly(): total_chunks = session.chunk_count del self._sessions[stream_id] - logger.info(f"Stream {stream_id} closed abruptly (no audio-stop), sent {total_chunks} chunks") + logger.info( + f"Stream {stream_id} closed abruptly (no audio-stop), sent {total_chunks} chunks" + ) return total_chunks + def send_button_event( + self, stream_id: str, button_state: str = "SINGLE_PRESS" + ) -> None: + """Send a button event to an open stream. + + Args: + stream_id: Stream session ID + button_state: Button state ("SINGLE_PRESS" or "DOUBLE_PRESS") + """ + session = self._sessions.get(stream_id) + if not session: + raise ValueError(f"Unknown stream_id: {stream_id}") + + async def _send(): + await session.client.send_button_event(button_state) + + future = asyncio.run_coroutine_threadsafe(_send(), session.loop) + future.result(timeout=5) + def get_session(self, stream_id: str) -> Optional[StreamSession]: """Get session info for a stream.""" return self._sessions.get(stream_id) diff --git a/backends/advanced/src/advanced_omi_backend/config_loader.py b/backends/advanced/src/advanced_omi_backend/config_loader.py index 1b8be9ee..bd1f9de4 100644 --- a/backends/advanced/src/advanced_omi_backend/config_loader.py +++ b/backends/advanced/src/advanced_omi_backend/config_loader.py @@ -26,7 +26,7 @@ def get_config_dir() -> Path: def get_plugins_yml_path() -> Path: """ Get path to plugins.yml file (single source of truth). - + Returns: Path to plugins.yml """ @@ -86,20 +86,34 @@ def load_config(force_reload: bool = False) -> DictConfig: # OmegaConf.merge replaces lists entirely, so we need custom merge # for the 'models' list: merge by name so defaults models that aren't # in user config are still available. - default_models = OmegaConf.to_container(defaults.get("models", []) or [], resolve=False) if defaults else [] - user_models = OmegaConf.to_container(user_config.get("models", []) or [], resolve=False) if user_config else [] + default_models = ( + OmegaConf.to_container(defaults.get("models", []) or [], resolve=False) + if defaults + else [] + ) + user_models = ( + OmegaConf.to_container(user_config.get("models", []) or [], resolve=False) + if user_config + else [] + ) merged = OmegaConf.merge(defaults, user_config) # Name-based merge: user models override defaults, but default-only models are kept if default_models and user_models: user_model_names = {m.get("name") for m in user_models if isinstance(m, dict)} - extra_defaults = [m for m in default_models if isinstance(m, dict) and m.get("name") not in user_model_names] + extra_defaults = [ + m + for m in default_models + if isinstance(m, dict) and m.get("name") not in user_model_names + ] if extra_defaults: all_models = user_models + extra_defaults merged["models"] = OmegaConf.create(all_models) - logger.info(f"Merged {len(extra_defaults)} default-only models into config: " - f"{[m.get('name') for m in extra_defaults]}") + logger.info( + f"Merged {len(extra_defaults)} default-only models into config: " + f"{[m.get('name') for m in extra_defaults]}" + ) # Cache result _config_cache = merged @@ -126,7 +140,7 @@ def get_backend_config(section: Optional[str] = None) -> DictConfig: DictConfig for backend section or subsection """ cfg = load_config() - if 'backend' not in cfg: + if "backend" not in cfg: return OmegaConf.create({}) backend_cfg = cfg.backend @@ -153,6 +167,9 @@ def save_config_section(section_path: str, values: dict) -> bool: """ Update a config section and save to config.yml. + Also updates the in-memory config cache so changes take effect immediately, + even when CONFIG_FILE points to a different file (e.g., in test environments). + Args: section_path: Dot-separated path (e.g., 'backend.diarization') values: Dict with new values @@ -174,8 +191,13 @@ def save_config_section(section_path: str, values: dict) -> bool: # Save back to file OmegaConf.save(existing_config, config_path) - # Invalidate cache - reload_config() + # Reload config from the primary config file (CONFIG_FILE env var) + merged = reload_config() + + # Also apply the values to the in-memory cache directly. + # This is needed when CONFIG_FILE points to a different file than config.yml + # (e.g., test configs), so the saved values still take effect at runtime. + OmegaConf.update(merged, section_path, values, merge=True) logger.info(f"Saved config section '{section_path}' to {config_path}") return True diff --git a/backends/advanced/src/advanced_omi_backend/controllers/conversation_controller.py b/backends/advanced/src/advanced_omi_backend/controllers/conversation_controller.py index 0fcb9d9c..1bf41dfc 100644 --- a/backends/advanced/src/advanced_omi_backend/controllers/conversation_controller.py +++ b/backends/advanced/src/advanced_omi_backend/controllers/conversation_controller.py @@ -17,6 +17,7 @@ client_belongs_to_user, get_client_manager, ) +from advanced_omi_backend.config import get_transcription_job_timeout from advanced_omi_backend.config_loader import get_service_config from advanced_omi_backend.controllers.queue_controller import ( JOB_RESULT_TTL, @@ -32,15 +33,14 @@ from advanced_omi_backend.models.conversation import Conversation from advanced_omi_backend.models.job import JobPriority from advanced_omi_backend.plugins.events import ConversationCloseReason, PluginEvent +from advanced_omi_backend.services.memory import get_memory_service from advanced_omi_backend.users import User from advanced_omi_backend.workers.conversation_jobs import generate_title_summary_job -from advanced_omi_backend.services.memory import get_memory_service from advanced_omi_backend.workers.memory_jobs import ( enqueue_memory_processing, process_memory_job, ) from advanced_omi_backend.workers.speaker_jobs import recognise_speakers_job -from advanced_omi_backend.config import get_transcription_job_timeout logger = logging.getLogger(__name__) audio_logger = logging.getLogger("audio_processing") @@ -73,7 +73,7 @@ async def close_current_conversation(client_id: str, user: User): status_code=404, ) - session_id = getattr(client_state, 'stream_session_id', None) + session_id = getattr(client_state, "stream_session_id", None) if not session_id: return JSONResponse( content={"error": "No active session"}, @@ -96,7 +96,9 @@ async def close_current_conversation(client_id: str, user: User): status_code=404, ) - logger.info(f"Conversation close requested for client {client_id} by user {user.user_id}") + logger.info( + f"Conversation close requested for client {client_id} by user {user.user_id}" + ) return JSONResponse( content={ @@ -111,9 +113,13 @@ async def get_conversation(conversation_id: str, user: User): """Get a single conversation with full transcript details.""" try: # Find the conversation using Beanie - conversation = await Conversation.find_one(Conversation.conversation_id == conversation_id) + conversation = await Conversation.find_one( + Conversation.conversation_id == conversation_id + ) if not conversation: - return JSONResponse(status_code=404, content={"error": "Conversation not found"}) + return JSONResponse( + status_code=404, content={"error": "Conversation not found"} + ) # Check ownership for non-admin users if not user.is_superuser and conversation.user_id != str(user.user_id): @@ -127,15 +133,23 @@ async def get_conversation(conversation_id: str, user: User): "audio_chunks_count": conversation.audio_chunks_count, "audio_total_duration": conversation.audio_total_duration, "audio_compression_ratio": conversation.audio_compression_ratio, - "created_at": conversation.created_at.isoformat() if conversation.created_at else None, + "created_at": ( + conversation.created_at.isoformat() if conversation.created_at else None + ), "deleted": conversation.deleted, "deletion_reason": conversation.deletion_reason, - "deleted_at": conversation.deleted_at.isoformat() if conversation.deleted_at else None, + "deleted_at": ( + conversation.deleted_at.isoformat() if conversation.deleted_at else None + ), "processing_status": conversation.processing_status, "always_persist": conversation.always_persist, - "end_reason": conversation.end_reason.value if conversation.end_reason else None, + "end_reason": ( + conversation.end_reason.value if conversation.end_reason else None + ), "completed_at": ( - conversation.completed_at.isoformat() if conversation.completed_at else None + conversation.completed_at.isoformat() + if conversation.completed_at + else None ), "title": conversation.title, "summary": conversation.summary, @@ -153,14 +167,18 @@ async def get_conversation(conversation_id: str, user: User): "active_transcript_version_number": conversation.active_transcript_version_number, "active_memory_version_number": conversation.active_memory_version_number, "starred": conversation.starred, - "starred_at": conversation.starred_at.isoformat() if conversation.starred_at else None, + "starred_at": ( + conversation.starred_at.isoformat() if conversation.starred_at else None + ), } return {"conversation": response} except Exception as e: logger.error(f"Error fetching conversation {conversation_id}: {e}") - return JSONResponse(status_code=500, content={"error": "Error fetching conversation"}) + return JSONResponse( + status_code=500, content={"error": "Error fetching conversation"} + ) async def get_conversation_memories(conversation_id: str, user: User, limit: int = 100): @@ -170,7 +188,9 @@ async def get_conversation_memories(conversation_id: str, user: User, limit: int Conversation.conversation_id == conversation_id ) if not conversation: - return JSONResponse(status_code=404, content={"error": "Conversation not found"}) + return JSONResponse( + status_code=404, content={"error": "Conversation not found"} + ) if not user.is_superuser and conversation.user_id != str(user.user_id): return JSONResponse(status_code=403, content={"error": "Access forbidden"}) @@ -364,21 +384,29 @@ async def get_conversations( if include_unprocessed: # Orphan type 1: always_persist stuck in pending/failed (not deleted) - conditions.append({ - "always_persist": True, - "processing_status": {"$in": ["pending_transcription", "transcription_failed"]}, - "deleted": False, - }) + conditions.append( + { + "always_persist": True, + "processing_status": { + "$in": ["pending_transcription", "transcription_failed"] + }, + "deleted": False, + } + ) # Orphan type 2: soft-deleted due to no speech but have audio data - conditions.append({ - "deleted": True, - "deletion_reason": {"$in": [ - "no_meaningful_speech", - "audio_file_not_ready", - "no_meaningful_speech_batch_transcription", - ]}, - "audio_chunks_count": {"$gt": 0}, - }) + conditions.append( + { + "deleted": True, + "deletion_reason": { + "$in": [ + "no_meaningful_speech", + "audio_file_not_ready", + "no_meaningful_speech_batch_transcription", + ] + }, + "audio_chunks_count": {"$gt": 0}, + } + ) # Assemble final query if len(conditions) == 1: @@ -406,12 +434,14 @@ async def get_conversations( conv_id = doc.get("conversation_id") is_orphan_type1 = ( doc.get("always_persist") - and doc.get("processing_status") in ("pending_transcription", "transcription_failed") + and doc.get("processing_status") + in ("pending_transcription", "transcription_failed") and not doc.get("deleted") ) is_orphan_type2 = ( doc.get("deleted") - and doc.get("deletion_reason") in ( + and doc.get("deletion_reason") + in ( "no_meaningful_speech", "audio_file_not_ready", "no_meaningful_speech_batch_transcription", @@ -437,7 +467,9 @@ async def get_conversations( except Exception as e: logger.exception(f"Error fetching conversations: {e}") - return JSONResponse(status_code=500, content={"error": "Error fetching conversations"}) + return JSONResponse( + status_code=500, content={"error": "Error fetching conversations"} + ) async def search_conversations( @@ -513,10 +545,14 @@ async def search_conversations( except Exception as e: logger.exception(f"Error searching conversations: {e}") - return JSONResponse(status_code=500, content={"error": "Error searching conversations"}) + return JSONResponse( + status_code=500, content={"error": "Error searching conversations"} + ) -async def _soft_delete_conversation(conversation: Conversation, user: User) -> JSONResponse: +async def _soft_delete_conversation( + conversation: Conversation, user: User +) -> JSONResponse: """Mark conversation and chunks as deleted (soft delete). Chunks are soft-deleted first so that a crash between the two writes @@ -533,7 +569,9 @@ async def _soft_delete_conversation(conversation: Conversation, user: User) -> J ).update_many({"$set": {"deleted": True, "deleted_at": deleted_at}}) deleted_chunks = result.modified_count - logger.info(f"Soft deleted {deleted_chunks} audio chunks for conversation {conversation_id}") + logger.info( + f"Soft deleted {deleted_chunks} audio chunks for conversation {conversation_id}" + ) # 2. Mark conversation as deleted conversation.deleted = True @@ -561,7 +599,9 @@ async def _soft_delete_conversation(conversation: Conversation, user: User) -> J "deleted_chunks": deleted_chunks, "conversation_id": conversation_id, "client_id": conversation.client_id, - "deleted_at": conversation.deleted_at.isoformat() if conversation.deleted_at else None, + "deleted_at": ( + conversation.deleted_at.isoformat() if conversation.deleted_at else None + ), }, ) @@ -582,7 +622,9 @@ async def _hard_delete_conversation(conversation: Conversation) -> JSONResponse: ).delete() deleted_chunks = result.deleted_count - logger.info(f"Hard deleted {deleted_chunks} audio chunks for conversation {conversation_id}") + logger.info( + f"Hard deleted {deleted_chunks} audio chunks for conversation {conversation_id}" + ) # 2. Delete conversation document try: @@ -607,7 +649,9 @@ async def _hard_delete_conversation(conversation: Conversation) -> JSONResponse: ) -async def delete_conversation(conversation_id: str, user: User, permanent: bool = False): +async def delete_conversation( + conversation_id: str, user: User, permanent: bool = False +): """ Soft delete a conversation (mark as deleted but keep data). @@ -628,11 +672,14 @@ async def delete_conversation(conversation_id: str, user: User, permanent: bool ) # Find the conversation using Beanie - conversation = await Conversation.find_one(Conversation.conversation_id == conversation_id) + conversation = await Conversation.find_one( + Conversation.conversation_id == conversation_id + ) if not conversation: return JSONResponse( - status_code=404, content={"error": f"Conversation '{conversation_id}' not found"} + status_code=404, + content={"error": f"Conversation '{conversation_id}' not found"}, ) # Check ownership for non-admin users @@ -658,7 +705,8 @@ async def delete_conversation(conversation_id: str, user: User, permanent: bool except Exception as e: logger.error(f"Error deleting conversation {conversation_id}: {e}") return JSONResponse( - status_code=500, content={"error": f"Failed to delete conversation: {str(e)}"} + status_code=500, + content={"error": f"Failed to delete conversation: {str(e)}"}, ) @@ -671,17 +719,23 @@ async def restore_conversation(conversation_id: str, user: User) -> JSONResponse user: Requesting user """ try: - conversation = await Conversation.find_one(Conversation.conversation_id == conversation_id) + conversation = await Conversation.find_one( + Conversation.conversation_id == conversation_id + ) if not conversation: - return JSONResponse(status_code=404, content={"error": "Conversation not found"}) + return JSONResponse( + status_code=404, content={"error": "Conversation not found"} + ) # Permission check if not user.is_superuser and conversation.user_id != str(user.user_id): return JSONResponse(status_code=403, content={"error": "Access denied"}) if not conversation.deleted: - return JSONResponse(status_code=400, content={"error": "Conversation is not deleted"}) + return JSONResponse( + status_code=400, content={"error": "Conversation is not deleted"} + ) # 1. Restore audio chunks FIRST (safe failure mode: restored chunks, conversation still deleted) original_deleted_at = conversation.deleted_at @@ -707,7 +761,9 @@ async def restore_conversation(conversation_id: str, user: User) -> JSONResponse await AudioChunkDocument.find( AudioChunkDocument.conversation_id == conversation_id, AudioChunkDocument.deleted == False, - ).update_many({"$set": {"deleted": True, "deleted_at": original_deleted_at}}) + ).update_many( + {"$set": {"deleted": True, "deleted_at": original_deleted_at}} + ) raise logger.info( @@ -727,16 +783,163 @@ async def restore_conversation(conversation_id: str, user: User) -> JSONResponse except Exception as e: logger.error(f"Error restoring conversation {conversation_id}: {e}") return JSONResponse( - status_code=500, content={"error": f"Failed to restore conversation: {str(e)}"} + status_code=500, + content={"error": f"Failed to restore conversation: {str(e)}"}, + ) + + +def _enqueue_transcript_reprocessing( + conversation_id: str, + user_id: str, + source: str, + job_id_prefix: str, + end_reason: str, +) -> tuple: + """Enqueue transcribe job + post-conversation chain. + + Returns (version_id, transcript_job, post_jobs dict). + """ + from advanced_omi_backend.workers.transcription_jobs import ( + transcribe_full_audio_job, + ) + + version_id = str(uuid.uuid4()) + + transcript_job = transcription_queue.enqueue( + transcribe_full_audio_job, + conversation_id, + version_id, + source, + job_timeout=get_transcription_job_timeout(), + result_ttl=JOB_RESULT_TTL, + job_id=f"{job_id_prefix}_{conversation_id[:8]}", + description=f"Transcribe audio for {conversation_id[:8]}", + meta={"conversation_id": conversation_id}, + ) + + post_jobs = start_post_conversation_jobs( + conversation_id=conversation_id, + user_id=user_id, + transcript_version_id=version_id, + depends_on_job=transcript_job, + end_reason=end_reason, + ) + + return version_id, transcript_job, post_jobs + + +def _resolve_transcript_version(conversation: Conversation, version_id: str) -> tuple: + """Resolve 'active' to real version ID and find the version object. + + Returns (error_response_or_None, resolved_version_id, version_object). + If error_response is not None, the caller should return it immediately. + """ + resolved_id = version_id + if resolved_id == "active": + active_id = conversation.active_transcript_version + if not active_id: + return ( + JSONResponse( + status_code=404, + content={"error": "No active transcript version found"}, + ), + None, + None, + ) + resolved_id = active_id + + version_obj = None + for v in conversation.transcript_versions: + if v.version_id == resolved_id: + version_obj = v + break + + if not version_obj: + return ( + JSONResponse( + status_code=404, + content={"error": f"Transcript version '{resolved_id}' not found"}, + ), + None, + None, ) + return None, resolved_id, version_obj + + +def _enqueue_speaker_reprocessing_chain( + conversation_id: str, + version_id: str, + source_version_id: str, +) -> dict: + """Enqueue speaker -> memory -> title_summary chain. + + Returns dict with keys: speaker, memory, title_summary (job IDs). + """ + speaker_job = transcription_queue.enqueue( + recognise_speakers_job, + conversation_id, + version_id, + job_timeout=1200, + result_ttl=JOB_RESULT_TTL, + job_id=f"reprocess_speaker_{conversation_id[:12]}", + description=f"Re-diarize speakers for {conversation_id[:8]}", + meta={ + "conversation_id": conversation_id, + "version_id": version_id, + "source_version_id": source_version_id, + "trigger": "reprocess", + }, + ) + logger.info( + f"Enqueued speaker reprocessing job {speaker_job.id} for version {version_id}" + ) + + memory_job = memory_queue.enqueue( + process_memory_job, + conversation_id, + depends_on=speaker_job, + job_timeout=1800, + result_ttl=JOB_RESULT_TTL, + job_id=f"memory_{conversation_id[:12]}", + description=f"Extract memories for {conversation_id[:8]}", + meta={"conversation_id": conversation_id, "trigger": "reprocess_after_speaker"}, + ) + logger.info( + f"Chained memory job {memory_job.id} after speaker job {speaker_job.id}" + ) + + title_summary_job = default_queue.enqueue( + generate_title_summary_job, + conversation_id, + job_timeout=300, + result_ttl=JOB_RESULT_TTL, + depends_on=memory_job, + job_id=f"title_summary_{conversation_id[:12]}", + description=f"Regenerate title/summary for {conversation_id[:8]}", + meta={"conversation_id": conversation_id, "trigger": "reprocess_after_speaker"}, + ) + logger.info( + f"Chained title/summary job {title_summary_job.id} after memory job {memory_job.id}" + ) + + return { + "speaker": speaker_job.id, + "memory": memory_job.id, + "title_summary": title_summary_job.id, + } + async def toggle_star(conversation_id: str, user: User): """Toggle the starred/favorite status of a conversation.""" try: - conversation = await Conversation.find_one(Conversation.conversation_id == conversation_id) + conversation = await Conversation.find_one( + Conversation.conversation_id == conversation_id + ) if not conversation: - return JSONResponse(status_code=404, content={"error": "Conversation not found"}) + return JSONResponse( + status_code=404, content={"error": "Conversation not found"} + ) if not user.is_superuser and conversation.user_id != str(user.user_id): return JSONResponse(status_code=403, content={"error": "Access forbidden"}) @@ -763,7 +966,11 @@ async def toggle_star(conversation_id: str, user: User): data={ "conversation_id": conversation_id, "starred": conversation.starred, - "starred_at": conversation.starred_at.isoformat() if conversation.starred_at else None, + "starred_at": ( + conversation.starred_at.isoformat() + if conversation.starred_at + else None + ), "title": conversation.title, }, ) @@ -773,7 +980,9 @@ async def toggle_star(conversation_id: str, user: User): return { "conversation_id": conversation_id, "starred": conversation.starred, - "starred_at": conversation.starred_at.isoformat() if conversation.starred_at else None, + "starred_at": ( + conversation.starred_at.isoformat() if conversation.starred_at else None + ), } except Exception as e: @@ -784,9 +993,13 @@ async def toggle_star(conversation_id: str, user: User): async def reprocess_orphan(conversation_id: str, user: User): """Reprocess an orphan audio session - restore if deleted and enqueue full processing chain.""" try: - conversation = await Conversation.find_one(Conversation.conversation_id == conversation_id) + conversation = await Conversation.find_one( + Conversation.conversation_id == conversation_id + ) if not conversation: - return JSONResponse(status_code=404, content={"error": "Conversation not found"}) + return JSONResponse( + status_code=404, content={"error": "Conversation not found"} + ) # Check ownership if not user.is_superuser and conversation.user_id != str(user.user_id): @@ -821,33 +1034,12 @@ async def reprocess_orphan(conversation_id: str, user: User): conversation.detailed_summary = None await conversation.save() - # Create new transcript version ID - version_id = str(uuid.uuid4()) - - # Enqueue the same 4-job chain as reprocess_transcript - from advanced_omi_backend.workers.transcription_jobs import ( - transcribe_full_audio_job, - ) - - # Job 1: Transcribe audio - transcript_job = transcription_queue.enqueue( - transcribe_full_audio_job, - conversation_id, - version_id, - "reprocess_orphan", - job_timeout=get_transcription_job_timeout(), - result_ttl=JOB_RESULT_TTL, - job_id=f"orphan_transcribe_{conversation_id[:8]}", - description=f"Transcribe orphan audio for {conversation_id[:8]}", - meta={"conversation_id": conversation_id}, - ) - - # Chain post-transcription jobs (speaker recognition ā†’ memory ā†’ title/summary ā†’ event dispatch) - post_jobs = start_post_conversation_jobs( + # Enqueue the same job chain as reprocess_transcript + version_id, transcript_job, post_jobs = _enqueue_transcript_reprocessing( conversation_id=conversation_id, user_id=str(user.user_id), - transcript_version_id=version_id, - depends_on_job=transcript_job, + source="reprocess_orphan", + job_id_prefix="orphan_transcribe", end_reason="reprocess_orphan", ) @@ -881,7 +1073,9 @@ async def reprocess_transcript(conversation_id: str, user: User): Conversation.conversation_id == conversation_id ) if not conversation_model: - return JSONResponse(status_code=404, content={"error": "Conversation not found"}) + return JSONResponse( + status_code=404, content={"error": "Conversation not found"} + ) # Check ownership for non-admin users if not user.is_superuser and conversation_model.user_id != str(user.user_id): @@ -907,34 +1101,12 @@ async def reprocess_transcript(conversation_id: str, user: User): }, ) - # Create new transcript version ID - version_id = str(uuid.uuid4()) - - # Enqueue job chain with RQ (transcription -> speaker recognition -> memory) - from advanced_omi_backend.workers.transcription_jobs import ( - transcribe_full_audio_job, - ) - - # Job 1: Transcribe audio to text (reconstructs from MongoDB chunks) - transcript_job = transcription_queue.enqueue( - transcribe_full_audio_job, - conversation_id, - version_id, - "reprocess", - job_timeout=get_transcription_job_timeout(), - result_ttl=JOB_RESULT_TTL, - job_id=f"reprocess_{conversation_id[:8]}", - description=f"Transcribe audio for {conversation_id[:8]}", - meta={"conversation_id": conversation_id}, - ) - logger.info(f"šŸ“„ RQ: Enqueued transcription job {transcript_job.id}") - - # Chain post-transcription jobs (speaker recognition ā†’ memory ā†’ title/summary ā†’ event dispatch) - post_jobs = start_post_conversation_jobs( + # Enqueue transcription + post-conversation job chain + version_id, transcript_job, post_jobs = _enqueue_transcript_reprocessing( conversation_id=conversation_id, user_id=str(user.user_id), - transcript_version_id=version_id, - depends_on_job=transcript_job, + source="reprocess", + job_id_prefix="reprocess", end_reason="reprocess_transcript", ) @@ -960,7 +1132,9 @@ async def reprocess_transcript(conversation_id: str, user: User): ) -async def reprocess_memory(conversation_id: str, transcript_version_id: str, user: User): +async def reprocess_memory( + conversation_id: str, transcript_version_id: str, user: User +): """Reprocess memory extraction for a specific transcript version. Users can only reprocess their own conversations.""" try: # Find the conversation using Beanie @@ -968,7 +1142,9 @@ async def reprocess_memory(conversation_id: str, transcript_version_id: str, use Conversation.conversation_id == conversation_id ) if not conversation_model: - return JSONResponse(status_code=404, content={"error": "Conversation not found"}) + return JSONResponse( + status_code=404, content={"error": "Conversation not found"} + ) # Check ownership for non-admin users if not user.is_superuser and conversation_model.user_id != str(user.user_id): @@ -979,28 +1155,12 @@ async def reprocess_memory(conversation_id: str, transcript_version_id: str, use }, ) - # Resolve transcript version ID - # Handle special "active" version ID - if transcript_version_id == "active": - active_version_id = conversation_model.active_transcript_version - if not active_version_id: - return JSONResponse( - status_code=404, content={"error": "No active transcript version found"} - ) - transcript_version_id = active_version_id - - # Find the specific transcript version - transcript_version = None - for version in conversation_model.transcript_versions: - if version.version_id == transcript_version_id: - transcript_version = version - break - - if not transcript_version: - return JSONResponse( - status_code=404, - content={"error": f"Transcript version '{transcript_version_id}' not found"}, - ) + # Resolve transcript version ID (handle "active" special case) + error, transcript_version_id, transcript_version = _resolve_transcript_version( + conversation_model, transcript_version_id + ) + if error: + return error # Create new memory version ID version_id = str(uuid.uuid4()) @@ -1033,7 +1193,9 @@ async def reprocess_memory(conversation_id: str, transcript_version_id: str, use ) -async def reprocess_speakers(conversation_id: str, transcript_version_id: str, user: User): +async def reprocess_speakers( + conversation_id: str, transcript_version_id: str, user: User +): """ Reprocess speaker identification for a specific transcript version. Users can only reprocess their own conversations. @@ -1047,7 +1209,9 @@ async def reprocess_speakers(conversation_id: str, transcript_version_id: str, u Conversation.conversation_id == conversation_id ) if not conversation_model: - return JSONResponse(status_code=404, content={"error": "Conversation not found"}) + return JSONResponse( + status_code=404, content={"error": "Conversation not found"} + ) # Check ownership for non-admin users if not user.is_superuser and conversation_model.user_id != str(user.user_id): @@ -1058,28 +1222,12 @@ async def reprocess_speakers(conversation_id: str, transcript_version_id: str, u }, ) - # 2. Resolve source transcript version ID (handle "active" special case) - source_version_id = transcript_version_id - if source_version_id == "active": - active_version_id = conversation_model.active_transcript_version - if not active_version_id: - return JSONResponse( - status_code=404, content={"error": "No active transcript version found"} - ) - source_version_id = active_version_id - - # 3. Find and validate the source transcript version - source_version = None - for version in conversation_model.transcript_versions: - if version.version_id == source_version_id: - source_version = version - break - - if not source_version: - return JSONResponse( - status_code=404, - content={"error": f"Transcript version '{source_version_id}' not found"}, - ) + # 2-3. Resolve source transcript version ID and find version object + error, source_version_id, source_version = _resolve_transcript_version( + conversation_model, transcript_version_id + ) + if error: + return error # 4. Validate transcript has content and words (or provider-diarized segments) if not source_version.transcript: @@ -1169,72 +1317,22 @@ async def reprocess_speakers(conversation_id: str, transcript_version_id: str, u f"for conversation {conversation_id}" ) - # 7. Enqueue speaker recognition job with NEW version_id - speaker_job = transcription_queue.enqueue( - recognise_speakers_job, + # 7-8. Enqueue speaker ā†’ memory ā†’ title/summary chain + job_ids = _enqueue_speaker_reprocessing_chain( conversation_id, - new_version_id, # NEW version (not source) - job_timeout=1200, # 20 minutes - result_ttl=JOB_RESULT_TTL, - job_id=f"reprocess_speaker_{conversation_id[:12]}", - description=f"Re-diarize speakers for {conversation_id[:8]}", - meta={ - "conversation_id": conversation_id, - "version_id": new_version_id, - "source_version_id": source_version_id, - "trigger": "reprocess", - }, - ) - - logger.info( - f"Enqueued speaker reprocessing job {speaker_job.id} " - f"for new version {new_version_id}" - ) - - # 8. Chain memory reprocessing (speaker changes affect memory context) - memory_job = memory_queue.enqueue( - process_memory_job, - conversation_id, - depends_on=speaker_job, - job_timeout=1800, # 30 minutes - result_ttl=JOB_RESULT_TTL, - job_id=f"memory_{conversation_id[:12]}", - description=f"Extract memories for {conversation_id[:8]}", - meta={"conversation_id": conversation_id, "trigger": "reprocess_after_speaker"}, - ) - - logger.info( - f"Chained memory reprocessing job {memory_job.id} " - f"after speaker job {speaker_job.id}" - ) - - # 8b. Chain title/summary regeneration after memory job - # Depends on memory_job to avoid race condition (both save conversation document) - # and to ensure fresh memories are available for context-enriched summaries - title_summary_job = default_queue.enqueue( - generate_title_summary_job, - conversation_id, - job_timeout=300, - result_ttl=JOB_RESULT_TTL, - depends_on=memory_job, - job_id=f"title_summary_{conversation_id[:12]}", - description=f"Regenerate title/summary for {conversation_id[:8]}", - meta={"conversation_id": conversation_id, "trigger": "reprocess_after_speaker"}, - ) - - logger.info( - f"Chained title/summary job {title_summary_job.id} " f"after memory job {memory_job.id}" + new_version_id, + source_version_id, ) # 9. Return job information return JSONResponse( content={ "message": "Speaker reprocessing started", - "job_id": speaker_job.id, - "memory_job_id": memory_job.id, - "title_summary_job_id": title_summary_job.id, - "version_id": new_version_id, # NEW version ID - "source_version_id": source_version_id, # Original version used as source + "job_id": job_ids["speaker"], + "memory_job_id": job_ids["memory"], + "title_summary_job_id": job_ids["title_summary"], + "version_id": new_version_id, + "source_version_id": source_version_id, "status": "queued", } ) @@ -1246,7 +1344,9 @@ async def reprocess_speakers(conversation_id: str, transcript_version_id: str, u ) -async def activate_transcript_version(conversation_id: str, version_id: str, user: User): +async def activate_transcript_version( + conversation_id: str, version_id: str, user: User +): """Activate a specific transcript version. Users can only modify their own conversations.""" try: # Find the conversation using Beanie @@ -1254,20 +1354,25 @@ async def activate_transcript_version(conversation_id: str, version_id: str, use Conversation.conversation_id == conversation_id ) if not conversation_model: - return JSONResponse(status_code=404, content={"error": "Conversation not found"}) + return JSONResponse( + status_code=404, content={"error": "Conversation not found"} + ) # Check ownership for non-admin users if not user.is_superuser and conversation_model.user_id != str(user.user_id): return JSONResponse( status_code=403, - content={"error": "Access forbidden. You can only modify your own conversations."}, + content={ + "error": "Access forbidden. You can only modify your own conversations." + }, ) # Activate the transcript version using Beanie model method success = conversation_model.set_active_transcript_version(version_id) if not success: return JSONResponse( - status_code=400, content={"error": "Failed to activate transcript version"} + status_code=400, + content={"error": "Failed to activate transcript version"}, ) await conversation_model.save() @@ -1301,13 +1406,17 @@ async def activate_memory_version(conversation_id: str, version_id: str, user: U Conversation.conversation_id == conversation_id ) if not conversation_model: - return JSONResponse(status_code=404, content={"error": "Conversation not found"}) + return JSONResponse( + status_code=404, content={"error": "Conversation not found"} + ) # Check ownership for non-admin users if not user.is_superuser and conversation_model.user_id != str(user.user_id): return JSONResponse( status_code=403, - content={"error": "Access forbidden. You can only modify your own conversations."}, + content={ + "error": "Access forbidden. You can only modify your own conversations." + }, ) # Activate the memory version using Beanie model method @@ -1332,7 +1441,9 @@ async def activate_memory_version(conversation_id: str, version_id: str, user: U except Exception as e: logger.error(f"Error activating memory version: {e}") - return JSONResponse(status_code=500, content={"error": "Error activating memory version"}) + return JSONResponse( + status_code=500, content={"error": "Error activating memory version"} + ) async def get_conversation_version_history(conversation_id: str, user: User): @@ -1343,13 +1454,17 @@ async def get_conversation_version_history(conversation_id: str, user: User): Conversation.conversation_id == conversation_id ) if not conversation_model: - return JSONResponse(status_code=404, content={"error": "Conversation not found"}) + return JSONResponse( + status_code=404, content={"error": "Conversation not found"} + ) # Check ownership for non-admin users if not user.is_superuser and conversation_model.user_id != str(user.user_id): return JSONResponse( status_code=403, - content={"error": "Access forbidden. You can only access your own conversations."}, + content={ + "error": "Access forbidden. You can only access your own conversations." + }, ) # Get version history from model @@ -1380,4 +1495,6 @@ async def get_conversation_version_history(conversation_id: str, user: User): except Exception as e: logger.error(f"Error fetching version history: {e}") - return JSONResponse(status_code=500, content={"error": "Error fetching version history"}) + return JSONResponse( + status_code=500, content={"error": "Error fetching version history"} + ) diff --git a/backends/advanced/src/advanced_omi_backend/controllers/websocket_controller.py b/backends/advanced/src/advanced_omi_backend/controllers/websocket_controller.py index 4bb1088a..bab956ec 100644 --- a/backends/advanced/src/advanced_omi_backend/controllers/websocket_controller.py +++ b/backends/advanced/src/advanced_omi_backend/controllers/websocket_controller.py @@ -1,4 +1,3 @@ - """ WebSocket controller for Chronicle backend. @@ -32,7 +31,6 @@ from advanced_omi_backend.services.audio_stream.producer import ( get_audio_stream_producer, ) -from advanced_omi_backend.utils.audio_utils import process_audio_chunk # Thread pool executors for audio decoding _DEC_IO_EXECUTOR = concurrent.futures.ThreadPoolExecutor( @@ -82,29 +80,34 @@ async def subscribe_to_interim_results(websocket: WebSocket, session_id: str) -> # Listen for messages while True: try: - message = await pubsub.get_message(ignore_subscribe_messages=True, timeout=1.0) + message = await pubsub.get_message( + ignore_subscribe_messages=True, timeout=1.0 + ) - if message and message['type'] == 'message': + if message and message["type"] == "message": # Parse result data try: - result_data = json.loads(message['data']) + result_data = json.loads(message["data"]) # Forward to client WebSocket - await websocket.send_json({ - "type": "interim_transcript", - "data": result_data - }) + await websocket.send_json( + {"type": "interim_transcript", "data": result_data} + ) # Log for debugging is_final = result_data.get("is_final", False) text_preview = result_data.get("text", "")[:50] result_type = "FINAL" if is_final else "interim" - logger.debug(f"āœ‰ļø Forwarded {result_type} result to client {session_id}: {text_preview}...") + logger.debug( + f"āœ‰ļø Forwarded {result_type} result to client {session_id}: {text_preview}..." + ) except json.JSONDecodeError as e: logger.error(f"Failed to parse interim result JSON: {e}") except Exception as send_error: - logger.error(f"Failed to send interim result to client {session_id}: {send_error}") + logger.error( + f"Failed to send interim result to client {session_id}: {send_error}" + ) # WebSocket might be closed, exit loop break @@ -112,14 +115,22 @@ async def subscribe_to_interim_results(websocket: WebSocket, session_id: str) -> # No message received, continue waiting continue except asyncio.CancelledError: - logger.info(f"Interim results subscriber cancelled for session {session_id}") + logger.info( + f"Interim results subscriber cancelled for session {session_id}" + ) break except Exception as e: - logger.error(f"Error in interim results subscriber for {session_id}: {e}", exc_info=True) + logger.error( + f"Error in interim results subscriber for {session_id}: {e}", + exc_info=True, + ) break except Exception as e: - logger.error(f"Failed to initialize interim results subscriber for {session_id}: {e}", exc_info=True) + logger.error( + f"Failed to initialize interim results subscriber for {session_id}: {e}", + exc_info=True, + ) finally: try: # Unsubscribe and close connections @@ -128,7 +139,9 @@ async def subscribe_to_interim_results(websocket: WebSocket, session_id: str) -> await redis_client.aclose() logger.info(f"šŸ”• Unsubscribed from interim results channel: {channel}") except Exception as cleanup_error: - logger.error(f"Error cleaning up interim results subscriber: {cleanup_error}") + logger.error( + f"Error cleaning up interim results subscriber: {cleanup_error}" + ) async def parse_wyoming_protocol(ws: WebSocket) -> tuple[dict, Optional[bytes]]: @@ -140,14 +153,18 @@ async def parse_wyoming_protocol(ws: WebSocket) -> tuple[dict, Optional[bytes]]: # Read data from WebSocket logger.debug(f"parse_wyoming_protocol: About to call ws.receive()") message = await ws.receive() - logger.debug(f"parse_wyoming_protocol: Received message with keys: {message.keys() if message else 'None'}") + logger.debug( + f"parse_wyoming_protocol: Received message with keys: {message.keys() if message else 'None'}" + ) # Handle WebSocket close frame if "type" in message and message["type"] == "websocket.disconnect": # This is a normal WebSocket close event code = message.get("code", 1000) reason = message.get("reason", "") - logger.info(f"šŸ““ WebSocket disconnect received in parse_wyoming_protocol. Code: {code}, Reason: {reason}") + logger.info( + f"šŸ““ WebSocket disconnect received in parse_wyoming_protocol. Code: {code}, Reason: {reason}" + ) raise WebSocketDisconnect(code=code, reason=reason) # Handle text message (JSON header) @@ -190,7 +207,10 @@ async def create_client_state(client_id: str, user, device_name: Optional[str] = # Directory where WAV chunks are written from pathlib import Path - CHUNK_DIR = Path("./audio_chunks") # This will be mounted to ./data/audio_chunks by Docker + + CHUNK_DIR = Path( + "./audio_chunks" + ) # This will be mounted to ./data/audio_chunks by Docker # Use ClientManager for atomic client creation and registration client_state = client_manager.create_client( @@ -199,10 +219,12 @@ async def create_client_state(client_id: str, user, device_name: Optional[str] = # Also track in persistent mapping (for database queries + cross-container Redis) from advanced_omi_backend.client_manager import track_client_user_relationship_async + await track_client_user_relationship_async(client_id, user.user_id) # Register client in user model (persistent) from advanced_omi_backend.users import register_client_to_user + await register_client_to_user(user, client_id, device_name) return client_state @@ -224,7 +246,9 @@ async def cleanup_client_state(client_id: str): # The speech detection job now monitors session status and completes naturally. import redis.asyncio as redis - logger.info(f"šŸ”„ Letting speech detection job complete naturally for client {client_id} (if running)") + logger.info( + f"šŸ”„ Letting speech detection job complete naturally for client {client_id} (if running)" + ) # Mark all active sessions for this client as complete AND delete Redis streams try: @@ -236,6 +260,7 @@ async def cleanup_client_state(client_id: str): from advanced_omi_backend.services.audio_stream.producer import ( get_audio_stream_producer, ) + audio_stream_producer = get_audio_stream_producer() # Find all session keys for this client and mark them complete @@ -258,18 +283,26 @@ async def cleanup_client_state(client_id: str): # If session is still active, finalize it first (sets status + completion_reason atomically) if status in ["active", None]: - logger.info(f"šŸ“Š Finalizing active session {session_id[:12]} due to WebSocket disconnect") - await audio_stream_producer.finalize_session(session_id, completion_reason="websocket_disconnect") + logger.info( + f"šŸ“Š Finalizing active session {session_id[:12]} due to WebSocket disconnect" + ) + await audio_stream_producer.finalize_session( + session_id, completion_reason="websocket_disconnect" + ) # Mark session as complete (WebSocket disconnected) - await mark_session_complete(async_redis, session_id, "websocket_disconnect") + await mark_session_complete( + async_redis, session_id, "websocket_disconnect" + ) sessions_closed += 1 if cursor == 0: break if sessions_closed > 0: - logger.info(f"āœ… Closed {sessions_closed} active session(s) for client {client_id}") + logger.info( + f"āœ… Closed {sessions_closed} active session(s) for client {client_id}" + ) # Set TTL on Redis Streams for this client (allows consumer groups to finish processing) stream_pattern = f"audio:stream:{client_id}" @@ -282,9 +315,11 @@ async def cleanup_client_state(client_id: str): pending_count = 0 try: # Check streaming-transcription consumer group for pending messages - pending_info = await async_redis.xpending(stream_pattern, "streaming-transcription") + pending_info = await async_redis.xpending( + stream_pattern, "streaming-transcription" + ) if pending_info: - pending_count = pending_info.get('pending', 0) + pending_count = pending_info.get("pending", 0) except Exception as e: # Consumer group might not exist yet - that's ok logger.debug(f"No consumer group for {stream_pattern}: {e}") @@ -295,7 +330,9 @@ async def cleanup_client_state(client_id: str): f"{stream_length} messages in stream, {pending_count} pending in consumer group" ) - await async_redis.expire(stream_pattern, 60) # 60 second TTL for consumer group fan-out + await async_redis.expire( + stream_pattern, 60 + ) # 60 second TTL for consumer group fan-out logger.info(f"ā° Set 60s TTL on Redis stream: {stream_pattern}") else: logger.debug(f"No Redis stream found for client {client_id}") @@ -303,7 +340,9 @@ async def cleanup_client_state(client_id: str): await async_redis.close() except Exception as session_error: - logger.warning(f"āš ļø Error marking sessions complete for client {client_id}: {session_error}") + logger.warning( + f"āš ļø Error marking sessions complete for client {client_id}: {session_error}" + ) # Use ClientManager for atomic client removal with cleanup client_manager = get_client_manager() @@ -321,7 +360,7 @@ async def _setup_websocket_connection( token: Optional[str], device_name: Optional[str], pending_client_id: str, - connection_type: str + connection_type: str, ) -> tuple[Optional[str], Optional[object], Optional[object]]: """ Setup WebSocket connection: accept, authenticate, create client state. @@ -344,12 +383,17 @@ async def _setup_websocket_connection( if not user: # Send error message to client before closing try: - error_msg = json.dumps({ - "type": "error", - "error": "authentication_failed", - "message": "Authentication failed. Please log in again and ensure your token is valid.", - "code": 1008 - }) + "\n" + error_msg = ( + json.dumps( + { + "type": "error", + "error": "authentication_failed", + "message": "Authentication failed. Please log in again and ensure your token is valid.", + "code": 1008, + } + ) + + "\n" + ) await ws.send_text(error_msg) application_logger.info("Sent authentication error message to client") except Exception as send_error: @@ -370,7 +414,10 @@ async def _setup_websocket_connection( # Send ready message to confirm connection is established try: - ready_msg = json.dumps({"type": "ready", "message": "WebSocket connection established"}) + "\n" + ready_msg = ( + json.dumps({"type": "ready", "message": "WebSocket connection established"}) + + "\n" + ) await ws.send_text(ready_msg) application_logger.debug(f"āœ… Sent ready message to {client_id}") except Exception as e: @@ -389,7 +436,7 @@ async def _initialize_streaming_session( user_email: str, client_id: str, audio_format: dict, - websocket: Optional[WebSocket] = None + websocket: Optional[WebSocket] = None, ) -> Optional[asyncio.Task]: """ Initialize streaming session with Redis and enqueue processing jobs. @@ -410,30 +457,38 @@ async def _initialize_streaming_session( f"šŸ”“ BACKEND: _initialize_streaming_session called for {client_id}" ) - if hasattr(client_state, 'stream_session_id'): + if hasattr(client_state, "stream_session_id"): application_logger.debug(f"Session already initialized for {client_id}") return None # Initialize stream session - use client_id as session_id for predictable lookup # All other session metadata goes to Redis (single source of truth) client_state.stream_session_id = client_state.client_id - application_logger.info(f"šŸ†” Created stream session: {client_state.stream_session_id}") + application_logger.info( + f"šŸ†” Created stream session: {client_state.stream_session_id}" + ) # Determine transcription provider from config.yml from advanced_omi_backend.model_registry import get_models_registry registry = get_models_registry() if not registry: - raise ValueError("config.yml not found - cannot determine transcription provider") + raise ValueError( + "config.yml not found - cannot determine transcription provider" + ) stt_model = registry.get_default("stt") if not stt_model: raise ValueError("No default STT model configured in config.yml (defaults.stt)") # Use model_provider for session tracking (generic, not validated against hardcoded list) - provider = stt_model.model_provider.lower() if stt_model.model_provider else stt_model.name + provider = ( + stt_model.model_provider.lower() if stt_model.model_provider else stt_model.name + ) - application_logger.info(f"šŸ“‹ Using STT provider: {provider} (model: {stt_model.name})") + application_logger.info( + f"šŸ“‹ Using STT provider: {provider} (model: {stt_model.name})" + ) # Initialize session tracking in Redis (SINGLE SOURCE OF TRUTH for session metadata) # This includes user_email, connection info, audio format, chunk counters, job IDs, etc. @@ -445,7 +500,7 @@ async def _initialize_streaming_session( user_email=user_email, connection_id=connection_id, mode="streaming", - provider=provider + provider=provider, ) # Store audio format in Redis session (not in ClientState) @@ -454,6 +509,7 @@ async def _initialize_streaming_session( from advanced_omi_backend.services.audio_stream.producer import ( get_audio_stream_producer, ) + session_key = f"audio:session:{client_state.stream_session_id}" redis_client = audio_stream_producer.redis_client await redis_client.hset(session_key, "audio_format", json.dumps(audio_format)) @@ -462,16 +518,14 @@ async def _initialize_streaming_session( from advanced_omi_backend.controllers.queue_controller import start_streaming_jobs job_ids = start_streaming_jobs( - session_id=client_state.stream_session_id, - user_id=user_id, - client_id=client_id + session_id=client_state.stream_session_id, user_id=user_id, client_id=client_id ) # Store job IDs in Redis session (not in ClientState) await audio_stream_producer.update_session_job_ids( session_id=client_state.stream_session_id, - speech_detection_job_id=job_ids['speech_detection'], - audio_persistence_job_id=job_ids['audio_persistence'] + speech_detection_job_id=job_ids["speech_detection"], + audio_persistence_job_id=job_ids["audio_persistence"], ) # Note: Placeholder conversation creation is handled by the audio persistence job, @@ -483,17 +537,15 @@ async def _initialize_streaming_session( subscriber_task = asyncio.create_task( subscribe_to_interim_results(websocket, client_state.stream_session_id) ) - application_logger.info(f"šŸ“” Launched interim results subscriber for session {client_state.stream_session_id}") + application_logger.info( + f"šŸ“” Launched interim results subscriber for session {client_state.stream_session_id}" + ) return subscriber_task async def _finalize_streaming_session( - client_state, - audio_stream_producer, - user_id: str, - user_email: str, - client_id: str + client_state, audio_stream_producer, user_id: str, user_email: str, client_id: str ) -> None: """ Finalize streaming session: flush buffer, signal workers, enqueue finalize job, cleanup. @@ -505,7 +557,7 @@ async def _finalize_streaming_session( user_email: User email client_id: Client ID """ - if not hasattr(client_state, 'stream_session_id'): + if not hasattr(client_state, "stream_session_id"): application_logger.debug(f"No active session to finalize for {client_id}") return @@ -513,19 +565,21 @@ async def _finalize_streaming_session( try: # Flush any remaining buffered audio - audio_format = getattr(client_state, 'stream_audio_format', {}) + audio_format = getattr(client_state, "stream_audio_format", {}) await audio_stream_producer.flush_session_buffer( session_id=session_id, sample_rate=audio_format.get("rate", 16000), channels=audio_format.get("channels", 1), - sample_width=audio_format.get("width", 2) + sample_width=audio_format.get("width", 2), ) # Send end-of-session signal to workers await audio_stream_producer.send_session_end_signal(session_id) # Mark session as finalizing with user_stopped reason (audio-stop event) - await audio_stream_producer.finalize_session(session_id, completion_reason="user_stopped") + await audio_stream_producer.finalize_session( + session_id, completion_reason="user_stopped" + ) # Store markers in Redis so open_conversation_job can persist them if client_state.markers: @@ -556,13 +610,12 @@ async def _finalize_streaming_session( # Clear session state from ClientState (only stream_session_id is stored there now) # All other session metadata lives in Redis (single source of truth) - if hasattr(client_state, 'stream_session_id'): - delattr(client_state, 'stream_session_id') + if hasattr(client_state, "stream_session_id"): + delattr(client_state, "stream_session_id") except Exception as finalize_error: application_logger.error( - f"āŒ Failed to finalize streaming session: {finalize_error}", - exc_info=True + f"āŒ Failed to finalize streaming session: {finalize_error}", exc_info=True ) @@ -574,7 +627,7 @@ async def _publish_audio_to_stream( client_id: str, sample_rate: int, channels: int, - sample_width: int + sample_width: int, ) -> None: """ Publish audio chunk to Redis Stream with chunk tracking. @@ -589,28 +642,23 @@ async def _publish_audio_to_stream( channels: Number of channels sample_width: Bytes per sample """ - if not hasattr(client_state, 'stream_session_id'): - application_logger.warning(f"āš ļø Received audio chunk before session initialized for {client_id}") + if not hasattr(client_state, "stream_session_id"): + application_logger.warning( + f"āš ļø Received audio chunk before session initialized for {client_id}" + ) return session_id = client_state.stream_session_id - # Increment chunk count in Redis (single source of truth) and format chunk ID - session_key = f"audio:session:{session_id}" - redis_client = audio_stream_producer.redis_client - chunk_count = await redis_client.hincrby(session_key, "chunks_published", 1) - chunk_id = f"{chunk_count:05d}" - - # Publish to Redis Stream using producer + # Publish to Redis Stream using producer (producer owns chunk counting) await audio_stream_producer.add_audio_chunk( audio_data=audio_data, session_id=session_id, - chunk_id=chunk_id, user_id=user_id, client_id=client_id, sample_rate=sample_rate, channels=channels, - sample_width=sample_width + sample_width=sample_width, ) @@ -621,7 +669,7 @@ async def _handle_omi_audio_chunk( decode_packet_fn, user_id: str, client_id: str, - packet_count: int + packet_count: int, ) -> None: """ Handle OMI audio chunk: decode Opus to PCM, then publish to stream. @@ -638,7 +686,9 @@ async def _handle_omi_audio_chunk( # Decode Opus to PCM start_time = time.time() loop = asyncio.get_running_loop() - pcm_data = await loop.run_in_executor(_DEC_IO_EXECUTOR, decode_packet_fn, opus_payload) + pcm_data = await loop.run_in_executor( + _DEC_IO_EXECUTOR, decode_packet_fn, opus_payload + ) decode_time = time.time() - start_time if pcm_data: @@ -657,7 +707,7 @@ async def _handle_omi_audio_chunk( client_id, OMI_SAMPLE_RATE, OMI_CHANNELS, - OMI_SAMPLE_WIDTH + OMI_SAMPLE_WIDTH, ) else: # Log decode failures for first 5 packets @@ -675,7 +725,7 @@ async def _handle_streaming_mode_audio( user_id: str, user_email: str, client_id: str, - websocket: Optional[WebSocket] = None + websocket: Optional[WebSocket] = None, ) -> Optional[asyncio.Task]: """ Handle audio chunk in streaming mode. @@ -695,7 +745,7 @@ async def _handle_streaming_mode_audio( """ # Initialize session if needed subscriber_task = None - if not hasattr(client_state, 'stream_session_id'): + if not hasattr(client_state, "stream_session_id"): subscriber_task = await _initialize_streaming_session( client_state, audio_stream_producer, @@ -703,7 +753,7 @@ async def _handle_streaming_mode_audio( user_email, client_id, audio_format, - websocket=websocket # Pass WebSocket to launch interim results subscriber + websocket=websocket, # Pass WebSocket to launch interim results subscriber ) # Publish to Redis Stream @@ -715,17 +765,14 @@ async def _handle_streaming_mode_audio( client_id, audio_format.get("rate", 16000), audio_format.get("channels", 1), - audio_format.get("width", 2) + audio_format.get("width", 2), ) return subscriber_task async def _handle_batch_mode_audio( - client_state, - audio_data: bytes, - audio_format: dict, - client_id: str + client_state, audio_data: bytes, audio_format: dict, client_id: str ) -> None: """ Handle audio chunk in batch mode with rolling 30-minute limit. @@ -737,7 +784,7 @@ async def _handle_batch_mode_audio( client_id: Client ID """ # Initialize batch accumulator if needed - if not hasattr(client_state, 'batch_audio_chunks'): + if not hasattr(client_state, "batch_audio_chunks"): client_state.batch_audio_chunks = [] client_state.batch_audio_format = audio_format client_state.batch_audio_bytes = 0 # Track total bytes @@ -774,7 +821,7 @@ async def _handle_batch_mode_audio( user_id=client_state.user_id, # Need to store these on session start user_email=client_state.user_email, client_id=client_state.client_id, - batch_number=client_state.batch_chunks_processed + 1 + batch_number=client_state.batch_chunks_processed + 1, ) # Clear buffer for next batch @@ -796,7 +843,7 @@ async def _handle_audio_chunk( user_id: str, user_email: str, client_id: str, - websocket: Optional[WebSocket] = None + websocket: Optional[WebSocket] = None, ) -> Optional[asyncio.Task]: """ Route audio chunk to appropriate mode handler (streaming or batch). @@ -814,13 +861,18 @@ async def _handle_audio_chunk( Returns: Interim results subscriber task if websocket provided and streaming mode, None otherwise """ - recording_mode = getattr(client_state, 'recording_mode', 'batch') + recording_mode = getattr(client_state, "recording_mode", "batch") if recording_mode == "streaming": return await _handle_streaming_mode_audio( - client_state, audio_stream_producer, audio_data, - audio_format, user_id, user_email, client_id, - websocket=websocket + client_state, + audio_stream_producer, + audio_data, + audio_format, + user_id, + user_email, + client_id, + websocket=websocket, ) else: await _handle_batch_mode_audio( @@ -833,7 +885,7 @@ async def _handle_audio_session_start( client_state, audio_format: dict, client_id: str, - websocket: Optional[WebSocket] = None + websocket: Optional[WebSocket] = None, ) -> tuple[bool, str]: """ Handle audio-start event - validate mode and set recording mode. @@ -878,14 +930,20 @@ async def _handle_audio_session_start( "type": "error", "error": "streaming_not_configured", "message": error_msg, - "code": 400 + "code": 400, } await websocket.send_json(error_response) - application_logger.info(f"šŸ“¤ Sent streaming error to WebUI client {client_id}") + application_logger.info( + f"šŸ“¤ Sent streaming error to WebUI client {client_id}" + ) # Close the websocket connection after sending error - await websocket.close(code=1008, reason="Streaming transcription not configured") - application_logger.info(f"šŸ”Œ Closed WebSocket connection for {client_id} due to streaming config error") + await websocket.close( + code=1008, reason="Streaming transcription not configured" + ) + application_logger.info( + f"šŸ”Œ Closed WebSocket connection for {client_id} due to streaming config error" + ) # Raise ValueError to exit the handler completely raise ValueError(error_msg) @@ -917,11 +975,7 @@ async def _handle_audio_session_start( async def _handle_audio_session_stop( - client_state, - audio_stream_producer, - user_id: str, - user_email: str, - client_id: str + client_state, audio_stream_producer, user_id: str, user_email: str, client_id: str ) -> bool: """ Handle audio-stop event - finalize session based on mode. @@ -936,13 +990,14 @@ async def _handle_audio_session_stop( Returns: False to switch back to control mode """ - recording_mode = getattr(client_state, 'recording_mode', 'batch') - application_logger.info(f"šŸ›‘ Audio session stopped for {client_id} (mode: {recording_mode})") + recording_mode = getattr(client_state, "recording_mode", "batch") + application_logger.info( + f"šŸ›‘ Audio session stopped for {client_id} (mode: {recording_mode})" + ) if recording_mode == "streaming": await _finalize_streaming_session( - client_state, audio_stream_producer, - user_id, user_email, client_id + client_state, audio_stream_producer, user_id, user_email, client_id ) else: await _process_batch_audio_complete( @@ -965,14 +1020,11 @@ async def _handle_button_event( Args: client_state: Client state object - button_state: Button state string (e.g., "SINGLE_TAP", "DOUBLE_TAP") + button_state: Button state string (e.g., "SINGLE_PRESS", "DOUBLE_PRESS") user_id: User ID client_id: Client ID """ - from advanced_omi_backend.plugins.events import ( - BUTTON_STATE_TO_EVENT, - ButtonState, - ) + from advanced_omi_backend.plugins.events import BUTTON_STATE_TO_EVENT, ButtonState from advanced_omi_backend.services.plugin_service import get_plugin_router timestamp = time.time() @@ -993,7 +1045,6 @@ async def _handle_button_event( } client_state.add_marker(marker) - # Map device button state to typed plugin event try: button_state_enum = ButtonState(button_state) @@ -1016,243 +1067,276 @@ async def _handle_button_event( "state": button_state_enum.value, "timestamp": timestamp, "audio_uuid": audio_uuid, - "session_id": getattr(client_state, 'stream_session_id', None), + "session_id": getattr(client_state, "stream_session_id", None), "client_id": client_id, }, ) -async def _process_rolling_batch( +async def _create_batch_conversation_and_enqueue( client_state, user_id: str, - user_email: str, client_id: str, - batch_number: int -) -> None: - """ - Process accumulated batch audio as a rolling segment. - - Creates conversation titled "Recording Part {batch_number}" and enqueues transcription. + title: str, + trigger: str, + job_id_prefix: str, + enqueue_post_jobs: bool = False, + attach_markers: bool = False, +) -> Optional[str]: + """Create conversation from batch audio, store chunks, enqueue transcription. Args: client_state: Client state with batch_audio_chunks user_id: User ID - user_email: User email client_id: Client ID - batch_number: Sequential batch number (1, 2, 3...) - """ - if not hasattr(client_state, 'batch_audio_chunks') or not client_state.batch_audio_chunks: - application_logger.warning(f"āš ļø No audio chunks to process for rolling batch") - return + title: Conversation title + trigger: Trigger string for transcription job + job_id_prefix: Prefix for the transcription job ID + enqueue_post_jobs: If True, chain post-conversation jobs after transcription + attach_markers: If True, copy client_state.markers to conversation - try: - from advanced_omi_backend.models.conversation import create_conversation - from advanced_omi_backend.utils.audio_chunk_utils import convert_audio_to_chunks + Returns: + conversation_id on success, None on failure. + """ + from advanced_omi_backend.config import get_transcription_job_timeout + from advanced_omi_backend.controllers.queue_controller import ( + JOB_RESULT_TTL, + transcription_queue, + ) + from advanced_omi_backend.models.conversation import create_conversation + from advanced_omi_backend.utils.audio_chunk_utils import convert_audio_to_chunks + from advanced_omi_backend.workers.transcription_jobs import ( + transcribe_full_audio_job, + ) - # Combine chunks - complete_audio = b''.join(client_state.batch_audio_chunks) - application_logger.info( - f"šŸ“¦ Rolling batch #{batch_number}: Combined {len(client_state.batch_audio_chunks)} chunks " - f"into {len(complete_audio)} bytes" - ) + complete_audio = b"".join(client_state.batch_audio_chunks) + audio_format = getattr(client_state, "batch_audio_format", {}) + sample_rate = audio_format.get("rate", 16000) + sample_width = audio_format.get("width", 2) + channels = audio_format.get("channels", 1) - # Get audio format - audio_format = getattr(client_state, 'batch_audio_format', {}) - sample_rate = audio_format.get("rate", 16000) - width = audio_format.get("width", 2) - channels = audio_format.get("channels", 1) + application_logger.info( + f"šŸ“¦ Batch: Combined {len(client_state.batch_audio_chunks)} chunks " + f"into {len(complete_audio)} bytes (title={title})" + ) - # Create conversation with batch number in title - conversation = create_conversation( - user_id=user_id, - client_id=client_id, - title=f"Recording Part {batch_number}", - summary="Rolling batch processing..." - ) - await conversation.insert() - conversation_id = conversation.conversation_id # Get the auto-generated ID + # Create conversation + conversation = create_conversation( + user_id=user_id, + client_id=client_id, + title=title, + summary="Processing batch audio...", + ) + if attach_markers and client_state.markers: + conversation.markers = list(client_state.markers) + client_state.markers.clear() + await conversation.insert() + conversation_id = conversation.conversation_id - # Convert to MongoDB chunks + # Convert audio to MongoDB chunks + try: num_chunks = await convert_audio_to_chunks( conversation_id=conversation_id, audio_data=complete_audio, sample_rate=sample_rate, channels=channels, - sample_width=width + sample_width=sample_width, ) - - # Enqueue transcription job - from advanced_omi_backend.controllers.queue_controller import ( - JOB_RESULT_TTL, - transcription_queue, + application_logger.info( + f"šŸ“¦ Batch: Converted to {num_chunks} MongoDB chunks ({conversation_id[:12]})" ) - from advanced_omi_backend.workers.transcription_jobs import ( - transcribe_full_audio_job, + except Exception as chunk_error: + application_logger.error( + f"Failed to convert batch audio to chunks: {chunk_error}", exc_info=True ) - version_id = str(uuid.uuid4()) - transcribe_job_id = f"transcribe_rolling_{conversation_id[:12]}_{batch_number}" - - from advanced_omi_backend.config import get_transcription_job_timeout - - transcription_job = transcription_queue.enqueue( - transcribe_full_audio_job, - conversation_id, - version_id, - f"rolling_batch_{batch_number}", # trigger - job_timeout=get_transcription_job_timeout(), - result_ttl=JOB_RESULT_TTL, - job_id=transcribe_job_id, - description=f"Transcribe rolling batch #{batch_number} {conversation_id[:8]}", - meta={'conversation_id': conversation_id, 'client_id': client_id, 'batch_number': batch_number} + # Enqueue transcription job + version_id = str(uuid.uuid4()) + transcription_job = transcription_queue.enqueue( + transcribe_full_audio_job, + conversation_id, + version_id, + trigger, + job_timeout=get_transcription_job_timeout(), + result_ttl=JOB_RESULT_TTL, + job_id=f"{job_id_prefix}_{conversation_id[:12]}", + description=f"Transcribe {title.lower()} {conversation_id[:8]}", + meta={"conversation_id": conversation_id, "client_id": client_id}, + ) + + application_logger.info( + f"šŸ“„ Batch: Enqueued transcription job {transcription_job.id}" + ) + + # Optionally chain post-conversation jobs + if enqueue_post_jobs: + from advanced_omi_backend.controllers.queue_controller import ( + start_post_conversation_jobs, ) + job_ids = start_post_conversation_jobs( + conversation_id=conversation_id, + user_id=None, + depends_on_job=transcription_job, + client_id=client_id, + ) application_logger.info( - f"āœ… Rolling batch #{batch_number} created conversation {conversation_id}, " - f"enqueued transcription job {transcription_job.id}" + f"āœ… Batch: Enqueued job chain for {conversation_id} ā€” " + f"transcription ({transcription_job.id}) ā†’ " + f"speaker ({job_ids['speaker_recognition']}) ā†’ " + f"memory ({job_ids['memory']})" ) - except Exception as e: - application_logger.error( - f"āŒ Failed to process rolling batch #{batch_number}: {e}", - exc_info=True - ) + return conversation_id -async def _process_batch_audio_complete( - client_state, - user_id: str, - user_email: str, - client_id: str +async def _process_rolling_batch( + client_state, user_id: str, user_email: str, client_id: str, batch_number: int ) -> None: - """ - Process completed batch audio: write file, create conversation, enqueue jobs. - - Args: - client_state: Client state with batch_audio_chunks - user_id: User ID - user_email: User email - client_id: Client ID - """ - if not hasattr(client_state, 'batch_audio_chunks') or not client_state.batch_audio_chunks: - application_logger.warning(f"āš ļø Batch mode: No audio chunks accumulated for {client_id}") + """Process accumulated batch audio as a rolling segment.""" + if ( + not hasattr(client_state, "batch_audio_chunks") + or not client_state.batch_audio_chunks + ): + application_logger.warning(f"āš ļø No audio chunks to process for rolling batch") return try: - from advanced_omi_backend.models.conversation import create_conversation - from advanced_omi_backend.utils.audio_chunk_utils import convert_audio_to_chunks - - # Combine all chunks - complete_audio = b''.join(client_state.batch_audio_chunks) - application_logger.info( - f"šŸ“¦ Batch mode: Combined {len(client_state.batch_audio_chunks)} chunks into {len(complete_audio)} bytes" + await _create_batch_conversation_and_enqueue( + client_state, + user_id=user_id, + client_id=client_id, + title=f"Recording Part {batch_number}", + trigger=f"rolling_batch_{batch_number}", + job_id_prefix=f"transcribe_rolling_{batch_number}", + ) + except Exception as e: + application_logger.error( + f"āŒ Failed to process rolling batch #{batch_number}: {e}", exc_info=True ) - # Timestamp for logging - timestamp = int(time.time() * 1000) - - # Get audio format from batch metadata (set during audio-start) - audio_format = getattr(client_state, 'batch_audio_format', {}) - sample_rate = audio_format.get('rate', OMI_SAMPLE_RATE) - sample_width = audio_format.get('width', OMI_SAMPLE_WIDTH) - channels = audio_format.get('channels', OMI_CHANNELS) - - # Calculate audio duration - duration = len(complete_audio) / (sample_rate * sample_width * channels) - application_logger.info( - f"āœ… Batch mode: Processing audio ({duration:.1f}s)" +async def _process_batch_audio_complete( + client_state, user_id: str, user_email: str, client_id: str +) -> None: + """Process completed batch audio: create conversation, enqueue full job chain.""" + if ( + not hasattr(client_state, "batch_audio_chunks") + or not client_state.batch_audio_chunks + ): + application_logger.warning( + f"āš ļø Batch mode: No audio chunks accumulated for {client_id}" ) + return - # Create conversation immediately for batch audio (conversation_id auto-generated) - version_id = str(uuid.uuid4()) - - conversation = create_conversation( + try: + await _create_batch_conversation_and_enqueue( + client_state, user_id=user_id, client_id=client_id, title="Batch Recording", - summary="Processing batch audio..." + trigger="batch", + job_id_prefix="transcribe", + enqueue_post_jobs=True, + attach_markers=True, + ) + client_state.batch_audio_chunks = [] + except Exception as batch_error: + application_logger.error( + f"āŒ Batch mode processing failed: {batch_error}", exc_info=True ) - # Attach any markers (e.g., button events) captured during the session - if client_state.markers: - conversation.markers = list(client_state.markers) - client_state.markers.clear() - await conversation.insert() - conversation_id = conversation.conversation_id # Get the auto-generated ID - application_logger.info(f"šŸ“ Batch mode: Created conversation {conversation_id}") - # Convert audio directly to MongoDB chunks (no disk intermediary) +async def _cleanup_websocket_connection( + client_id: Optional[str], + pending_client_id: str, + interim_subscriber_task: Optional[asyncio.Task], +) -> None: + """ + Shared cleanup for WebSocket handlers (OMI and PCM). + + Cancels the interim results subscriber, removes the pending connection + tracking entry, and tears down client state. + + Args: + client_id: Actual client ID (may be None if auth failed) + pending_client_id: Temporary tracking ID to discard + interim_subscriber_task: Background task forwarding interim transcripts + """ + # Cancel interim results subscriber task if running + if interim_subscriber_task and not interim_subscriber_task.done(): + interim_subscriber_task.cancel() try: - num_chunks = await convert_audio_to_chunks( - conversation_id=conversation_id, - audio_data=complete_audio, - sample_rate=sample_rate, - channels=channels, - sample_width=sample_width, - ) + await interim_subscriber_task + except asyncio.CancelledError: application_logger.info( - f"šŸ“¦ Batch mode: Converted to {num_chunks} MongoDB chunks " - f"(conversation {conversation_id[:12]})" + f"Interim subscriber task cancelled for {client_id}" ) - except Exception as chunk_error: + except Exception as task_error: application_logger.error( - f"Failed to convert batch audio to chunks: {chunk_error}", - exc_info=True + f"Error cancelling interim subscriber task: {task_error}" ) - # Continue anyway - transcription job will handle it - # Enqueue batch transcription job first (file uploads need transcription) - from advanced_omi_backend.controllers.queue_controller import ( - JOB_RESULT_TTL, - start_post_conversation_jobs, - transcription_queue, - ) - from advanced_omi_backend.workers.transcription_jobs import ( - transcribe_full_audio_job, - ) + # Clean up pending connection tracking + pending_connections.discard(pending_client_id) - version_id = str(uuid.uuid4()) - transcribe_job_id = f"transcribe_{conversation_id[:12]}" - - from advanced_omi_backend.config import get_transcription_job_timeout - - transcription_job = transcription_queue.enqueue( - transcribe_full_audio_job, - conversation_id, - version_id, - "batch", # trigger - job_timeout=get_transcription_job_timeout(), - result_ttl=JOB_RESULT_TTL, - job_id=transcribe_job_id, - description=f"Transcribe batch audio {conversation_id[:8]}", - meta={'conversation_id': conversation_id, 'client_id': client_id} - ) + # Ensure cleanup happens even if client_id is None + if client_id: + try: + await cleanup_client_state(client_id) + except Exception as cleanup_error: + application_logger.error( + f"Error during cleanup for client {client_id}: {cleanup_error}", + exc_info=True, + ) - application_logger.info(f"šŸ“„ Batch mode: Enqueued transcription job {transcription_job.id}") - # Enqueue post-conversation processing job chain (depends on transcription) - job_ids = start_post_conversation_jobs( - conversation_id=conversation_id, - user_id=None, # Will be read from conversation in DB by jobs - depends_on_job=transcription_job, # Wait for transcription to complete - client_id=client_id # Pass client_id for UI tracking - ) +from contextlib import asynccontextmanager - application_logger.info( - f"āœ… Batch mode: Enqueued job chain for {conversation_id} - " - f"transcription ({transcription_job.id}) ā†’ " - f"speaker ({job_ids['speaker_recognition']}) ā†’ " - f"memory ({job_ids['memory']})" + +@asynccontextmanager +async def _websocket_session(ws, token, device_name, connection_type): + """Lifecycle wrapper: pending tracking, auth, client setup, cleanup. + + Yields (client_id, client_state, user, audio_stream_producer, interim_holder) + on success, or None if auth failed. + interim_holder is a mutable list ā€” the inner loop sets interim_holder[0] = task. + """ + pending_client_id = f"pending_{uuid.uuid4()}" + pending_connections.add(pending_client_id) + + client_id = None + interim_holder = [None] # mutable so inner loop can update + + try: + client_id, client_state, user = await _setup_websocket_connection( + ws, token, device_name, pending_client_id, connection_type ) + if not user: + yield None + return - # Clear accumulated chunks - client_state.batch_audio_chunks = [] + # Store user context on client state up front (shared by all handlers) + client_state.user_id = user.user_id + client_state.user_email = user.email + client_state.client_id = client_id - except Exception as batch_error: + audio_stream_producer = get_audio_stream_producer() + + yield (client_id, client_state, user, audio_stream_producer, interim_holder) + + except WebSocketDisconnect: + application_logger.info( + f"šŸ”Œ {connection_type} WebSocket disconnected ā€” Client: {client_id}" + ) + except Exception as e: application_logger.error( - f"āŒ Batch mode processing failed: {batch_error}", - exc_info=True + f"āŒ {connection_type} WebSocket error for client {client_id}: {e}", + exc_info=True, + ) + finally: + await _cleanup_websocket_connection( + client_id, pending_client_id, interim_holder[0] ) @@ -1262,29 +1346,15 @@ async def handle_omi_websocket( device_name: Optional[str] = None, ): """Handle OMI WebSocket connections with Opus decoding.""" - # Generate pending client_id to track connection even if auth fails - pending_client_id = f"pending_{uuid.uuid4()}" - pending_connections.add(pending_client_id) - - client_id = None - client_state = None - interim_subscriber_task = None - - try: - # Setup connection (accept, auth, create client state) - client_id, client_state, user = await _setup_websocket_connection( - ws, token, device_name, pending_client_id, "OMI" - ) - if not user: + async with _websocket_session(ws, token, device_name, "OMI") as session: + if session is None: return + client_id, client_state, user, audio_stream_producer, interim_holder = session # OMI-specific: Setup Opus decoder decoder = OmiOpusDecoder() _decode_packet = partial(decoder.decode_packet, strip_header=False) - # Get singleton audio stream producer - audio_stream_producer = get_audio_stream_producer() - packet_count = 0 total_bytes = 0 @@ -1293,36 +1363,37 @@ async def handle_omi_websocket( header, payload = await parse_wyoming_protocol(ws) if header["type"] == "audio-start": - # Handle audio session start - application_logger.info(f"šŸ”“ BACKEND: Received audio-start in OMI MODE for {client_id} (header={header})") + application_logger.info( + f"šŸ”“ BACKEND: Received audio-start in OMI MODE for {client_id} (header={header})" + ) application_logger.info(f"šŸŽ™ļø OMI audio session started for {client_id}") - # Store user context on client state - client_state.user_id = user.user_id - client_state.user_email = user.email - client_state.client_id = client_id - - interim_subscriber_task = await _initialize_streaming_session( + interim_holder[0] = await _initialize_streaming_session( client_state, audio_stream_producer, user.user_id, user.email, client_id, - header.get("data", {"rate": OMI_SAMPLE_RATE, "width": OMI_SAMPLE_WIDTH, "channels": OMI_CHANNELS}), - websocket=ws # Pass WebSocket to launch interim results subscriber + header.get( + "data", + { + "rate": OMI_SAMPLE_RATE, + "width": OMI_SAMPLE_WIDTH, + "channels": OMI_CHANNELS, + }, + ), + websocket=ws, ) elif header["type"] == "audio-chunk" and payload: packet_count += 1 total_bytes += len(payload) - # Log progress if packet_count <= 5 or packet_count % 1000 == 0: application_logger.info( f"šŸŽµ Received OMI audio chunk #{packet_count}: {len(payload)} bytes" ) - # Handle OMI audio chunk (Opus decode + publish to stream) await _handle_omi_audio_chunk( client_state, audio_stream_producer, @@ -1330,32 +1401,28 @@ async def handle_omi_websocket( _decode_packet, user.user_id, client_id, - packet_count + packet_count, ) - # Log progress every 1000th packet if packet_count % 1000 == 0: application_logger.info( f"šŸ“Š Processed {packet_count} OMI packets ({total_bytes} bytes total)" ) elif header["type"] == "audio-stop": - # Handle audio session stop application_logger.info( f"šŸ›‘ OMI audio session stopped for {client_id} - " f"Total chunks: {packet_count}, Total bytes: {total_bytes}" ) - # Finalize session using helper function await _finalize_streaming_session( client_state, audio_stream_producer, user.user_id, user.email, - client_id + client_id, ) - # Reset counters for next session packet_count = 0 total_bytes = 0 @@ -1367,114 +1434,75 @@ async def handle_omi_websocket( ) else: - # Unknown event type application_logger.debug( f"Ignoring Wyoming event type '{header['type']}' for OMI client {client_id}" ) - except WebSocketDisconnect: - application_logger.info( - f"šŸ”Œ WebSocket disconnected - Client: {client_id}, Packets: {packet_count}, Total bytes: {total_bytes}" - ) - except Exception as e: - application_logger.error(f"āŒ WebSocket error for client {client_id}: {e}", exc_info=True) - finally: - # Cancel interim results subscriber task if running - if interim_subscriber_task and not interim_subscriber_task.done(): - interim_subscriber_task.cancel() - try: - await interim_subscriber_task - except asyncio.CancelledError: - application_logger.info(f"Interim subscriber task cancelled for {client_id}") - except Exception as task_error: - application_logger.error(f"Error cancelling interim subscriber task: {task_error}") - - # Clean up pending connection tracking - pending_connections.discard(pending_client_id) - - # Ensure cleanup happens even if client_id is None - if client_id: - try: - # Clean up client state - await cleanup_client_state(client_id) - except Exception as cleanup_error: - application_logger.error( - f"Error during cleanup for client {client_id}: {cleanup_error}", exc_info=True - ) - async def handle_pcm_websocket( - ws: WebSocket, - token: Optional[str] = None, - device_name: Optional[str] = None + ws: WebSocket, token: Optional[str] = None, device_name: Optional[str] = None ): """Handle PCM WebSocket connections with batch and streaming mode support.""" - # Generate pending client_id to track connection even if auth fails - pending_client_id = f"pending_{uuid.uuid4()}" - pending_connections.add(pending_client_id) - - client_id = None - client_state = None - interim_subscriber_task = None - - try: - # Setup connection (accept, auth, create client state) - client_id, client_state, user = await _setup_websocket_connection( - ws, token, device_name, pending_client_id, "PCM" - ) - if not user: + async with _websocket_session(ws, token, device_name, "PCM") as session: + if session is None: return - - # Get singleton audio stream producer - audio_stream_producer = get_audio_stream_producer() + client_id, client_state, user, audio_stream_producer, interim_holder = session packet_count = 0 total_bytes = 0 - audio_streaming = False # Track if audio session is active + audio_streaming = False while True: try: if not audio_streaming: # Control message mode - parse Wyoming protocol - application_logger.debug(f"šŸ”„ Control mode for {client_id}, WebSocket state: {ws.client_state if hasattr(ws, 'client_state') else 'unknown'}") - application_logger.debug(f"šŸ“Ø About to receive control message for {client_id}") + application_logger.debug( + f"šŸ”„ Control mode for {client_id}, WebSocket state: {ws.client_state if hasattr(ws, 'client_state') else 'unknown'}" + ) + application_logger.debug( + f"šŸ“Ø About to receive control message for {client_id}" + ) header, payload = await parse_wyoming_protocol(ws) - application_logger.debug(f"āœ… Received message type: {header.get('type')} for {client_id}") + application_logger.debug( + f"āœ… Received message type: {header.get('type')} for {client_id}" + ) if header["type"] == "audio-start": - application_logger.info(f"šŸ”“ BACKEND: Received audio-start in CONTROL MODE for {client_id}") - application_logger.debug(f"šŸŽ™ļø Processing audio-start for {client_id}") - - # Store user context on client state for rolling batch processing - client_state.user_id = user.user_id - client_state.user_email = user.email - client_state.client_id = client_id - - # Handle audio session start using helper function (pass websocket for error handling) - audio_streaming, recording_mode = await _handle_audio_session_start( - client_state, - header.get("data", {}), - client_id, - websocket=ws # Pass websocket for WebUI error display + application_logger.info( + f"šŸ”“ BACKEND: Received audio-start in CONTROL MODE for {client_id}" + ) + application_logger.debug( + f"šŸŽ™ļø Processing audio-start for {client_id}" + ) + + # Handle audio session start (pass websocket for error handling) + audio_streaming, recording_mode = ( + await _handle_audio_session_start( + client_state, + header.get("data", {}), + client_id, + websocket=ws, + ) ) # Initialize streaming session if recording_mode == "streaming": - application_logger.info(f"šŸ”“ BACKEND: Initializing streaming session for {client_id}") - interim_subscriber_task = await _initialize_streaming_session( + application_logger.info( + f"šŸ”“ BACKEND: Initializing streaming session for {client_id}" + ) + interim_holder[0] = await _initialize_streaming_session( client_state, audio_stream_producer, user.user_id, user.email, client_id, header.get("data", {}), - websocket=ws + websocket=ws, ) - continue # Continue to audio streaming mode - + continue + elif header["type"] == "ping": - # Handle keepalive ping from frontend application_logger.debug(f"šŸ“ Received ping from {client_id}") continue @@ -1487,67 +1515,73 @@ async def handle_pcm_websocket( continue else: - # Unknown control message type application_logger.debug( f"Ignoring Wyoming control event type '{header['type']}' for {client_id}" ) continue - + else: - # Audio streaming mode - receive raw bytes (like speaker recognition) - application_logger.debug(f"šŸŽµ Audio streaming mode for {client_id} - waiting for audio data") - + # Audio streaming mode + application_logger.debug( + f"šŸŽµ Audio streaming mode for {client_id} - waiting for audio data" + ) + try: - # Receive raw audio bytes or check for control messages message = await ws.receive() - - - # Check if it's a disconnect - if "type" in message and message["type"] == "websocket.disconnect": + + if ( + "type" in message + and message["type"] == "websocket.disconnect" + ): code = message.get("code", 1000) reason = message.get("reason", "") - application_logger.info(f"šŸ”Œ WebSocket disconnect during audio streaming for {client_id}. Code: {code}, Reason: {reason}") + application_logger.info( + f"šŸ”Œ WebSocket disconnect during audio streaming for {client_id}. Code: {code}, Reason: {reason}" + ) break - - # Check if it's a text message (control message like audio-stop) + if "text" in message: try: control_header = json.loads(message["text"].strip()) if control_header.get("type") == "audio-stop": - # Handle audio session stop using helper function audio_streaming = await _handle_audio_session_stop( client_state, audio_stream_producer, user.user_id, user.email, - client_id + client_id, ) - # Reset counters for next session packet_count = 0 total_bytes = 0 continue elif control_header.get("type") == "ping": - application_logger.debug(f"šŸ“ Received ping during streaming from {client_id}") + application_logger.debug( + f"šŸ“ Received ping during streaming from {client_id}" + ) continue elif control_header.get("type") == "audio-start": - # Handle duplicate audio-start messages gracefully (idempotent behavior) - application_logger.info(f"šŸ”„ Ignoring duplicate audio-start message during streaming for {client_id}") + application_logger.info( + f"šŸ”„ Ignoring duplicate audio-start message during streaming for {client_id}" + ) continue elif control_header.get("type") == "audio-chunk": - # Handle Wyoming protocol audio-chunk with binary payload - payload_length = control_header.get("payload_length") + payload_length = control_header.get( + "payload_length" + ) if payload_length and payload_length > 0: - # Receive the binary audio data payload_msg = await ws.receive() if "bytes" in payload_msg: audio_data = payload_msg["bytes"] packet_count += 1 total_bytes += len(audio_data) - application_logger.debug(f"šŸŽµ Received audio chunk #{packet_count}: {len(audio_data)} bytes") + application_logger.debug( + f"šŸŽµ Received audio chunk #{packet_count}: {len(audio_data)} bytes" + ) - # Route to appropriate mode handler - audio_format = control_header.get("data", {}) + audio_format = control_header.get( + "data", {} + ) task = await _handle_audio_chunk( client_state, audio_stream_producer, @@ -1556,41 +1590,50 @@ async def handle_pcm_websocket( user.user_id, user.email, client_id, - websocket=ws + websocket=ws, ) - # Store subscriber task if it was created (first streaming chunk) - if task and not interim_subscriber_task: - interim_subscriber_task = task + if task and not interim_holder[0]: + interim_holder[0] = task else: - application_logger.warning(f"Expected binary payload for audio-chunk, got: {payload_msg.keys()}") + application_logger.warning( + f"Expected binary payload for audio-chunk, got: {payload_msg.keys()}" + ) else: - application_logger.warning(f"audio-chunk missing payload_length: {payload_length}") + application_logger.warning( + f"audio-chunk missing payload_length: {payload_length}" + ) continue elif control_header.get("type") == "button-event": button_data = control_header.get("data", {}) button_state = button_data.get("state", "unknown") await _handle_button_event( - client_state, button_state, user.user_id, client_id + client_state, + button_state, + user.user_id, + client_id, ) continue else: - application_logger.warning(f"Unknown control message during streaming: {control_header.get('type')}") + application_logger.warning( + f"Unknown control message during streaming: {control_header.get('type')}" + ) continue except json.JSONDecodeError: - application_logger.warning(f"Invalid control message during streaming for {client_id}") + application_logger.warning( + f"Invalid control message during streaming for {client_id}" + ) continue - - # Check if it's binary data (raw audio without Wyoming protocol) + elif "bytes" in message: - # Raw binary audio data (legacy support) audio_data = message["bytes"] packet_count += 1 total_bytes += len(audio_data) - application_logger.debug(f"šŸŽµ Received raw audio chunk #{packet_count}: {len(audio_data)} bytes") + application_logger.debug( + f"šŸŽµ Received raw audio chunk #{packet_count}: {len(audio_data)} bytes" + ) - # Route to appropriate mode handler with default format default_format = {"rate": 16000, "width": 2, "channels": 1} task = await _handle_audio_chunk( client_state, @@ -1600,18 +1643,21 @@ async def handle_pcm_websocket( user.user_id, user.email, client_id, - websocket=ws + websocket=ws, ) - # Store subscriber task if it was created (first streaming chunk) - if task and not interim_subscriber_task: - interim_subscriber_task = task - + if task and not interim_holder[0]: + interim_holder[0] = task + else: - application_logger.warning(f"Unexpected message format in streaming mode: {message.keys()}") + application_logger.warning( + f"Unexpected message format in streaming mode: {message.keys()}" + ) continue - + except Exception as streaming_error: - application_logger.error(f"Error in audio streaming mode: {streaming_error}") + application_logger.error( + f"Error in audio streaming mode: {streaming_error}" + ) if "disconnect" in str(streaming_error).lower(): break continue @@ -1621,24 +1667,21 @@ async def handle_pcm_websocket( f"šŸ”Œ WebSocket disconnected during message processing for {client_id}. " f"Code: {e.code}, Reason: {e.reason}" ) - break # Exit the loop on disconnect + break except json.JSONDecodeError as e: application_logger.error( f"āŒ JSON decode error in Wyoming protocol for {client_id}: {e}" ) - continue # Skip this message but don't disconnect + continue except ValueError as e: - application_logger.error( - f"āŒ Protocol error for {client_id}: {e}" - ) - continue # Skip this message but don't disconnect + application_logger.error(f"āŒ Protocol error for {client_id}: {e}") + continue except RuntimeError as e: - # Handle "Cannot call receive once a disconnect message has been received" if "disconnect" in str(e).lower(): application_logger.info( f"šŸ”Œ WebSocket already disconnected for {client_id}: {e}" ) - break # Exit the loop on disconnect + break else: application_logger.error( f"āŒ Runtime error for {client_id}: {e}", exc_info=True @@ -1646,46 +1689,18 @@ async def handle_pcm_websocket( continue except Exception as e: application_logger.error( - f"āŒ Unexpected error processing message for {client_id}: {e}", exc_info=True + f"āŒ Unexpected error processing message for {client_id}: {e}", + exc_info=True, ) - # Check if it's a connection-related error error_msg = str(e).lower() - if "disconnect" in error_msg or "closed" in error_msg or "receive" in error_msg: + if ( + "disconnect" in error_msg + or "closed" in error_msg + or "receive" in error_msg + ): application_logger.info( f"šŸ”Œ Connection issue detected for {client_id}, exiting loop" ) break else: - continue # Skip this message for other errors - - except WebSocketDisconnect: - application_logger.info( - f"šŸ”Œ PCM WebSocket disconnected - Client: {client_id}, Packets: {packet_count}, Total bytes: {total_bytes}" - ) - except Exception as e: - application_logger.error( - f"āŒ PCM WebSocket error for client {client_id}: {e}", exc_info=True - ) - finally: - # Cancel interim results subscriber task if running - if interim_subscriber_task and not interim_subscriber_task.done(): - interim_subscriber_task.cancel() - try: - await interim_subscriber_task - except asyncio.CancelledError: - application_logger.info(f"Interim subscriber task cancelled for {client_id}") - except Exception as task_error: - application_logger.error(f"Error cancelling interim subscriber task: {task_error}") - - # Clean up pending connection tracking - pending_connections.discard(pending_client_id) - - # Ensure cleanup happens even if client_id is None - if client_id: - try: - # Clean up client state - await cleanup_client_state(client_id) - except Exception as cleanup_error: - application_logger.error( - f"Error during cleanup for client {client_id}: {cleanup_error}", exc_info=True - ) + continue diff --git a/backends/advanced/src/advanced_omi_backend/observability/__init__.py b/backends/advanced/src/advanced_omi_backend/observability/__init__.py new file mode 100644 index 00000000..e69de29b diff --git a/backends/advanced/src/advanced_omi_backend/observability/otel_setup.py b/backends/advanced/src/advanced_omi_backend/observability/otel_setup.py new file mode 100644 index 00000000..9c891aeb --- /dev/null +++ b/backends/advanced/src/advanced_omi_backend/observability/otel_setup.py @@ -0,0 +1,129 @@ +"""OpenTelemetry setup and session management. + +Uses OpenInference semantic conventions (session.id) so that any +compatible observability backend (Galileo, Arize Phoenix, Langfuse, etc.) +can group traces by session. +""" + +import contextvars +import logging +import os +from functools import lru_cache + +logger = logging.getLogger(__name__) + +_otel_initialised = False + +# Per-task/thread token so concurrent conversations don't clobber each other. +_session_token_var: contextvars.ContextVar[object | None] = contextvars.ContextVar( + "_otel_session_token", default=None +) + + +@lru_cache(maxsize=1) +def is_galileo_enabled() -> bool: + """Check if Galileo OTEL is configured.""" + return bool(os.getenv("GALILEO_API_KEY")) + + +def is_otel_enabled() -> bool: + """Check if any OTel exporter has been initialised.""" + return _otel_initialised + + +def set_otel_session(session_id: str) -> None: + """Attach *session_id* to the OTel context (OpenInference ``session.id``). + + All subsequent spans on this thread/context will carry the session ID, + regardless of which observability backend is consuming them. + Safe to call concurrently from different asyncio tasks or threads. + """ + if not is_otel_enabled(): + return + try: + from openinference.semconv.trace import SpanAttributes + from opentelemetry.context import attach, get_current, set_value + + clear_otel_session() + ctx = set_value(SpanAttributes.SESSION_ID, session_id, get_current()) + _session_token_var.set(attach(ctx)) + except ImportError: + pass + + +def clear_otel_session() -> None: + """Detach the current session from the OTel context.""" + token = _session_token_var.get() + if token is None: + return + try: + from opentelemetry.context import detach + + detach(token) + _session_token_var.set(None) + except Exception: + _session_token_var.set(None) + + +def init_otel() -> None: + """Initialize OTEL with Galileo exporter and OpenAI instrumentor. + + Call once at app startup. Safe to call if Galileo is not configured (no-op). + Filters out embedding spans ā€” only LLM (chat completion) calls are exported. + """ + if not is_galileo_enabled(): + logger.info("Galileo not configured, skipping OTEL initialization") + return + + try: + from galileo import otel + from openinference.instrumentation.openai import OpenAIInstrumentor + from opentelemetry import context + from opentelemetry.sdk import trace as trace_sdk + from opentelemetry.sdk.trace import ReadableSpan, Span, SpanProcessor + + project = os.getenv("GALILEO_PROJECT", "chronicle") + logstream = os.getenv("GALILEO_LOG_STREAM", "default") + + class _LLMOnlyProcessor(SpanProcessor): + """Wraps GalileoSpanProcessor, dropping EMBEDDING spans.""" + + def __init__(self, inner: SpanProcessor): + self._inner = inner + + def on_start( + self, span: Span, parent_context: context.Context | None = None + ) -> None: + self._inner.on_start(span, parent_context) + + def on_end(self, span: ReadableSpan) -> None: + kind = span.attributes.get("openinference.span.kind", "") + if kind == "EMBEDDING": + return # drop + self._inner.on_end(span) + + def shutdown(self) -> None: + self._inner.shutdown() + + def force_flush(self, timeout_millis: int = 30000) -> bool: + return self._inner.force_flush(timeout_millis) + + tracer_provider = trace_sdk.TracerProvider() + galileo_processor = otel.GalileoSpanProcessor( + project=project, logstream=logstream + ) + tracer_provider.add_span_processor(_LLMOnlyProcessor(galileo_processor)) + + # Auto-instrument all OpenAI SDK calls + OpenAIInstrumentor().instrument(tracer_provider=tracer_provider) + + global _otel_initialised + _otel_initialised = True + logger.info("OTEL initialized with Galileo exporter + OpenAI instrumentor") + except ImportError: + logger.warning( + "Galileo/OTEL packages not installed. " + "Install with: uv pip install '.[galileo]'" + ) + except Exception as e: + logger.error(f"Failed to initialize OTEL: {e}") diff --git a/backends/advanced/src/advanced_omi_backend/plugins/base.py b/backends/advanced/src/advanced_omi_backend/plugins/base.py index 5c9b668d..92e926f6 100644 --- a/backends/advanced/src/advanced_omi_backend/plugins/base.py +++ b/backends/advanced/src/advanced_omi_backend/plugins/base.py @@ -6,6 +6,7 @@ - PluginResult: Result from plugin execution - BasePlugin: Abstract base class for all plugins """ + from abc import ABC, abstractmethod from dataclasses import dataclass, field from typing import Any, Dict, List, Optional @@ -14,16 +15,20 @@ @dataclass class PluginContext: """Context passed to plugin execution""" + user_id: str event: str # Event name (e.g., "transcript.streaming", "conversation.complete") data: Dict[str, Any] # Event-specific data metadata: Dict[str, Any] = field(default_factory=dict) - services: Optional[Any] = None # PluginServices instance for system/cross-plugin calls + services: Optional[Any] = ( + None # PluginServices instance for system/cross-plugin calls + ) @dataclass class PluginResult: """Result from plugin execution""" + success: bool data: Optional[Dict[str, Any]] = None message: Optional[str] = None @@ -58,9 +63,9 @@ def __init__(self, config: Dict[str, Any]): Contains: enabled, events, condition, and plugin-specific config """ self.config = config - self.enabled = config.get('enabled', False) - self.events = config.get('events', []) - self.condition = config.get('condition', {'type': 'always'}) + self.enabled = config.get("enabled", False) + self.events = config.get("events", []) + self.condition = config.get("condition", {"type": "always"}) def register_prompts(self, registry) -> None: """Register plugin prompts with the prompt registry. @@ -122,7 +127,9 @@ async def on_transcript(self, context: PluginContext) -> Optional[PluginResult]: """ pass - async def on_conversation_complete(self, context: PluginContext) -> Optional[PluginResult]: + async def on_conversation_complete( + self, context: PluginContext + ) -> Optional[PluginResult]: """ Called when conversation processing completes. @@ -137,7 +144,9 @@ async def on_conversation_complete(self, context: PluginContext) -> Optional[Plu """ pass - async def on_memory_processed(self, context: PluginContext) -> Optional[PluginResult]: + async def on_memory_processed( + self, context: PluginContext + ) -> Optional[PluginResult]: """ Called after memory extraction finishes. @@ -152,7 +161,9 @@ async def on_memory_processed(self, context: PluginContext) -> Optional[PluginRe """ pass - async def on_conversation_starred(self, context: PluginContext) -> Optional[PluginResult]: + async def on_conversation_starred( + self, context: PluginContext + ) -> Optional[PluginResult]: """ Called when a conversation is starred or unstarred. @@ -172,7 +183,7 @@ async def on_button_event(self, context: PluginContext) -> Optional[PluginResult Called when a device button event is received. Context data contains: - - state: str - Button state (e.g., "SINGLE_TAP", "DOUBLE_TAP", "LONG_PRESS") + - state: str - Button state (e.g., "SINGLE_PRESS", "DOUBLE_PRESS", "LONG_PRESS") - timestamp: float - Unix timestamp of the event - audio_uuid: str - Current audio session UUID (may be None) diff --git a/backends/advanced/src/advanced_omi_backend/plugins/events.py b/backends/advanced/src/advanced_omi_backend/plugins/events.py index 3d7ec284..7a732a04 100644 --- a/backends/advanced/src/advanced_omi_backend/plugins/events.py +++ b/backends/advanced/src/advanced_omi_backend/plugins/events.py @@ -24,32 +24,50 @@ def __new__(cls, value: str, description: str = ""): return obj # Conversation lifecycle - CONVERSATION_COMPLETE = ("conversation.complete", "Fires when conversation processing finishes (transcript ready)") - TRANSCRIPT_STREAMING = ("transcript.streaming", "Real-time transcript segments during a live conversation") - TRANSCRIPT_BATCH = ("transcript.batch", "Batch transcript from file upload processing") - MEMORY_PROCESSED = ("memory.processed", "After memories are extracted from a conversation") - CONVERSATION_STARRED = ("conversation.starred", "Fires when a conversation is starred or unstarred") + CONVERSATION_COMPLETE = ( + "conversation.complete", + "Fires when conversation processing finishes (transcript ready)", + ) + TRANSCRIPT_STREAMING = ( + "transcript.streaming", + "Real-time transcript segments during a live conversation", + ) + TRANSCRIPT_BATCH = ( + "transcript.batch", + "Batch transcript from file upload processing", + ) + MEMORY_PROCESSED = ( + "memory.processed", + "After memories are extracted from a conversation", + ) + CONVERSATION_STARRED = ( + "conversation.starred", + "Fires when a conversation is starred or unstarred", + ) # Button events (from OMI device) BUTTON_SINGLE_PRESS = ("button.single_press", "OMI device button single press") BUTTON_DOUBLE_PRESS = ("button.double_press", "OMI device button double press") # Cross-plugin communication (dispatched by PluginServices.call_plugin) - PLUGIN_ACTION = ("plugin_action", "Cross-plugin dispatch via PluginServices.call_plugin()") + PLUGIN_ACTION = ( + "plugin_action", + "Cross-plugin dispatch via PluginServices.call_plugin()", + ) class ButtonState(str, Enum): """Raw button states from OMI device firmware.""" - SINGLE_TAP = "SINGLE_TAP" - DOUBLE_TAP = "DOUBLE_TAP" + SINGLE_PRESS = "SINGLE_PRESS" + DOUBLE_PRESS = "DOUBLE_PRESS" LONG_PRESS = "LONG_PRESS" # Maps device button states to plugin events BUTTON_STATE_TO_EVENT: Dict[ButtonState, PluginEvent] = { - ButtonState.SINGLE_TAP: PluginEvent.BUTTON_SINGLE_PRESS, - ButtonState.DOUBLE_TAP: PluginEvent.BUTTON_DOUBLE_PRESS, + ButtonState.SINGLE_PRESS: PluginEvent.BUTTON_SINGLE_PRESS, + ButtonState.DOUBLE_PRESS: PluginEvent.BUTTON_DOUBLE_PRESS, } diff --git a/backends/advanced/src/advanced_omi_backend/plugins/router.py b/backends/advanced/src/advanced_omi_backend/plugins/router.py index c70ad73b..e06d3043 100644 --- a/backends/advanced/src/advanced_omi_backend/plugins/router.py +++ b/backends/advanced/src/advanced_omi_backend/plugins/router.py @@ -37,13 +37,59 @@ def normalize_text_for_wake_word(text: str) -> str: # Lowercase text = text.lower() # Replace punctuation with spaces (instead of removing, to preserve word boundaries) - text = text.translate(str.maketrans(string.punctuation, ' ' * len(string.punctuation))) + text = text.translate( + str.maketrans(string.punctuation, " " * len(string.punctuation)) + ) # Normalize whitespace (collapse multiple spaces to single space) - text = re.sub(r'\s+', ' ', text) + text = re.sub(r"\s+", " ", text) # Strip leading/trailing whitespace return text.strip() +def extract_command_around_keyword(transcript: str, keyword: str) -> str: + """ + Extract command by removing a keyword from anywhere in the transcript. + + Handles punctuation and spacing around the keyword gracefully. + + Example: + transcript: "Turn off the lights, Vivi" + keyword: "vivi" + -> "Turn off the lights" + + transcript: "Vivi, turn off the lights in the hall" + keyword: "vivi" + -> "turn off the lights in the hall" + + transcript: "Turn off the hall lights, Vivi, please" + keyword: "vivi" + -> "Turn off the hall lights, please" + + Args: + transcript: Original transcript text + keyword: Keyword to remove (will be normalized) + + Returns: + Command text with keyword removed + """ + keyword_parts = normalize_text_for_wake_word(keyword).split() + if not keyword_parts: + return transcript.strip() + + pattern_parts = [re.escape(part) for part in keyword_parts] + # Match keyword with optional surrounding punctuation/whitespace + kw_pattern = r"[\s,.\-!?;:]*".join(pattern_parts) + # Consume adjacent punctuation/whitespace on both sides + full_pattern = r"[\s,.\-!?;:]*" + kw_pattern + r"[\s,.\-!?;:]*" + + command = re.sub( + full_pattern, " ", transcript, count=1, flags=re.IGNORECASE + ).strip() + # Collapse any doubled spaces left behind + command = re.sub(r"\s{2,}", " ", command) + return command + + def extract_command_after_wake_word(transcript: str, wake_word: str) -> str: """ Intelligently extract command after wake word in original transcript. @@ -73,25 +119,28 @@ def extract_command_after_wake_word(transcript: str, wake_word: str) -> str: # The pattern matches the wake word parts with optional punctuation/whitespace between and after pattern_parts = [re.escape(part) for part in wake_word_parts] # Allow optional punctuation/whitespace between parts - pattern = r'[\s,.\-!?;:]*'.join(pattern_parts) + pattern = r"[\s,.\-!?;:]*".join(pattern_parts) # Add trailing punctuation/whitespace consumption after last wake word part - pattern = '^' + pattern + r'[\s,.\-!?;:]*' + pattern = "^" + pattern + r"[\s,.\-!?;:]*" # Try to match wake word at start of transcript (case-insensitive) match = re.match(pattern, transcript, re.IGNORECASE) if match: # Extract everything after the matched wake word (including trailing punctuation) - command = transcript[match.end():].strip() + command = transcript[match.end() :].strip() return command else: # Fallback: couldn't find wake word boundary, return full transcript - logger.warning(f"Could not find wake word boundary for '{wake_word}' in '{transcript}', using full transcript") + logger.warning( + f"Could not find wake word boundary for '{wake_word}' in '{transcript}', using full transcript" + ) return transcript.strip() class ConditionResult(NamedTuple): """Result of a plugin condition check.""" + execute: bool extra: Dict[str, Any] = {} @@ -100,9 +149,9 @@ class PluginHealth: """Health status for a single plugin.""" # Possible status values - REGISTERED = "registered" # Registered but not yet initialized - INITIALIZED = "initialized" # Successfully initialized - FAILED = "failed" # initialize() raised an exception + REGISTERED = "registered" # Registered but not yet initialized + INITIALIZED = "initialized" # Successfully initialized + FAILED = "failed" # initialize() raised an exception def __init__(self, plugin_id: str): self.plugin_id = plugin_id @@ -182,11 +231,7 @@ def get_health_summary(self) -> Dict[str, Any]: } async def dispatch_event( - self, - event: str, - user_id: str, - data: Dict, - metadata: Optional[Dict] = None + self, event: str, user_id: str, data: Dict, metadata: Optional[Dict] = None ) -> List[PluginResult]: """ Dispatch event to all subscribed plugins. @@ -212,7 +257,9 @@ async def dispatch_event( if not plugin_ids: logger.info(f"šŸ”Œ ROUTER: No plugins subscribed to event '{event}'") else: - logger.info(f"šŸ”Œ ROUTER: Found {len(plugin_ids)} subscribed plugin(s): {plugin_ids}") + logger.info( + f"šŸ”Œ ROUTER: Found {len(plugin_ids)} subscribed plugin(s): {plugin_ids}" + ) for plugin_id in plugin_ids: plugin = self.plugins[plugin_id] @@ -251,22 +298,34 @@ async def dispatch_event( f"success={result.success}, message={result.message}" ) results.append(result) - executed.append({"plugin_id": plugin_id, "success": result.success, "message": result.message}) + executed.append( + { + "plugin_id": plugin_id, + "success": result.success, + "message": result.message, + } + ) # If plugin says stop processing, break if not result.should_continue: - logger.info(f" āŠ— Plugin '{plugin_id}' stopped further processing") + logger.info( + f" āŠ— Plugin '{plugin_id}' stopped further processing" + ) break else: - logger.info(f" āŠ˜ Plugin '{plugin_id}' returned no result for '{event}'") + logger.info( + f" āŠ˜ Plugin '{plugin_id}' returned no result for '{event}'" + ) except Exception as e: # CRITICAL: Log exception details logger.error( f" āœ— Plugin '{plugin_id}' FAILED with exception: {e}", - exc_info=True + exc_info=True, + ) + executed.append( + {"plugin_id": plugin_id, "success": False, "message": str(e)} ) - executed.append({"plugin_id": plugin_id, "success": False, "message": str(e)}) # Add at end logger.info( @@ -287,7 +346,9 @@ async def dispatch_event( _SKIP = ConditionResult(execute=False) _PASS = ConditionResult(execute=True) - async def _should_execute(self, plugin: BasePlugin, data: Dict, event: Optional[str] = None) -> ConditionResult: + async def _should_execute( + self, plugin: BasePlugin, data: Dict, event: Optional[str] = None + ) -> ConditionResult: """Check if plugin should be executed based on condition configuration. Returns a ConditionResult. The ``extra`` dict contains per-plugin data @@ -297,53 +358,85 @@ async def _should_execute(self, plugin: BasePlugin, data: Dict, event: Optional[ Button events bypass transcript-based conditions (wake_word) since they have no transcript to match against. """ - condition_type = plugin.condition.get('type', 'always') + condition_type = plugin.condition.get("type", "always") - if condition_type == 'always': + if condition_type == "always": return self._PASS # Button and starred events bypass transcript-based conditions (no transcript to match) - if event and event in (PluginEvent.BUTTON_SINGLE_PRESS, PluginEvent.BUTTON_DOUBLE_PRESS, PluginEvent.CONVERSATION_STARRED): + if event and event in ( + PluginEvent.BUTTON_SINGLE_PRESS, + PluginEvent.BUTTON_DOUBLE_PRESS, + PluginEvent.CONVERSATION_STARRED, + ): return self._PASS - elif condition_type == 'wake_word': + elif condition_type == "wake_word": # Normalize transcript for matching (handles punctuation and spacing) - transcript = data.get('transcript', '') + transcript = data.get("transcript", "") normalized_transcript = normalize_text_for_wake_word(transcript) # Support both singular 'wake_word' and plural 'wake_words' (list) - wake_words = plugin.condition.get('wake_words', []) + wake_words = plugin.condition.get("wake_words", []) if not wake_words: # Fallback to singular wake_word for backward compatibility - wake_word = plugin.condition.get('wake_word', '') + wake_word = plugin.condition.get("wake_word", "") if wake_word: wake_words = [wake_word] # Check if transcript starts with any wake word (after normalization) for wake_word in wake_words: normalized_wake_word = normalize_text_for_wake_word(wake_word) - if normalized_wake_word and normalized_transcript.startswith(normalized_wake_word): + if normalized_wake_word and normalized_transcript.startswith( + normalized_wake_word + ): # Smart extraction: find where wake word actually ends in original text command = extract_command_after_wake_word(transcript, wake_word) - logger.debug(f"Wake word '{wake_word}' detected. Original: '{transcript}', Command: '{command}'") + logger.debug( + f"Wake word '{wake_word}' detected. Original: '{transcript}', Command: '{command}'" + ) return ConditionResult( execute=True, - extra={'command': command, 'original_transcript': transcript}, + extra={"command": command, "original_transcript": transcript}, ) return self._SKIP - elif condition_type == 'conditional': + elif condition_type == "keyword_anywhere": + # Trigger when keyword appears anywhere in the transcript. + # Command is the transcript with the keyword removed. + transcript = data.get("transcript", "") + normalized_transcript = normalize_text_for_wake_word(transcript) + + keywords = plugin.condition.get("keywords", []) + if not keywords: + keyword = plugin.condition.get("keyword", "") + if keyword: + keywords = [keyword] + + for keyword in keywords: + normalized_keyword = normalize_text_for_wake_word(keyword) + if normalized_keyword and normalized_keyword in normalized_transcript: + command = extract_command_around_keyword(transcript, keyword) + logger.debug( + f"Keyword '{keyword}' found in transcript. " + f"Original: '{transcript}', Command: '{command}'" + ) + return ConditionResult( + execute=True, + extra={"command": command, "original_transcript": transcript}, + ) + + return self._SKIP + + elif condition_type == "conditional": # Future: Custom condition checking return self._PASS return self._SKIP async def _execute_plugin( - self, - plugin: BasePlugin, - event: str, - context: PluginContext + self, plugin: BasePlugin, event: str, context: PluginContext ) -> Optional[PluginResult]: """Execute plugin method for specified event""" # Map events to plugin callback methods using enums @@ -356,7 +449,10 @@ async def _execute_plugin( return await plugin.on_memory_processed(context) elif event == PluginEvent.CONVERSATION_STARRED: return await plugin.on_conversation_starred(context) - elif event in (PluginEvent.BUTTON_SINGLE_PRESS, PluginEvent.BUTTON_DOUBLE_PRESS): + elif event in ( + PluginEvent.BUTTON_SINGLE_PRESS, + PluginEvent.BUTTON_DOUBLE_PRESS, + ): return await plugin.on_button_event(context) elif event == PluginEvent.PLUGIN_ACTION: return await plugin.on_plugin_action(context) @@ -377,14 +473,16 @@ def _log_event( if not self._event_redis: return try: - record = json.dumps({ - "timestamp": time.time(), - "event": event, - "user_id": user_id, - "plugins_subscribed": plugins_subscribed, - "plugins_executed": plugins_executed, - "metadata": metadata or {}, - }) + record = json.dumps( + { + "timestamp": time.time(), + "event": event, + "user_id": user_id, + "plugins_subscribed": plugins_subscribed, + "plugins_executed": plugins_executed, + "metadata": metadata or {}, + } + ) pipe = self._event_redis.pipeline() pipe.lpush(self._EVENT_LOG_KEY, record) pipe.ltrim(self._EVENT_LOG_KEY, 0, self._EVENT_LOG_MAX - 1) @@ -404,8 +502,9 @@ def clear_events(self) -> int: logger.debug("Failed to clear events from Redis", exc_info=True) return 0 - - def get_recent_events(self, limit: int = 50, event_type: Optional[str] = None) -> List[Dict]: + def get_recent_events( + self, limit: int = 50, event_type: Optional[str] = None + ) -> List[Dict]: """Read recent events from the Redis log.""" if not self._event_redis: return [] @@ -439,9 +538,15 @@ async def check_connectivity(self) -> Dict[str, Dict[str, Any]]: result = await asyncio.wait_for(plugin.health_check(), timeout=10.0) results[plugin_id] = result except asyncio.TimeoutError: - results[plugin_id] = {"ok": False, "message": "Health check timed out (10s)"} + results[plugin_id] = { + "ok": False, + "message": "Health check timed out (10s)", + } except Exception as e: - results[plugin_id] = {"ok": False, "message": f"Health check error: {e}"} + results[plugin_id] = { + "ok": False, + "message": f"Health check error: {e}", + } return results diff --git a/backends/advanced/src/advanced_omi_backend/plugins/services.py b/backends/advanced/src/advanced_omi_backend/plugins/services.py index dbddfb21..5dc09a98 100644 --- a/backends/advanced/src/advanced_omi_backend/plugins/services.py +++ b/backends/advanced/src/advanced_omi_backend/plugins/services.py @@ -10,7 +10,6 @@ import redis.asyncio as aioredis - from .base import PluginContext, PluginResult from .events import ConversationCloseReason, PluginEvent @@ -44,18 +43,36 @@ async def close_conversation( Signals the open_conversation_job to close the current conversation and trigger post-processing. The session stays active for new conversations. + Only succeeds when open_conversation_job is actively running and polling + (indicated by the conversation:current:{session_id} Redis key). During + speech detection phase, no conversation is open ā€” the flag would go unread. + Args: session_id: The streaming session ID (typically same as client_id) reason: Why the conversation is being closed Returns: - True if the close request was set successfully + True if the close request was set successfully, False if no + conversation is currently open for this session """ + # Gate: only set the flag when open_conversation_job is running and will read it. + # The conversation:current key is set right before the polling loop starts. + conversation_id = await self._async_redis.get( + f"conversation:current:{session_id}" + ) + if not conversation_id: + logger.warning( + f"No open conversation for session {session_id} ā€” close request ignored" + ) + return False + from advanced_omi_backend.controllers.session_controller import ( request_conversation_close, ) - return await request_conversation_close(self._async_redis, session_id, reason=reason.value) + return await request_conversation_close( + self._async_redis, session_id, reason=reason.value + ) async def star_conversation(self, session_id: str) -> bool: """Toggle the star on the current conversation for a session. @@ -73,13 +90,17 @@ async def star_conversation(self, session_id: str) -> bool: from advanced_omi_backend.users import User # Look up current conversation_id from Redis - conversation_id = await self._async_redis.get(f"conversation:current:{session_id}") + conversation_id = await self._async_redis.get( + f"conversation:current:{session_id}" + ) if not conversation_id: logger.warning(f"No current conversation for session {session_id}") return False # Find conversation to get user_id - conversation = await Conversation.find_one(Conversation.conversation_id == conversation_id) + conversation = await Conversation.find_one( + Conversation.conversation_id == conversation_id + ) if not conversation: logger.warning(f"Conversation {conversation_id} not found for starring") return False @@ -115,10 +136,14 @@ async def call_plugin( plugin = self._router.plugins.get(plugin_id) if not plugin: logger.warning(f"Plugin '{plugin_id}' not found for cross-plugin call") - return PluginResult(success=False, message=f"Plugin '{plugin_id}' not found") + return PluginResult( + success=False, message=f"Plugin '{plugin_id}' not found" + ) if not plugin.enabled: logger.warning(f"Plugin '{plugin_id}' is disabled, cannot call") - return PluginResult(success=False, message=f"Plugin '{plugin_id}' is disabled") + return PluginResult( + success=False, message=f"Plugin '{plugin_id}' is disabled" + ) context = PluginContext( user_id=user_id, @@ -136,5 +161,7 @@ async def call_plugin( ) return result except Exception as e: - logger.error(f"Cross-plugin call to {plugin_id}.{action} failed: {e}", exc_info=True) + logger.error( + f"Cross-plugin call to {plugin_id}.{action} failed: {e}", exc_info=True + ) return PluginResult(success=False, message=f"Plugin action failed: {e}") diff --git a/backends/advanced/src/advanced_omi_backend/prompt_defaults.py b/backends/advanced/src/advanced_omi_backend/prompt_defaults.py index 75d89eee..94b03d31 100644 --- a/backends/advanced/src/advanced_omi_backend/prompt_defaults.py +++ b/backends/advanced/src/advanced_omi_backend/prompt_defaults.py @@ -524,7 +524,7 @@ def register_all_defaults(registry: PromptRegistry) -> None: # ------------------------------------------------------------------ registry.register_default( "asr.hot_words", - template="hey vivi, chronicle, omi", + template="vivi, chronicle, omi", name="ASR Hot Words", description="Comma-separated hot words for speech recognition. " "For Deepgram: boosts keyword recognition via keyterm. " diff --git a/backends/advanced/src/advanced_omi_backend/prompt_registry.py b/backends/advanced/src/advanced_omi_backend/prompt_registry.py index eae9c248..5aacb151 100644 --- a/backends/advanced/src/advanced_omi_backend/prompt_registry.py +++ b/backends/advanced/src/advanced_omi_backend/prompt_registry.py @@ -39,6 +39,7 @@ def _get_client(self): if self._langfuse is None: try: from langfuse import Langfuse + self._langfuse = Langfuse() except Exception as e: logger.warning(f"LangFuse client init failed: {e}") @@ -77,9 +78,10 @@ async def get_prompt(self, prompt_id: str, **variables) -> str: return template_text async def seed_prompts(self) -> None: - """Create prompts in LangFuse if they don't already exist. + """Create or update prompts in LangFuse, skipping unchanged ones. Called once at startup after all defaults have been registered. + Only creates a new version when the prompt text has actually changed. """ client = self._get_client() if client is None: @@ -90,6 +92,19 @@ async def seed_prompts(self) -> None: skipped = 0 for prompt_id, template_text in self._defaults.items(): try: + # Check if the prompt already exists with the same text + existing = None + try: + existing = client.get_prompt(prompt_id) + except Exception: + pass # Prompt doesn't exist yet + + if existing is not None: + existing_text = getattr(existing, "prompt", None) + if existing_text == template_text: + skipped += 1 + continue + client.create_prompt( name=prompt_id, type="text", @@ -98,13 +113,9 @@ async def seed_prompts(self) -> None: ) seeded += 1 except Exception as e: - err_msg = str(e).lower() - if "already exists" in err_msg or "409" in err_msg: - skipped += 1 - else: - logger.warning(f"Failed to seed prompt '{prompt_id}': {e}") + logger.warning(f"Failed to seed prompt '{prompt_id}': {e}") - logger.info(f"Prompt seeding complete: {seeded} created, {skipped} already existed") + logger.info(f"Prompt seeding complete: {seeded} created, {skipped} unchanged") # --------------------------------------------------------------------------- diff --git a/backends/advanced/src/advanced_omi_backend/routers/modules/queue_routes.py b/backends/advanced/src/advanced_omi_backend/routers/modules/queue_routes.py index 4410665a..29719566 100644 --- a/backends/advanced/src/advanced_omi_backend/routers/modules/queue_routes.py +++ b/backends/advanced/src/advanced_omi_backend/routers/modules/queue_routes.py @@ -32,11 +32,17 @@ async def list_jobs( queue_name: str = Query(None, description="Filter by queue name"), job_type: str = Query(None, description="Filter by job type (matches func_name)"), client_id: str = Query(None, description="Filter by client_id in meta"), - current_user: User = Depends(current_active_user) + current_user: User = Depends(current_active_user), ): """List jobs with pagination and filtering.""" try: - result = get_jobs(limit=limit, offset=offset, queue_name=queue_name, job_type=job_type, client_id=client_id) + result = get_jobs( + limit=limit, + offset=offset, + queue_name=queue_name, + job_type=job_type, + client_id=client_id, + ) # Filter jobs by user if not admin if not current_user.is_superuser: @@ -54,14 +60,20 @@ async def list_jobs( except Exception as e: logger.error(f"Failed to list jobs: {e}") - return {"error": "Failed to list jobs", "jobs": [], "pagination": {"total": 0, "limit": limit, "offset": offset, "has_more": False}} + return { + "error": "Failed to list jobs", + "jobs": [], + "pagination": { + "total": 0, + "limit": limit, + "offset": offset, + "has_more": False, + }, + } @router.get("/jobs/{job_id}/status") -async def get_job_status( - job_id: str, - current_user: User = Depends(current_active_user) -): +async def get_job_status(job_id: str, current_user: User = Depends(current_active_user)): """Get just the status of a specific job (lightweight endpoint).""" try: job = Job.fetch(job_id, connection=redis_conn) @@ -79,10 +91,7 @@ async def get_job_status( logger.error(f"Failed to determine status for job {job_id}: {e}") raise HTTPException(status_code=500, detail=str(e)) - response = { - "job_id": job.id, - "status": status - } + response = {"job_id": job.id, "status": status} # Include error information for failed jobs if status == "failed" and job.exc_info: @@ -100,10 +109,7 @@ async def get_job_status( @router.get("/jobs/{job_id}") -async def get_job( - job_id: str, - current_user: User = Depends(current_active_user) -): +async def get_job(job_id: str, current_user: User = Depends(current_active_user)): """Get detailed job information including result.""" try: job = Job.fetch(job_id, connection=redis_conn) @@ -128,7 +134,7 @@ async def get_job( "started_at": job.started_at.isoformat() if job.started_at else None, "ended_at": job.ended_at.isoformat() if job.ended_at else None, "description": job.description or "", - "func_name": job.func_name if hasattr(job, 'func_name') else "", + "func_name": job.func_name if hasattr(job, "func_name") else "", "args": job.args, "kwargs": job.kwargs, "meta": job.meta if job.meta else {}, @@ -145,10 +151,7 @@ async def get_job( @router.delete("/jobs/{job_id}") -async def cancel_job( - job_id: str, - current_user: User = Depends(current_active_user) -): +async def cancel_job(job_id: str, current_user: User = Depends(current_active_user)): """Cancel or delete a job.""" try: job = Job.fetch(job_id, connection=redis_conn) @@ -167,7 +170,7 @@ async def cancel_job( return { "job_id": job_id, "action": "canceled", - "message": f"Job {job_id} has been canceled" + "message": f"Job {job_id} has been canceled", } else: # Delete finished/failed jobs @@ -176,7 +179,7 @@ async def cancel_job( return { "job_id": job_id, "action": "deleted", - "message": f"Job {job_id} has been deleted" + "message": f"Job {job_id} has been deleted", } except HTTPException: @@ -184,14 +187,13 @@ async def cancel_job( raise except Exception as e: logger.error(f"Failed to cancel/delete job {job_id}: {e}") - raise HTTPException(status_code=404, detail=f"Job not found or could not be canceled: {str(e)}") + raise HTTPException( + status_code=404, detail=f"Job not found or could not be canceled: {str(e)}" + ) @router.get("/jobs/by-client/{client_id}") -async def get_jobs_by_client( - client_id: str, - current_user: User = Depends(current_active_user) -): +async def get_jobs_by_client(client_id: str, current_user: User = Depends(current_active_user)): """Get all jobs associated with a specific client device.""" try: from rq.registry import ( @@ -237,27 +239,31 @@ def process_job_and_dependents(job, queue_name, base_status): status = get_job_status(job, {}) # Add this job to results - all_jobs.append({ - "job_id": job.id, - "job_type": job.func_name.split('.')[-1] if job.func_name else "unknown", - "queue": queue_name, - "status": status, - "created_at": job.created_at.isoformat() if job.created_at else None, - "started_at": job.started_at.isoformat() if job.started_at else None, - "ended_at": job.ended_at.isoformat() if job.ended_at else None, - "description": job.description or "", - "result": job.result, - "meta": job.meta if job.meta else {}, - "args": job.args, - "kwargs": job.kwargs if job.kwargs else {}, - "error_message": str(job.exc_info) if job.exc_info else None, - }) + all_jobs.append( + { + "job_id": job.id, + "job_type": (job.func_name.split(".")[-1] if job.func_name else "unknown"), + "queue": queue_name, + "status": status, + "created_at": (job.created_at.isoformat() if job.created_at else None), + "started_at": (job.started_at.isoformat() if job.started_at else None), + "ended_at": job.ended_at.isoformat() if job.ended_at else None, + "description": job.description or "", + "result": job.result, + "meta": job.meta if job.meta else {}, + "args": job.args, + "kwargs": job.kwargs if job.kwargs else {}, + "error_message": str(job.exc_info) if job.exc_info else None, + } + ) # Check for dependent jobs (jobs that depend on this one) try: dependent_ids = job.dependent_ids if dependent_ids: - logger.debug(f"Job {job.id} has {len(dependent_ids)} dependents: {dependent_ids}") + logger.debug( + f"Job {job.id} has {len(dependent_ids)} dependents: {dependent_ids}" + ) for dep_id in dependent_ids: try: @@ -276,12 +282,21 @@ def process_job_and_dependents(job, queue_name, base_status): # Check all registries (using RQ standard status names) registries = [ ("queued", queue.job_ids), - ("started", StartedJobRegistry(queue=queue).get_job_ids()), # RQ standard - ("finished", FinishedJobRegistry(queue=queue).get_job_ids()), # RQ standard + ( + "started", + StartedJobRegistry(queue=queue).get_job_ids(), + ), # RQ standard + ( + "finished", + FinishedJobRegistry(queue=queue).get_job_ids(), + ), # RQ standard ("failed", FailedJobRegistry(queue=queue).get_job_ids()), - ("canceled", CanceledJobRegistry(queue=queue).get_job_ids()), # RQ standard (US spelling) + ( + "canceled", + CanceledJobRegistry(queue=queue).get_job_ids(), + ), # RQ standard (US spelling) ("deferred", DeferredJobRegistry(queue=queue).get_job_ids()), - ("scheduled", ScheduledJobRegistry(queue=queue).get_job_ids()) + ("scheduled", ScheduledJobRegistry(queue=queue).get_job_ids()), ] for status_name, job_ids in registries: @@ -293,8 +308,8 @@ def process_job_and_dependents(job, queue_name, base_status): matches_client = False # Check job.meta for client_id (current standard) - if job.meta and 'client_id' in job.meta: - if job.meta['client_id'] == client_id: + if job.meta and "client_id" in job.meta: + if job.meta["client_id"] == client_id: matches_client = True if matches_client: @@ -310,11 +325,7 @@ def process_job_and_dependents(job, queue_name, base_status): logger.info(f"Found {len(all_jobs)} jobs for client {client_id} (including dependents)") - return { - "client_id": client_id, - "jobs": all_jobs, - "total": len(all_jobs) - } + return {"client_id": client_id, "jobs": all_jobs, "total": len(all_jobs)} except Exception as e: logger.error(f"Failed to get jobs for client {client_id}: {e}") @@ -408,11 +419,8 @@ async def clear_events( raise HTTPException(status_code=500, detail=f"Failed to clear events: {str(e)}") - @router.get("/stats") -async def get_queue_stats_endpoint( - current_user: User = Depends(current_active_user) -): +async def get_queue_stats_endpoint(current_user: User = Depends(current_active_user)): """Get queue statistics.""" try: stats = get_job_stats() @@ -420,13 +428,19 @@ async def get_queue_stats_endpoint( except Exception as e: logger.error(f"Failed to get queue stats: {e}") - return {"total_jobs": 0, "queued_jobs": 0, "started_jobs": 0, "finished_jobs": 0, "failed_jobs": 0, "canceled_jobs": 0, "deferred_jobs": 0} + return { + "total_jobs": 0, + "queued_jobs": 0, + "started_jobs": 0, + "finished_jobs": 0, + "failed_jobs": 0, + "canceled_jobs": 0, + "deferred_jobs": 0, + } @router.get("/worker-details") -async def get_queue_worker_details( - current_user: User = Depends(current_active_user) -): +async def get_queue_worker_details(current_user: User = Depends(current_active_user)): """Get detailed queue and worker status including task manager health.""" try: import time @@ -443,10 +457,10 @@ async def get_queue_worker_details( "total": queue_health.get("total_workers", 0), "active": queue_health.get("active_workers", 0), "idle": queue_health.get("idle_workers", 0), - "details": queue_health.get("workers", []) + "details": queue_health.get("workers", []), }, "queues": queue_health.get("queues", {}), - "redis_connection": queue_health.get("redis_connection", "unknown") + "redis_connection": queue_health.get("redis_connection", "unknown"), } return status @@ -459,18 +473,16 @@ async def get_queue_worker_details( @router.get("/streams") async def get_stream_stats( limit: int = Query(default=10, ge=1, le=100), # Max 100 streams to prevent timeouts - current_user: User = Depends(current_active_user) + current_user: User = Depends(current_active_user), ): """Get Redis Streams statistics with consumer group information.""" try: from advanced_omi_backend.services.audio_service import get_audio_stream_service + audio_service = get_audio_stream_service() if not audio_service.redis: - return { - "error": "Audio stream service not connected", - "streams": [] - } + return {"error": "Audio stream service not connected", "streams": []} # Get audio streams with limit stream_keys = [] @@ -479,7 +491,7 @@ async def get_stream_stats( cursor, keys = await audio_service.redis.scan( cursor, match=f"{audio_service.audio_stream_prefix}*", count=limit ) - stream_keys.extend(keys[:limit - len(stream_keys)]) + stream_keys.extend(keys[: limit - len(stream_keys)]) # Use asyncio.gather to fetch stream info in parallel import asyncio @@ -499,9 +511,13 @@ async def get_stream_info(stream_key): group_dict = {} # Parse group info (alternating key-value pairs) for i in range(0, len(group), 2): - if i+1 < len(group): - key = group[i].decode() if isinstance(group[i], bytes) else str(group[i]) - value = group[i+1] + if i + 1 < len(group): + key = ( + group[i].decode() + if isinstance(group[i], bytes) + else str(group[i]) + ) + value = group[i + 1] if isinstance(value, bytes): try: value = value.decode() @@ -512,13 +528,19 @@ async def get_stream_info(stream_key): # Get consumers for this group consumers = [] try: - consumers_raw = await audio_service.redis.xinfo_consumers(stream_name, group_dict.get('name', '')) + consumers_raw = await audio_service.redis.xinfo_consumers( + stream_name, group_dict.get("name", "") + ) for consumer in consumers_raw: consumer_dict = {} for i in range(0, len(consumer), 2): - if i+1 < len(consumer): - key = consumer[i].decode() if isinstance(consumer[i], bytes) else str(consumer[i]) - value = consumer[i+1] + if i + 1 < len(consumer): + key = ( + consumer[i].decode() + if isinstance(consumer[i], bytes) + else str(consumer[i]) + ) + value = consumer[i + 1] if isinstance(value, bytes): try: value = value.decode() @@ -527,24 +549,32 @@ async def get_stream_info(stream_key): consumer_dict[key] = value consumers.append(consumer_dict) except Exception as ce: - logger.debug(f"Could not fetch consumers for group {group_dict.get('name')}: {ce}") - - groups_info.append({ - "name": group_dict.get('name', 'unknown'), - "consumers": group_dict.get('consumers', 0), - "pending": group_dict.get('pending', 0), - "last_delivered_id": group_dict.get('last-delivered-id', 'N/A'), - "consumer_details": consumers - }) + logger.debug( + f"Could not fetch consumers for group {group_dict.get('name')}: {ce}" + ) + + groups_info.append( + { + "name": group_dict.get("name", "unknown"), + "consumers": group_dict.get("consumers", 0), + "pending": group_dict.get("pending", 0), + "last_delivered_id": group_dict.get("last-delivered-id", "N/A"), + "consumer_details": consumers, + } + ) except Exception as ge: logger.debug(f"No consumer groups for stream {stream_name}: {ge}") return { "stream_name": stream_name, "length": info[b"length"], - "first_entry_id": info[b"first-entry"][0].decode() if info[b"first-entry"] else None, - "last_entry_id": info[b"last-entry"][0].decode() if info[b"last-entry"] else None, - "groups": groups_info + "first_entry_id": ( + info[b"first-entry"][0].decode() if info[b"first-entry"] else None + ), + "last_entry_id": ( + info[b"last-entry"][0].decode() if info[b"last-entry"] else None + ), + "groups": groups_info, } except Exception as e: logger.error(f"Error getting info for stream {stream_key}: {e}") @@ -557,16 +587,12 @@ async def get_stream_info(stream_key): return { "total_streams": len(streams_info), "streams": streams_info, - "limited": len(stream_keys) >= limit + "limited": len(stream_keys) >= limit, } except Exception as e: logger.error(f"Failed to get stream stats: {e}", exc_info=True) - return { - "error": str(e), - "total_streams": 0, - "streams": [] - } + return {"error": str(e), "total_streams": 0, "streams": []} class FlushJobsRequest(BaseModel): @@ -581,10 +607,7 @@ class FlushAllJobsRequest(BaseModel): @router.post("/flush") -async def flush_jobs( - request: FlushJobsRequest, - current_user: User = Depends(current_active_user) -): +async def flush_jobs(request: FlushJobsRequest, current_user: User = Depends(current_active_user)): """Flush old inactive jobs based on age and status.""" if not current_user.is_superuser: raise HTTPException(status_code=403, detail="Admin access required") @@ -646,7 +669,7 @@ async def flush_jobs( return { "total_removed": total_removed, "cutoff_time": cutoff_time.isoformat(), - "statuses": request.statuses + "statuses": request.statuses, } except Exception as e: @@ -656,8 +679,7 @@ async def flush_jobs( @router.post("/flush-all") async def flush_all_jobs( - request: FlushAllJobsRequest, - current_user: User = Depends(current_active_user) + request: FlushAllJobsRequest, current_user: User = Depends(current_active_user) ): """ Flush jobs from queues and registries. @@ -696,10 +718,16 @@ async def flush_all_jobs( # Build list of registries to flush based on request parameters registries = [ - ("started", StartedJobRegistry(queue=queue)), # Always flush in-progress + ( + "started", + StartedJobRegistry(queue=queue), + ), # Always flush in-progress ("deferred", DeferredJobRegistry(queue=queue)), # Always flush deferred - ("scheduled", ScheduledJobRegistry(queue=queue)), # Always flush scheduled - ("canceled", CanceledJobRegistry(queue=queue)) # Always flush canceled + ( + "scheduled", + ScheduledJobRegistry(queue=queue), + ), # Always flush scheduled + ("canceled", CanceledJobRegistry(queue=queue)), # Always flush canceled ] # Conditionally add failed and finished registries @@ -729,13 +757,16 @@ async def flush_all_jobs( # This lets the worker clean up gracefully and prevents deadlock try: from rq.command import send_stop_job_command + send_stop_job_command(redis_conn, job_id) logger.info(f"Sent stop command to worker for job {job_id}") # Don't delete yet - let worker move it to canceled/failed registry # It will be cleaned up on next flush or by worker cleanup continue except Exception as stop_error: - logger.warning(f"Could not send stop command to job {job_id}: {stop_error}") + logger.warning( + f"Could not send stop command to job {job_id}: {stop_error}" + ) # If stop fails, try to cancel it (may already be finishing) try: job.cancel() @@ -752,7 +783,9 @@ async def flush_all_jobs( logger.warning(f"Error deleting job {job_id}: {e}") try: registry.remove(job_id) - logger.info(f"Removed stale job reference {job_id} from {registry_name} registry") + logger.info( + f"Removed stale job reference {job_id} from {registry_name} registry" + ) except Exception as reg_error: logger.error(f"Could not remove {job_id} from registry: {reg_error}") @@ -761,6 +794,7 @@ async def flush_all_jobs( # Get async Redis connection for scanning from advanced_omi_backend.controllers.queue_controller import REDIS_URL + async_redis = await aioredis.from_url(REDIS_URL) try: @@ -793,13 +827,15 @@ async def flush_all_jobs( preserved.append("finished jobs") preserved_msg = f" (preserved {', '.join(preserved)})" if preserved else "" - logger.info(f"Flushed {total_removed} jobs and {deleted_keys} Redis keys from all queues{preserved_msg}") + logger.info( + f"Flushed {total_removed} jobs and {deleted_keys} Redis keys from all queues{preserved_msg}" + ) return { "total_removed": total_removed, "deleted_keys": deleted_keys, "preserved": preserved, - "message": f"Flushed {total_removed} jobs{preserved_msg}" + "message": f"Flushed {total_removed} jobs{preserved_msg}", } except Exception as e: @@ -810,7 +846,7 @@ async def flush_all_jobs( @router.get("/sessions") async def get_redis_sessions( limit: int = Query(default=20, ge=1, le=100), - current_user: User = Depends(current_active_user) + current_user: User = Depends(current_active_user), ): """Get Redis session tracking information.""" try: @@ -819,51 +855,54 @@ async def get_redis_sessions( from advanced_omi_backend.controllers.queue_controller import REDIS_URL redis_client = aioredis.from_url(REDIS_URL) - - # Get session keys - session_keys = [] - cursor = b"0" - while cursor and len(session_keys) < limit: - cursor, keys = await redis_client.scan( - cursor, match="audio:session:*", count=limit - ) - session_keys.extend(keys[:limit - len(session_keys)]) - - # Get session info - sessions = [] - for key in session_keys: - try: - session_data = await redis_client.hgetall(key) - if session_data: - session_id = key.decode().replace("audio:session:", "") - - # Get conversation count for this session - conversation_count_key = f"session:conversation_count:{session_id}" - conversation_count_bytes = await redis_client.get(conversation_count_key) - conversation_count = int(conversation_count_bytes.decode()) if conversation_count_bytes else 0 - - sessions.append({ - "session_id": session_id, - "user_id": session_data.get(b"user_id", b"").decode(), - "client_id": session_data.get(b"client_id", b"").decode(), - "stream_name": session_data.get(b"stream_name", b"").decode(), - "provider": session_data.get(b"provider", b"").decode(), - "mode": session_data.get(b"mode", b"").decode(), - "status": session_data.get(b"status", b"").decode(), - "started_at": session_data.get(b"started_at", b"").decode(), - "chunks_published": int(session_data.get(b"chunks_published", b"0").decode() or 0), - "last_chunk_at": session_data.get(b"last_chunk_at", b"").decode(), - "conversation_count": conversation_count - }) - except Exception as e: - logger.error(f"Error getting session info for {key}: {e}") - - await redis_client.close() - - return { - "total_sessions": len(sessions), - "sessions": sessions - } + try: + # Get session keys + session_keys = [] + cursor = b"0" + while cursor and len(session_keys) < limit: + cursor, keys = await redis_client.scan(cursor, match="audio:session:*", count=limit) + session_keys.extend(keys[: limit - len(session_keys)]) + + # Get session info + sessions = [] + for key in session_keys: + try: + session_data = await redis_client.hgetall(key) + if session_data: + session_id = key.decode().replace("audio:session:", "") + + # Get conversation count for this session + conversation_count_key = f"session:conversation_count:{session_id}" + conversation_count_bytes = await redis_client.get(conversation_count_key) + conversation_count = ( + int(conversation_count_bytes.decode()) + if conversation_count_bytes + else 0 + ) + + sessions.append( + { + "session_id": session_id, + "user_id": session_data.get(b"user_id", b"").decode(), + "client_id": session_data.get(b"client_id", b"").decode(), + "stream_name": session_data.get(b"stream_name", b"").decode(), + "provider": session_data.get(b"provider", b"").decode(), + "mode": session_data.get(b"mode", b"").decode(), + "status": session_data.get(b"status", b"").decode(), + "started_at": session_data.get(b"started_at", b"").decode(), + "chunks_published": int( + session_data.get(b"chunks_published", b"0").decode() or 0 + ), + "last_chunk_at": session_data.get(b"last_chunk_at", b"").decode(), + "conversation_count": conversation_count, + } + ) + except Exception as e: + logger.error(f"Error getting session info for {key}: {e}") + + return {"total_sessions": len(sessions), "sessions": sessions} + finally: + await redis_client.aclose() except Exception as e: logger.error(f"Failed to get sessions: {e}", exc_info=True) @@ -872,8 +911,10 @@ async def get_redis_sessions( @router.post("/sessions/clear") async def clear_old_sessions( - older_than_seconds: int = Query(default=3600, description="Clear sessions older than N seconds"), - current_user: User = Depends(current_active_user) + older_than_seconds: int = Query( + default=3600, description="Clear sessions older than N seconds" + ), + current_user: User = Depends(current_active_user), ): """Clear old Redis sessions that are stuck or inactive.""" if not current_user.is_superuser: @@ -887,38 +928,36 @@ async def clear_old_sessions( from advanced_omi_backend.controllers.queue_controller import REDIS_URL redis_client = aioredis.from_url(REDIS_URL) - current_time = time.time() - cutoff_time = current_time - older_than_seconds - - # Get all session keys - session_keys = [] - cursor = b"0" - while cursor: - cursor, keys = await redis_client.scan(cursor, match="audio:session:*", count=100) - session_keys.extend(keys) - - # Check each session and delete if old - deleted_count = 0 - for key in session_keys: - try: - session_data = await redis_client.hgetall(key) - if session_data: - last_chunk_at = session_data.get(b"last_chunk_at", b"").decode() - if last_chunk_at: - last_chunk_time = float(last_chunk_at) - if last_chunk_time < cutoff_time: - await redis_client.delete(key) - deleted_count += 1 - logger.info(f"Deleted old session: {key.decode()}") - except Exception as e: - logger.error(f"Error processing session {key}: {e}") - - await redis_client.close() - - return { - "deleted_count": deleted_count, - "cutoff_seconds": older_than_seconds - } + try: + current_time = time.time() + cutoff_time = current_time - older_than_seconds + + # Get all session keys + session_keys = [] + cursor = b"0" + while cursor: + cursor, keys = await redis_client.scan(cursor, match="audio:session:*", count=100) + session_keys.extend(keys) + + # Check each session and delete if old + deleted_count = 0 + for key in session_keys: + try: + session_data = await redis_client.hgetall(key) + if session_data: + last_chunk_at = session_data.get(b"last_chunk_at", b"").decode() + if last_chunk_at: + last_chunk_time = float(last_chunk_at) + if last_chunk_time < cutoff_time: + await redis_client.delete(key) + deleted_count += 1 + logger.info(f"Deleted old session: {key.decode()}") + except Exception as e: + logger.error(f"Error processing session {key}: {e}") + + return {"deleted_count": deleted_count, "cutoff_seconds": older_than_seconds} + finally: + await redis_client.aclose() except Exception as e: logger.error(f"Failed to clear sessions: {e}", exc_info=True) @@ -928,8 +967,10 @@ async def clear_old_sessions( @router.get("/dashboard") async def get_dashboard_data( request: Request, - expanded_clients: str = Query(default="", description="Comma-separated list of client IDs to fetch jobs for"), - current_user: User = Depends(current_active_user) + expanded_clients: str = Query( + default="", description="Comma-separated list of client IDs to fetch jobs for" + ), + current_user: User = Depends(current_active_user), ): """Get all data needed for the Queue dashboard in a single API call. @@ -950,7 +991,11 @@ async def get_dashboard_data( from advanced_omi_backend.controllers.queue_controller import get_queue # Parse expanded clients list - expanded_client_ids = [c.strip() for c in expanded_clients.split(",") if c.strip()] if expanded_clients else [] + expanded_client_ids = ( + [c.strip() for c in expanded_clients.split(",") if c.strip()] + if expanded_clients + else [] + ) # Fetch all data in parallel import asyncio @@ -988,26 +1033,40 @@ async def fetch_jobs_by_status(status_name: str, limit: int = 100): continue # Add job with metadata - all_jobs.append({ - "job_id": job.id, - "job_type": job.func_name.split('.')[-1] if job.func_name else "unknown", - "user_id": job.kwargs.get("user_id") if job.kwargs else None, - "status": status_name, - "priority": "normal", # RQ doesn't have priority concept - "data": {"description": job.description or ""}, - "result": job.result, - "meta": job.meta if job.meta else {}, - "kwargs": job.kwargs if job.kwargs else {}, - "error_message": str(job.exc_info) if job.exc_info else None, - "created_at": job.created_at.isoformat() if job.created_at else None, - "started_at": job.started_at.isoformat() if job.started_at else None, - "ended_at": job.ended_at.isoformat() if job.ended_at else None, - "retry_count": 0, # RQ doesn't track this by default - "max_retries": 0, - "progress_percent": 0, - "progress_message": "", - "queue": queue_name - }) + all_jobs.append( + { + "job_id": job.id, + "job_type": ( + job.func_name.split(".")[-1] if job.func_name else "unknown" + ), + "user_id": (job.kwargs.get("user_id") if job.kwargs else None), + "status": status_name, + "priority": "normal", # RQ doesn't have priority concept + "data": {"description": job.description or ""}, + "result": job.result, + "meta": job.meta if job.meta else {}, + "kwargs": job.kwargs if job.kwargs else {}, + "error_message": (str(job.exc_info) if job.exc_info else None), + "created_at": ( + job.created_at.isoformat() if job.created_at else None + ), + "started_at": ( + job.started_at.isoformat() if job.started_at else None + ), + "ended_at": ( + job.ended_at.isoformat() if job.ended_at else None + ), + "retry_count": 0, # RQ doesn't track this by default + "max_retries": 0, + "progress_percent": (job.meta or {}) + .get("batch_progress", {}) + .get("percent", 0), + "progress_message": (job.meta or {}) + .get("batch_progress", {}) + .get("message", ""), + "queue": queue_name, + } + ) except Exception as e: logger.debug(f"Error fetching job {job_id}: {e}") continue @@ -1023,7 +1082,13 @@ async def fetch_stats(): return get_job_stats() except Exception as e: logger.error(f"Error fetching stats: {e}") - return {"total_jobs": 0, "queued_jobs": 0, "started_jobs": 0, "finished_jobs": 0, "failed_jobs": 0} + return { + "total_jobs": 0, + "queued_jobs": 0, + "started_jobs": 0, + "finished_jobs": 0, + "failed_jobs": 0, + } async def fetch_streaming_status(): """Fetch streaming status.""" @@ -1072,9 +1137,15 @@ def get_job_status(job): registries = [ ("queued", queue.job_ids), - ("started", StartedJobRegistry(queue=queue).get_job_ids()), # RQ standard - ("finished", FinishedJobRegistry(queue=queue).get_job_ids()), # RQ standard - ("failed", FailedJobRegistry(queue=queue).get_job_ids()) + ( + "started", + StartedJobRegistry(queue=queue).get_job_ids(), + ), # RQ standard + ( + "finished", + FinishedJobRegistry(queue=queue).get_job_ids(), + ), # RQ standard + ("failed", FailedJobRegistry(queue=queue).get_job_ids()), ] for status_name, job_ids in registries: @@ -1087,7 +1158,11 @@ def get_job_status(job): # Check if job belongs to this client matches_client = False - if job.meta and 'client_id' in job.meta and job.meta['client_id'] == client_id: + if ( + job.meta + and "client_id" in job.meta + and job.meta["client_id"] == client_id + ): matches_client = True if not matches_client: @@ -1100,19 +1175,33 @@ def get_job_status(job): continue processed_job_ids.add(job_id) - all_jobs.append({ - "job_id": job.id, - "job_type": job.func_name.split('.')[-1] if job.func_name else "unknown", - "queue": queue_name, - "status": get_job_status(job), - "created_at": job.created_at.isoformat() if job.created_at else None, - "started_at": job.started_at.isoformat() if job.started_at else None, - "ended_at": job.ended_at.isoformat() if job.ended_at else None, - "description": job.description or "", - "result": job.result, - "meta": job.meta if job.meta else {}, - "error_message": str(job.exc_info) if job.exc_info else None - }) + all_jobs.append( + { + "job_id": job.id, + "job_type": ( + job.func_name.split(".")[-1] + if job.func_name + else "unknown" + ), + "queue": queue_name, + "status": get_job_status(job), + "created_at": ( + job.created_at.isoformat() if job.created_at else None + ), + "started_at": ( + job.started_at.isoformat() if job.started_at else None + ), + "ended_at": ( + job.ended_at.isoformat() if job.ended_at else None + ), + "description": job.description or "", + "result": job.result, + "meta": job.meta if job.meta else {}, + "error_message": ( + str(job.exc_info) if job.exc_info else None + ), + } + ) except Exception as e: logger.debug(f"Error fetching job {job_id}: {e}") continue @@ -1127,7 +1216,9 @@ async def fetch_events(): if not current_user.is_superuser: return [] try: - from advanced_omi_backend.services.plugin_service import get_plugin_router + from advanced_omi_backend.services.plugin_service import ( + get_plugin_router, + ) router_instance = get_plugin_router() if not router_instance: @@ -1139,8 +1230,12 @@ async def fetch_events(): # Execute all fetches in parallel (using RQ standard status names) queued_jobs_task = fetch_jobs_by_status("queued", limit=100) - started_jobs_task = fetch_jobs_by_status("started", limit=100) # RQ standard, not "processing" - finished_jobs_task = fetch_jobs_by_status("finished", limit=50) # RQ standard, not "completed" + started_jobs_task = fetch_jobs_by_status( + "started", limit=100 + ) # RQ standard, not "processing" + finished_jobs_task = fetch_jobs_by_status( + "finished", limit=50 + ) # RQ standard, not "completed" failed_jobs_task = fetch_jobs_by_status("failed", limit=50) stats_task = fetch_stats() streaming_status_task = fetch_streaming_status() @@ -1156,7 +1251,7 @@ async def fetch_events(): streaming_status_task, events_task, *client_jobs_tasks, - return_exceptions=True + return_exceptions=True, ) queued_jobs = results[0] if not isinstance(results[0], Exception) else [] @@ -1164,7 +1259,9 @@ async def fetch_events(): finished_jobs = results[2] if not isinstance(results[2], Exception) else [] # RQ standard failed_jobs = results[3] if not isinstance(results[3], Exception) else [] stats = results[4] if not isinstance(results[4], Exception) else {"total_jobs": 0} - streaming_status = results[5] if not isinstance(results[5], Exception) else {"active_sessions": []} + streaming_status = ( + results[5] if not isinstance(results[5], Exception) else {"active_sessions": []} + ) events = results[6] if not isinstance(results[6], Exception) else [] recent_conversations = [] client_jobs_results = results[7:] if len(results) > 7 else [] @@ -1178,28 +1275,34 @@ async def fetch_events(): # Convert conversations to dict format for frontend conversations_list = [] for conv in recent_conversations: - conversations_list.append({ - "conversation_id": conv.conversation_id, - "user_id": str(conv.user_id) if conv.user_id else None, - "created_at": conv.created_at.isoformat() if conv.created_at else None, - "title": conv.title, - "summary": conv.summary, - "transcript_text": conv.get_active_transcript_text() if hasattr(conv, 'get_active_transcript_text') else None, - }) + conversations_list.append( + { + "conversation_id": conv.conversation_id, + "user_id": str(conv.user_id) if conv.user_id else None, + "created_at": (conv.created_at.isoformat() if conv.created_at else None), + "title": conv.title, + "summary": conv.summary, + "transcript_text": ( + conv.get_active_transcript_text() + if hasattr(conv, "get_active_transcript_text") + else None + ), + } + ) return { "jobs": { "queued": queued_jobs, "started": started_jobs, # RQ standard status name "finished": finished_jobs, # RQ standard status name - "failed": failed_jobs + "failed": failed_jobs, }, "stats": stats, "streaming_status": streaming_status, "recent_conversations": conversations_list, "client_jobs": client_jobs, "events": events, - "timestamp": asyncio.get_event_loop().time() + "timestamp": asyncio.get_event_loop().time(), } except Exception as e: diff --git a/backends/advanced/src/advanced_omi_backend/services/audio_stream/producer.py b/backends/advanced/src/advanced_omi_backend/services/audio_stream/producer.py index dc5e9b27..61244f43 100644 --- a/backends/advanced/src/advanced_omi_backend/services/audio_stream/producer.py +++ b/backends/advanced/src/advanced_omi_backend/services/audio_stream/producer.py @@ -2,14 +2,12 @@ Audio stream producer - publishes audio chunks to Redis Streams. """ +import json import logging import time -import json import redis.asyncio as redis -from advanced_omi_backend.services.transcription.base import TranscriptionProvider - logger = logging.getLogger(__name__) @@ -45,7 +43,7 @@ async def init_session( user_email: str = "", connection_id: str = "", mode: str = "streaming", - provider: str = "deepgram" + provider: str = "deepgram", ): """ Initialize session tracking metadata in Redis. @@ -66,35 +64,32 @@ async def init_session( stream_name = f"audio:stream:{client_id}" session_key = f"audio:session:{session_id}" - await self.redis_client.hset(session_key, mapping={ - # User & Client tracking - "user_id": user_id, - "user_email": user_email, - "client_id": client_id, - "connection_id": connection_id, - - # Stream configuration - "stream_name": stream_name, - "provider": provider, - "mode": mode, - - # Timestamps - "started_at": str(time.time()), - "last_chunk_at": str(time.time()), - - # Counters - "chunks_published": "0", - - # Job tracking (populated by queue_controller when jobs start) - "speech_detection_job_id": "", - "audio_persistence_job_id": "", - - # Connection state - "websocket_connected": "true", - - # Session status - "status": "active" - }) + await self.redis_client.hset( + session_key, + mapping={ + # User & Client tracking + "user_id": user_id, + "user_email": user_email, + "client_id": client_id, + "connection_id": connection_id, + # Stream configuration + "stream_name": stream_name, + "provider": provider, + "mode": mode, + # Timestamps + "started_at": str(time.time()), + "last_chunk_at": str(time.time()), + # Counters + "chunks_published": "0", + # Job tracking (populated by queue_controller when jobs start) + "speech_detection_job_id": "", + "audio_persistence_job_id": "", + # Connection state + "websocket_connected": "true", + # Session status + "status": "active", + }, + ) # Set TTL of 1 hour await self.redis_client.expire(session_key, 3600) @@ -106,10 +101,12 @@ async def init_session( "user_id": user_id, "client_id": client_id, "stream_name": stream_name, - "provider": provider + "provider": provider, } - logger.info(f"šŸ“Š Initialized session {session_id} ā†’ stream {stream_name} (provider: {provider})") + logger.info( + f"šŸ“Š Initialized session {session_id} ā†’ stream {stream_name} (provider: {provider})" + ) async def update_session_chunk_count(self, session_id: str): """ @@ -166,10 +163,7 @@ async def send_session_end_signal(self, session_id: str): } await self.redis_client.xadd( - stream_name, - end_signal, - maxlen=25000, - approximate=True + stream_name, end_signal, maxlen=25000, approximate=True ) logger.info(f"šŸ“” Sent end-of-session signal for {session_id} to {stream_name}") @@ -187,14 +181,22 @@ async def get_session(self, session_id: str) -> dict: session_data = await self.redis_client.hgetall(session_key) # Convert bytes to strings for easier handling - return {k.decode() if isinstance(k, bytes) else k: v.decode() if isinstance(v, bytes) else v - for k, v in session_data.items()} if session_data else {} + return ( + { + k.decode() if isinstance(k, bytes) else k: ( + v.decode() if isinstance(v, bytes) else v + ) + for k, v in session_data.items() + } + if session_data + else {} + ) async def update_session_job_ids( self, session_id: str, speech_detection_job_id: str = None, - audio_persistence_job_id: str = None + audio_persistence_job_id: str = None, ): """ Update job IDs in session metadata. @@ -224,10 +226,13 @@ async def mark_websocket_disconnected(self, session_id: str): session_id: Session identifier """ session_key = f"audio:session:{session_id}" - await self.redis_client.hset(session_key, mapping={ - "websocket_connected": "false", - "disconnected_at": str(time.time()) - }) + await self.redis_client.hset( + session_key, + mapping={ + "websocket_connected": "false", + "disconnected_at": str(time.time()), + }, + ) logger.info(f"šŸ”Œ Marked websocket disconnected for session {session_id}") async def finalize_session(self, session_id: str, completion_reason: str = None): @@ -242,15 +247,14 @@ async def finalize_session(self, session_id: str, completion_reason: str = None) session_key = f"audio:session:{session_id}" # Build mapping with status and optional completion_reason - mapping = { - "status": "finalizing", - "finalized_at": str(time.time()) - } + mapping = {"status": "finalizing", "finalized_at": str(time.time())} # Set completion_reason atomically with status to prevent race conditions if completion_reason: mapping["completion_reason"] = completion_reason - logger.info(f"šŸ“Š Finalizing session {session_id} with reason: {completion_reason}") + logger.info( + f"šŸ“Š Finalizing session {session_id} with reason: {completion_reason}" + ) await self.redis_client.hset(session_key, mapping=mapping) @@ -269,10 +273,7 @@ async def finalize_session(self, session_id: str, completion_reason: str = None) } await self.redis_client.xadd( - stream_name, - end_marker_data, - maxlen=25000, - approximate=True + stream_name, end_marker_data, maxlen=25000, approximate=True ) logger.info(f"šŸ“” Sent end_marker to {stream_name} for session {session_id}") @@ -286,12 +287,11 @@ async def add_audio_chunk( self, audio_data: bytes, session_id: str, - chunk_id: str, user_id: str, client_id: str, sample_rate: int = 16000, channels: int = 1, - sample_width: int = 2 + sample_width: int = 2, ) -> list[str]: """ Add audio data to session buffer and publish fixed-size chunks. @@ -302,7 +302,6 @@ async def add_audio_chunk( Args: audio_data: Raw PCM audio bytes (arbitrary size from WebSocket) session_id: Session identifier - chunk_id: Base chunk identifier (will increment for multiple chunks) user_id: User identifier client_id: Client identifier (used for stream naming) sample_rate: Audio sample rate (Hz) @@ -321,7 +320,7 @@ async def add_audio_chunk( "user_id": user_id, "client_id": client_id, "stream_name": stream_name, - "provider": "deepgram" + "provider": "deepgram", } session_buffer = self.session_buffers[session_id] @@ -366,7 +365,7 @@ async def add_audio_chunk( stream_name, chunk_data, maxlen=25000, # Keep max 25k chunks (~104 minutes at 250ms/chunk) - approximate=True + approximate=True, ) message_ids.append(message_id.decode()) @@ -374,7 +373,10 @@ async def add_audio_chunk( await self.update_session_chunk_count(session_id) # Log every 10th chunk to avoid spam - if session_buffer["chunk_count"] % 10 == 0 or session_buffer["chunk_count"] <= 5: + if ( + session_buffer["chunk_count"] % 10 == 0 + or session_buffer["chunk_count"] <= 5 + ): logger.debug( f"šŸ“¤ Added fixed-size chunk {chunk_id_formatted} to {stream_name} " f"({len(chunk_audio)} bytes = {len(chunk_audio)/bytes_per_second:.3f}s, " @@ -396,7 +398,7 @@ async def flush_session_buffer( session_id: str, sample_rate: int = 16000, channels: int = 1, - sample_width: int = 2 + sample_width: int = 2, ) -> str | None: """ Flush any remaining audio in session buffer. @@ -443,10 +445,7 @@ async def flush_session_buffer( # Add to stream with MAXLEN limit message_id = await self.redis_client.xadd( - stream_name, - chunk_data, - maxlen=25000, - approximate=True + stream_name, chunk_data, maxlen=25000, approximate=True ) # Update session tracking @@ -463,7 +462,6 @@ async def flush_session_buffer( return None - # Singleton instance _producer_instance = None @@ -486,12 +484,12 @@ def get_audio_stream_producer() -> AudioStreamProducer: # Create async Redis client (synchronous call, connection happens on first use) redis_client = redis_async.from_url( - redis_url, - encoding="utf-8", - decode_responses=False + redis_url, encoding="utf-8", decode_responses=False ) _producer_instance = AudioStreamProducer(redis_client) - logger.info(f"Created AudioStreamProducer singleton with Redis URL: {redis_url}") + logger.info( + f"Created AudioStreamProducer singleton with Redis URL: {redis_url}" + ) return _producer_instance diff --git a/backends/advanced/src/advanced_omi_backend/services/transcription/__init__.py b/backends/advanced/src/advanced_omi_backend/services/transcription/__init__.py index 84e26f11..804077f4 100644 --- a/backends/advanced/src/advanced_omi_backend/services/transcription/__init__.py +++ b/backends/advanced/src/advanced_omi_backend/services/transcription/__init__.py @@ -42,13 +42,10 @@ def _parse_hot_words_to_keyterm(hot_words_str: str) -> str: terms = [] for word in re.split(r"[,\n]+", hot_words_str): - word = word.strip() + word = word.strip().lower() if not word: continue terms.append(word) - capitalized = word.title() - if capitalized != word: - terms.append(capitalized) return " ".join(terms) @@ -61,11 +58,11 @@ def _dotted_get(d: dict | list | None, dotted: Optional[str]): if d is None or not dotted: return None cur = d - for part in dotted.split('.'): + for part in dotted.split("."): if not part: continue - if '[' in part and part.endswith(']'): - name, idx_str = part[:-1].split('[', 1) + if "[" in part and part.endswith("]"): + name, idx_str = part[:-1].split("[", 1) if name: cur = cur.get(name, {}) if isinstance(cur, dict) else {} try: @@ -83,6 +80,40 @@ def _dotted_get(d: dict | list | None, dotted: Optional[str]): return cur +def _normalize_provider_segments(segments: list) -> list: + """Normalize provider-specific segment formats to a standard shape. + + Handles Deepgram paragraph format where: + - text is nested in ``sentences[].text`` instead of a top-level ``text`` field + - ``speaker`` is an integer (0, 1) instead of a string ("Speaker 0") + + After normalization every segment dict will have: + - ``text`` (str): combined sentence text + - ``speaker`` (str): "Speaker N" label + - ``start`` / ``end`` (float): time span (preserved from original) + """ + if not segments: + return segments + + for seg in segments: + if not isinstance(seg, dict): + continue + + # Deepgram paragraphs: text lives inside sentences[], not top-level + if "text" not in seg and "sentences" in seg: + sentences = seg.get("sentences", []) + seg["text"] = " ".join( + s.get("text", "") for s in sentences if isinstance(s, dict) + ) + + # Normalise integer speaker IDs to "Speaker N" strings + speaker = seg.get("speaker") + if isinstance(speaker, (int, float)): + seg["speaker"] = f"Speaker {int(speaker)}" + + return segments + + class RegistryBatchTranscriptionProvider(BatchTranscriptionProvider): """Batch transcription provider driven by config.yml.""" @@ -124,24 +155,33 @@ def get_capabilities_dict(self) -> dict: """ return {cap: True for cap in self._capabilities} - async def transcribe(self, audio_data: bytes, sample_rate: int, diarize: bool = False, context_info: Optional[str] = None, **kwargs) -> dict: + async def transcribe( + self, + audio_data: bytes, + sample_rate: int, + diarize: bool = False, + context_info: Optional[str] = None, + progress_callback=None, + **kwargs, + ) -> dict: # Special handling for mock provider (no HTTP server needed) if self.model.model_provider == "mock": from .mock_provider import MockTranscriptionProvider + mock = MockTranscriptionProvider(fail_mode=False) return await mock.transcribe(audio_data, sample_rate, diarize) op = (self.model.operations or {}).get("stt_transcribe") or {} method = (op.get("method") or "POST").upper() - path = (op.get("path") or "/listen") + path = op.get("path") or "/listen" # Build URL base = self.model.model_url.rstrip("/") url = base + ("/" + path.lstrip("/")) - + # Check if we should use multipart file upload (for Parakeet) content_type = op.get("content_type", "audio/raw") use_multipart = content_type == "multipart/form-data" - + # Build headers (skip Content-Type for multipart as httpx will set it) headers = {} if not use_multipart: @@ -152,7 +192,7 @@ async def transcribe(self, audio_data: bytes, sample_rate: int, diarize: bool = headers["Content-Type"] = "audio/wav" else: headers["Content-Type"] = "audio/raw" - + if self.model.api_key: # Allow templated header, otherwise fallback to Bearer/Token conventions by config hdrs = op.get("headers") or {} @@ -167,7 +207,10 @@ async def transcribe(self, audio_data: bytes, sample_rate: int, diarize: bool = hdrs = op.get("headers") or {} for k, v in hdrs.items(): # Skip Authorization headers with empty/invalid values - if k.lower() == "authorization" and (not v or v.strip().lower() in ["token", "token ", "bearer", "bearer "]): + if k.lower() == "authorization" and ( + not v + or v.strip().lower() in ["token", "token ", "bearer", "bearer "] + ): continue headers[k] = v @@ -196,24 +239,72 @@ async def transcribe(self, audio_data: bytes, sample_rate: int, diarize: bool = if keyterm: query["keyterm"] = keyterm + # NOTE: PULSE (smallest.ai) does NOT support keywords on WebSocket or + # batch HTTP ā€” any `keywords` query param causes 0 responses or HTTP 400. + # Hot-word boosting for PULSE is not injected here. + timeout = op.get("timeout", 300) + # Use a longer read timeout for NDJSON progress responses ā€” each + # batch window can take minutes but the service keeps sending + # progress lines between windows. + read_timeout = op.get("read_timeout", timeout) try: - async with httpx.AsyncClient(timeout=timeout) as client: + timeouts = httpx.Timeout(timeout, read=read_timeout) + async with httpx.AsyncClient(timeout=timeouts) as client: if method == "POST": if use_multipart: # Send as multipart file upload (for Parakeet/VibeVoice) files = {"file": ("audio.wav", audio_data, "audio/wav")} - data = {} + form_data = {} if hot_words_str and hot_words_str.strip(): - data["context_info"] = hot_words_str.strip() - resp = await client.post(url, headers=headers, params=query, files=files, data=data) + form_data["context_info"] = hot_words_str.strip() + + # Use streaming to handle NDJSON progress responses + async with client.stream( + "POST", + url, + headers=headers, + params=query, + files=files, + data=form_data, + ) as resp: + resp.raise_for_status() + content_type = resp.headers.get("content-type", "") + + if "application/x-ndjson" in content_type: + # Batch progress: read events line by line + data = None + async for line in resp.aiter_lines(): + line = line.strip() + if not line: + continue + event = json.loads(line) + if ( + event.get("type") == "progress" + and progress_callback + ): + progress_callback(event) + elif event.get("type") == "result": + data = event + if data is None: + raise RuntimeError( + f"NDJSON stream from '{self._name}' ended without a result event" + ) + else: + # Normal JSON response + await resp.aread() + data = resp.json() else: # Send as raw audio data (for Deepgram) - resp = await client.post(url, headers=headers, params=query, content=audio_data) + resp = await client.post( + url, headers=headers, params=query, content=audio_data + ) + resp.raise_for_status() + data = resp.json() else: resp = await client.get(url, headers=headers, params=query) - resp.raise_for_status() - data = resp.json() + resp.raise_for_status() + data = resp.json() except httpx.ConnectError as e: raise ConnectionError( f"Cannot reach transcription service '{self._name}' at {url}. " @@ -233,7 +324,9 @@ async def transcribe(self, audio_data: bytes, sample_rate: int, diarize: bool = channels = data["results"]["channels"] if channels and "alternatives" in channels[0]: alt = channels[0]["alternatives"][0] - logger.debug(f"DEBUG Registry: Deepgram alternative keys: {list(alt.keys())}") + logger.debug( + f"DEBUG Registry: Deepgram alternative keys: {list(alt.keys())}" + ) # Extract normalized shape text, words, segments = "", [], [] @@ -242,26 +335,36 @@ async def transcribe(self, audio_data: bytes, sample_rate: int, diarize: bool = text = _dotted_get(data, extract.get("text")) or "" words = _dotted_get(data, extract.get("words")) or [] segments = _dotted_get(data, extract.get("segments")) or [] + segments = _normalize_provider_segments(segments) # Check config to decide whether to keep or discard provider segments transcription_config = get_backend_config("transcription") - use_provider_segments = transcription_config.get("use_provider_segments", False) + use_provider_segments = transcription_config.get( + "use_provider_segments", False + ) if not use_provider_segments: segments = [] - logger.debug(f"Transcription: Extracted {len(words)} words, ignoring provider segments (use_provider_segments=false)") + logger.debug( + f"Transcription: Extracted {len(words)} words, ignoring provider segments (use_provider_segments=false)" + ) else: - logger.debug(f"Transcription: Extracted {len(words)} words, keeping {len(segments)} provider segments (use_provider_segments=true)") + logger.debug( + f"Transcription: Extracted {len(words)} words, keeping {len(segments)} provider segments (use_provider_segments=true)" + ) return {"text": text, "words": words, "segments": segments} + class RegistryStreamingTranscriptionProvider(StreamingTranscriptionProvider): """Streaming transcription provider using a config-driven WebSocket template.""" def __init__(self): registry = get_models_registry() if not registry: - raise RuntimeError("config.yml not found; cannot configure streaming STT provider") + raise RuntimeError( + "config.yml not found; cannot configure streaming STT provider" + ) model = registry.get_default("stt_stream") if not model: raise RuntimeError("No default stt_stream model defined in config.yml") @@ -281,9 +384,13 @@ def capabilities(self) -> set: async def transcribe(self, audio_data: bytes, sample_rate: int, **kwargs) -> dict: """Not used for streaming providers - use start_stream/process_audio_chunk/end_stream instead.""" - raise NotImplementedError("Streaming providers do not support batch transcription") + raise NotImplementedError( + "Streaming providers do not support batch transcription" + ) - async def start_stream(self, client_id: str, sample_rate: int = 16000, diarize: bool = False): + async def start_stream( + self, client_id: str, sample_rate: int = 16000, diarize: bool = False + ): base_url = self.model.model_url ops = self.model.operations or {} @@ -309,6 +416,9 @@ async def start_stream(self, client_id: str, sample_rate: int = 16000, diarize: except Exception as e: logger.debug(f"Failed to fetch asr.hot_words for streaming: {e}") + # NOTE: PULSE/wave (smallest.ai) does NOT support keywords on WebSocket ā€” + # any `keywords` query param causes 0 responses or HTTP 400. + # Normalize boolean values to lowercase strings (Deepgram expects "true"/"false", not "True"/"False") normalized_query = {} for k, v in query_dict.items(): @@ -351,9 +461,16 @@ async def start_stream(self, client_id: str, sample_rate: int = 16000, diarize: except Exception: pass - self._streams[client_id] = {"ws": ws, "sample_rate": sample_rate, "final": None, "interim": []} + self._streams[client_id] = { + "ws": ws, + "sample_rate": sample_rate, + "final": None, + "interim": [], + } - async def process_audio_chunk(self, client_id: str, audio_chunk: bytes) -> dict | None: + async def process_audio_chunk( + self, client_id: str, audio_chunk: bytes + ) -> dict | None: if client_id not in self._streams: return None ws = self._streams[client_id]["ws"] @@ -372,7 +489,7 @@ async def process_audio_chunk(self, client_id: str, audio_chunk: bytes) -> dict await ws.send(audio_chunk) # Non-blocking read for results - expect = (ops.get("expect", {}) or {}) + expect = ops.get("expect", {}) or {} extract = expect.get("extract", {}) interim_type = expect.get("interim_type") final_type = expect.get("final_type") @@ -392,16 +509,33 @@ async def process_audio_chunk(self, client_id: str, audio_chunk: bytes) -> dict # Fallback: check is_final directly (for providers that don't use a type field) is_final = data.get("is_final", False) - # Extract result data - text = _dotted_get(data, extract.get("text")) if extract.get("text") else data.get("text", "") - words = _dotted_get(data, extract.get("words")) if extract.get("words") else data.get("words", []) - segments = _dotted_get(data, extract.get("segments")) if extract.get("segments") else data.get("segments", []) + # Extract result data (guard against None from _dotted_get) + text = ( + _dotted_get(data, extract.get("text")) + if extract.get("text") + else data.get("text", "") + ) or "" + words = ( + _dotted_get(data, extract.get("words")) + if extract.get("words") + else data.get("words", []) + ) or [] + segments = ( + _dotted_get(data, extract.get("segments")) + if extract.get("segments") + else data.get("segments", []) + ) or [] + segments = _normalize_provider_segments(segments) # Calculate confidence if available confidence = data.get("confidence", 0.0) if not confidence and words and isinstance(words, list): # Calculate average word confidence - confidences = [w.get("confidence", 0.0) for w in words if isinstance(w, dict) and "confidence" in w] + confidences = [ + w.get("confidence", 0.0) + for w in words + if isinstance(w, dict) and "confidence" in w + ] if confidences: confidence = sum(confidences) / len(confidences) @@ -412,7 +546,7 @@ async def process_audio_chunk(self, client_id: str, audio_chunk: bytes) -> dict "words": words, "segments": segments, "is_final": is_final, - "confidence": confidence + "confidence": confidence, } except asyncio.TimeoutError: @@ -430,7 +564,7 @@ async def end_stream(self, client_id: str) -> dict: end_msg = (ops.get("end", {}) or {}).get("message", {"type": "stop"}) await ws.send(json.dumps(end_msg)) - expect = (ops.get("expect", {}) or {}) + expect = ops.get("expect", {}) or {} final_type = expect.get("final_type") extract = expect.get("extract", {}) @@ -457,14 +591,29 @@ async def end_stream(self, client_id: str) -> dict: if not isinstance(final, dict): return {"text": "", "words": [], "segments": []} + segments = ( + _dotted_get(final, extract.get("segments")) + if extract + else final.get("segments", []) + ) or [] return { - "text": _dotted_get(final, extract.get("text")) if extract else final.get("text", ""), - "words": _dotted_get(final, extract.get("words")) if extract else final.get("words", []), - "segments": _dotted_get(final, extract.get("segments")) if extract else final.get("segments", []), + "text": ( + _dotted_get(final, extract.get("text")) + if extract + else final.get("text", "") + ), + "words": ( + _dotted_get(final, extract.get("words")) + if extract + else final.get("words", []) + ), + "segments": _normalize_provider_segments(segments), } -def get_transcription_provider(provider_name: Optional[str] = None, mode: Optional[str] = None) -> Optional[BaseTranscriptionProvider]: +def get_transcription_provider( + provider_name: Optional[str] = None, mode: Optional[str] = None +) -> Optional[BaseTranscriptionProvider]: """Return a registry-driven transcription provider. - mode="batch": HTTP-based STT (default) @@ -503,7 +652,9 @@ def is_transcription_available(mode: str = "batch") -> bool: return provider is not None -def get_mock_transcription_provider(fail_mode: bool = False) -> BaseTranscriptionProvider: +def get_mock_transcription_provider( + fail_mode: bool = False, +) -> BaseTranscriptionProvider: """Return a mock transcription provider (for testing only). Args: @@ -513,6 +664,7 @@ def get_mock_transcription_provider(fail_mode: bool = False) -> BaseTranscriptio MockTranscriptionProvider instance """ from .mock_provider import MockTranscriptionProvider + return MockTranscriptionProvider(fail_mode=fail_mode) diff --git a/backends/advanced/src/advanced_omi_backend/services/transcription/mock_provider.py b/backends/advanced/src/advanced_omi_backend/services/transcription/mock_provider.py index f596214f..02e5b37c 100644 --- a/backends/advanced/src/advanced_omi_backend/services/transcription/mock_provider.py +++ b/backends/advanced/src/advanced_omi_backend/services/transcription/mock_provider.py @@ -33,7 +33,15 @@ def name(self) -> str: """Return the provider name for logging.""" return "mock" - async def transcribe(self, audio_data: bytes, sample_rate: int, diarize: bool = False) -> dict: + async def transcribe( + self, + audio_data: bytes, + sample_rate: int, + diarize: bool = False, + context_info=None, + progress_callback=None, + **kwargs, + ) -> dict: """ Return a predefined mock transcript or raise exception in fail mode. @@ -41,6 +49,9 @@ async def transcribe(self, audio_data: bytes, sample_rate: int, diarize: bool = audio_data: Raw audio bytes (ignored in mock) sample_rate: Audio sample rate (ignored in mock) diarize: Whether to enable speaker diarization (ignored in mock) + context_info: Optional ASR context (ignored in mock) + progress_callback: Optional callback for batch progress (ignored in mock) + **kwargs: Additional parameters (ignored in mock) Returns: Dictionary containing predefined transcript with words and segments @@ -88,20 +99,9 @@ async def transcribe(self, audio_data: bytes, sample_rate: int, diarize: bool = ] # Mock segments (single speaker for simplicity) - segments = [ - { - "speaker": 0, - "start": 0.0, - "end": 6.5, - "text": mock_transcript - } - ] + segments = [{"speaker": 0, "start": 0.0, "end": 6.5, "text": mock_transcript}] - return { - "text": mock_transcript, - "words": words, - "segments": segments if diarize else [] - } + return {"text": mock_transcript, "words": words, "segments": segments if diarize else []} async def connect(self, client_id: Optional[str] = None): """Initialize the mock provider (no-op).""" diff --git a/backends/advanced/src/advanced_omi_backend/workers/conversation_jobs.py b/backends/advanced/src/advanced_omi_backend/workers/conversation_jobs.py index b1bbb8cd..2142ce07 100644 --- a/backends/advanced/src/advanced_omi_backend/workers/conversation_jobs.py +++ b/backends/advanced/src/advanced_omi_backend/workers/conversation_jobs.py @@ -21,6 +21,7 @@ ) from advanced_omi_backend.controllers.session_controller import mark_session_complete from advanced_omi_backend.models.job import async_job +from advanced_omi_backend.observability.otel_setup import set_otel_session from advanced_omi_backend.plugins.events import PluginEvent from advanced_omi_backend.services.plugin_service import ( ensure_plugin_router, @@ -106,7 +107,9 @@ async def handle_end_of_conversation( from advanced_omi_backend.models.conversation import Conversation - conversation = await Conversation.find_one(Conversation.conversation_id == conversation_id) + conversation = await Conversation.find_one( + Conversation.conversation_id == conversation_id + ) if conversation: # Convert string to enum try: @@ -121,7 +124,9 @@ async def handle_end_of_conversation( f"šŸ’¾ Saved conversation {conversation_id[:12]} end_reason: {conversation.end_reason}" ) else: - logger.warning(f"āš ļø Conversation {conversation_id} not found for end reason tracking") + logger.warning( + f"āš ļø Conversation {conversation_id} not found for end reason tracking" + ) # Increment conversation count for this session conversation_count_key = f"session:conversation_count:{session_id}" @@ -134,7 +139,9 @@ async def handle_end_of_conversation( session_status = await redis_client.hget(session_key, "status") if session_status: status_str = ( - session_status.decode() if isinstance(session_status, bytes) else session_status + session_status.decode() + if isinstance(session_status, bytes) + else session_status ) if status_str == "active": @@ -257,7 +264,9 @@ async def open_conversation_job( conversation = None if existing_conversation_id_bytes: existing_conversation_id = existing_conversation_id_bytes.decode() - logger.info(f"šŸ” Found Redis key with conversation_id={existing_conversation_id}") + logger.info( + f"šŸ” Found Redis key with conversation_id={existing_conversation_id}" + ) # Try to fetch the existing conversation by conversation_id conversation = await Conversation.find_one( @@ -272,13 +281,16 @@ async def open_conversation_job( f"processing_status={processing_status}" ) else: - logger.warning(f"āš ļø Conversation {existing_conversation_id} not found in database!") + logger.warning( + f"āš ļø Conversation {existing_conversation_id} not found in database!" + ) # Verify it's a placeholder conversation (always_persist=True, processing_status='pending_transcription') if ( conversation and getattr(conversation, "always_persist", False) - and getattr(conversation, "processing_status", None) == "pending_transcription" + and getattr(conversation, "processing_status", None) + == "pending_transcription" ): logger.info( f"šŸ”„ Reusing placeholder conversation {conversation.conversation_id} for session {session_id}" @@ -297,7 +309,9 @@ async def open_conversation_job( ) conversation = None else: - logger.info(f"šŸ” No Redis key found for {conversation_key}, creating new conversation") + logger.info( + f"šŸ” No Redis key found for {conversation_key}, creating new conversation" + ) # If no valid placeholder found, create new conversation if not conversation: @@ -309,17 +323,23 @@ async def open_conversation_job( ) await conversation.insert() conversation_id = conversation.conversation_id - logger.info(f"āœ… Created streaming conversation {conversation_id} for session {session_id}") + logger.info( + f"āœ… Created streaming conversation {conversation_id} for session {session_id}" + ) # Attach markers from Redis session (e.g., button events captured during streaming) session_key = f"audio:session:{session_id}" markers_json = await redis_client.hget(session_key, "markers") if markers_json: try: - markers_data = markers_json if isinstance(markers_json, str) else markers_json.decode() + markers_data = ( + markers_json if isinstance(markers_json, str) else markers_json.decode() + ) conversation.markers = json.loads(markers_data) await conversation.save() - logger.info(f"šŸ“Œ Attached {len(conversation.markers)} markers to conversation {conversation_id}") + logger.info( + f"šŸ“Œ Attached {len(conversation.markers)} markers to conversation {conversation_id}" + ) except Exception as marker_err: logger.warning(f"āš ļø Failed to parse markers from Redis: {marker_err}") @@ -334,7 +354,9 @@ async def open_conversation_job( speaker_check_job_id = speech_job.meta.get("speaker_check_job_id") if speaker_check_job_id: try: - speaker_check_job = Job.fetch(speaker_check_job_id, connection=redis_conn) + speaker_check_job = Job.fetch( + speaker_check_job_id, connection=redis_conn + ) speaker_check_job.meta["conversation_id"] = conversation_id speaker_check_job.save_meta() except Exception as e: @@ -358,7 +380,9 @@ async def open_conversation_job( # Signal audio persistence job to rotate to this conversation's file rotation_signal_key = f"conversation:current:{session_id}" - await redis_client.set(rotation_signal_key, conversation_id, ex=86400) # 24 hour TTL + await redis_client.set( + rotation_signal_key, conversation_id, ex=86400 + ) # 24 hour TTL logger.info( f"šŸ”„ Signaled audio persistence to rotate file for conversation {conversation_id[:12]}" ) @@ -368,20 +392,26 @@ async def open_conversation_job( # Job control session_key = f"audio:session:{session_id}" - max_runtime = 10740 # 3 hours - 60 seconds (single conversations shouldn't exceed 3 hours) + max_runtime = ( + 10740 # 3 hours - 60 seconds (single conversations shouldn't exceed 3 hours) + ) start_time = time.time() last_result_count = 0 finalize_received = False # Inactivity timeout configuration - inactivity_timeout_seconds = float(os.getenv("SPEECH_INACTIVITY_THRESHOLD_SECONDS", "60")) + inactivity_timeout_seconds = float( + os.getenv("SPEECH_INACTIVITY_THRESHOLD_SECONDS", "60") + ) inactivity_timeout_minutes = inactivity_timeout_seconds / 60 last_meaningful_speech_time = ( 0.0 # Initialize with audio time 0 (will be updated with first speech) ) timeout_triggered = False # Track if closure was due to timeout - close_requested_reason = None # Track if closure was requested via API/plugin/button + close_requested_reason = ( + None # Track if closure was requested via API/plugin/button + ) last_inactivity_log_time = ( time.time() ) # Track when we last logged inactivity (wall-clock for logging) @@ -389,7 +419,9 @@ async def open_conversation_job( # Test mode: wait for audio queue to drain before timing out # In real usage, ambient noise keeps connection alive. In tests, chunks arrive in bursts. - wait_for_queue_drain = os.getenv("WAIT_FOR_AUDIO_QUEUE_DRAIN", "false").lower() == "true" + wait_for_queue_drain = ( + os.getenv("WAIT_FOR_AUDIO_QUEUE_DRAIN", "false").lower() == "true" + ) logger.info( f"šŸ“Š Conversation timeout configured: {inactivity_timeout_minutes} minutes ({inactivity_timeout_seconds}s)" @@ -413,7 +445,9 @@ async def open_conversation_job( finalize_received = True # Get completion reason (guaranteed to exist with unified API) - completion_reason = await redis_client.hget(session_key, "completion_reason") + completion_reason = await redis_client.hget( + session_key, "completion_reason" + ) completion_reason_str = ( completion_reason.decode() if completion_reason else "unknown" ) @@ -433,11 +467,19 @@ async def open_conversation_job( # Check for conversation close request (set by API, plugins, button press) if not finalize_received: - close_reason = await redis_client.hget(session_key, "conversation_close_requested") + close_reason = await redis_client.hget( + session_key, "conversation_close_requested" + ) if close_reason: await redis_client.hdel(session_key, "conversation_close_requested") - close_requested_reason = close_reason.decode() if isinstance(close_reason, bytes) else close_reason - logger.info(f"šŸ”’ Conversation close requested: {close_requested_reason}") + close_requested_reason = ( + close_reason.decode() + if isinstance(close_reason, bytes) + else close_reason + ) + logger.info( + f"šŸ”’ Conversation close requested: {close_requested_reason}" + ) timeout_triggered = True # Session stays active (same restart behavior as inactivity timeout) finalize_received = True break @@ -484,9 +526,12 @@ async def open_conversation_job( estimated_duration = len(text.split()) * 0.5 # ~0.5 seconds per word seg["end"] = start + estimated_duration - # Ensure speaker field exists - if "speaker" not in seg or not seg["speaker"]: + # Ensure speaker field exists and is a string + speaker = seg.get("speaker") + if speaker is None or speaker == "": seg["speaker"] = "SPEAKER_00" + elif isinstance(speaker, (int, float)): + seg["speaker"] = f"Speaker {int(speaker)}" validated_segments.append(seg) @@ -528,7 +573,9 @@ async def open_conversation_job( # Can't reliably detect inactivity, so skip timeout check this iteration inactivity_duration = 0 if speech_analysis.get("fallback", False): - logger.debug("āš ļø Skipping inactivity check (no audio timestamps available)") + logger.debug( + "āš ļø Skipping inactivity check (no audio timestamps available)" + ) current_time = time.time() @@ -651,7 +698,9 @@ async def open_conversation_job( else: end_reason = "user_stopped" - logger.info(f"šŸ“Š Conversation {conversation_id[:12]} end_reason determined: {end_reason}") + logger.info( + f"šŸ“Š Conversation {conversation_id[:12]} end_reason determined: {end_reason}" + ) # Wrap all post-processing in try/finally to guarantee handle_end_of_conversation() # is always called, even if an exception occurs during transcript saving, job @@ -686,7 +735,9 @@ async def open_conversation_job( f"āš ļø Streaming transcription ended with error for {session_id}, proceeding anyway" ) else: - logger.info(f"āœ… Streaming transcription confirmed complete for {session_id}") + logger.info( + f"āœ… Streaming transcription confirmed complete for {session_id}" + ) break await asyncio.sleep(0.5) waited_streaming += 0.5 @@ -725,14 +776,20 @@ async def open_conversation_job( end_reason=end_reason, ) - logger.info(f"šŸ“¦ MongoDB audio chunks ready for conversation {conversation_id[:12]}") + logger.info( + f"šŸ“¦ MongoDB audio chunks ready for conversation {conversation_id[:12]}" + ) # Get final streaming transcript and save to conversation - logger.info(f"šŸ“ Retrieving final streaming transcript for conversation {conversation_id[:12]}") + logger.info( + f"šŸ“ Retrieving final streaming transcript for conversation {conversation_id[:12]}" + ) final_transcript = await aggregator.get_combined_results(session_id) # Fetch conversation from database to ensure we have latest state - conversation = await Conversation.find_one(Conversation.conversation_id == conversation_id) + conversation = await Conversation.find_one( + Conversation.conversation_id == conversation_id + ) if not conversation: logger.error(f"āŒ Conversation {conversation_id} not found in database") raise ValueError(f"Conversation {conversation_id} not found") @@ -810,7 +867,8 @@ async def open_conversation_job( # Update placeholder conversation if it exists if ( getattr(conversation, "always_persist", False) - and getattr(conversation, "processing_status", None) == "pending_transcription" + and getattr(conversation, "processing_status", None) + == "pending_transcription" ): # Keep placeholder status - will be updated by title_summary_job logger.info( @@ -837,12 +895,13 @@ async def open_conversation_job( # Check if always_batch_retranscribe is enabled from advanced_omi_backend.config_loader import get_backend_config - transcription_cfg = get_backend_config('transcription') + transcription_cfg = get_backend_config("transcription") batch_retranscribe = False if transcription_cfg: from omegaconf import OmegaConf + cfg_dict = OmegaConf.to_container(transcription_cfg, resolve=True) - batch_retranscribe = cfg_dict.get('always_batch_retranscribe', False) + batch_retranscribe = cfg_dict.get("always_batch_retranscribe", False) if batch_retranscribe: # BATCH PATH: Streaming transcript saved as preview ā€” user sees it immediately @@ -852,7 +911,9 @@ async def open_conversation_job( JOB_RESULT_TTL, transcription_queue, ) - from advanced_omi_backend.workers.transcription_jobs import transcribe_full_audio_job + from advanced_omi_backend.workers.transcription_jobs import ( + transcribe_full_audio_job, + ) batch_version_id = f"batch_{conversation_id[:12]}" batch_job = transcription_queue.enqueue( @@ -864,7 +925,7 @@ async def open_conversation_job( result_ttl=JOB_RESULT_TTL, job_id=f"batch_retranscribe_{conversation_id[:12]}", description=f"Batch re-transcription for {conversation_id[:8]}", - meta={'conversation_id': conversation_id, 'client_id': client_id}, + meta={"conversation_id": conversation_id, "client_id": client_id}, ) logger.info( @@ -948,7 +1009,9 @@ async def open_conversation_job( @async_job(redis=True, beanie=True) -async def generate_title_summary_job(conversation_id: str, *, redis_client=None) -> Dict[str, Any]: +async def generate_title_summary_job( + conversation_id: str, *, redis_client=None +) -> Dict[str, Any]: """ Generate title, short summary, and detailed summary for a conversation using LLM. @@ -971,12 +1034,17 @@ async def generate_title_summary_job(conversation_id: str, *, redis_client=None) generate_title_and_summary, ) - logger.info(f"šŸ“ Starting title/summary generation for conversation {conversation_id}") + set_otel_session(conversation_id) + logger.info( + f"šŸ“ Starting title/summary generation for conversation {conversation_id}" + ) start_time = time.time() # Get the conversation - conversation = await Conversation.find_one(Conversation.conversation_id == conversation_id) + conversation = await Conversation.find_one( + Conversation.conversation_id == conversation_id + ) if not conversation: logger.error(f"Conversation {conversation_id} not found") return {"success": False, "error": "Conversation not found"} @@ -986,7 +1054,9 @@ async def generate_title_summary_job(conversation_id: str, *, redis_client=None) segments = conversation.segments or [] if not transcript_text and (not segments or len(segments) == 0): - logger.warning(f"āš ļø No transcript or segments available for conversation {conversation_id}") + logger.warning( + f"āš ļø No transcript or segments available for conversation {conversation_id}" + ) return { "success": False, "error": "No transcript or segments available", @@ -1018,18 +1088,24 @@ async def generate_title_summary_job(conversation_id: str, *, redis_client=None) else: logger.info(f"šŸ“š No memories found for context enrichment") except Exception as mem_error: - logger.warning(f"āš ļø Could not fetch memory context (continuing without): {mem_error}") + logger.warning( + f"āš ļø Could not fetch memory context (continuing without): {mem_error}" + ) # Generate title+summary (one call) and detailed summary in parallel import asyncio (title, short_summary), detailed_summary = await asyncio.gather( generate_title_and_summary( - transcript_text, segments=segments, user_id=conversation.user_id, + transcript_text, + segments=segments, + user_id=conversation.user_id, langfuse_session_id=conversation_id, ), generate_detailed_summary( - transcript_text, segments=segments, memory_context=memory_context, + transcript_text, + segments=segments, + memory_context=memory_context, langfuse_session_id=conversation_id, ), ) @@ -1040,10 +1116,15 @@ async def generate_title_summary_job(conversation_id: str, *, redis_client=None) logger.info(f"āœ… Generated title: '{conversation.title}'") logger.info(f"āœ… Generated summary: '{conversation.summary}'") - logger.info(f"āœ… Generated detailed summary: {len(conversation.detailed_summary)} chars") + logger.info( + f"āœ… Generated detailed summary: {len(conversation.detailed_summary)} chars" + ) # Update processing status for placeholder/reprocessing conversations - if getattr(conversation, "processing_status", None) in ["pending_transcription", "reprocessing"]: + if getattr(conversation, "processing_status", None) in [ + "pending_transcription", + "reprocessing", + ]: conversation.processing_status = "completed" logger.info( f"āœ… Updated placeholder conversation {conversation_id} " @@ -1054,7 +1135,10 @@ async def generate_title_summary_job(conversation_id: str, *, redis_client=None) logger.error(f"āŒ Title/summary generation failed: {gen_error}") # Mark placeholder/reprocessing conversation as failed - if getattr(conversation, "processing_status", None) in ["pending_transcription", "reprocessing"]: + if getattr(conversation, "processing_status", None) in [ + "pending_transcription", + "reprocessing", + ]: conversation.title = "Audio Recording (Transcription Failed)" conversation.summary = f"Title/summary generation failed: {str(gen_error)}" conversation.processing_status = "transcription_failed" @@ -1089,7 +1173,9 @@ async def generate_title_summary_job(conversation_id: str, *, redis_client=None) "title": conversation.title, "summary": conversation.summary, "detailed_summary_length": ( - len(conversation.detailed_summary) if conversation.detailed_summary else 0 + len(conversation.detailed_summary) + if conversation.detailed_summary + else 0 ), "segment_count": len(segments), "processing_time": processing_time, @@ -1140,12 +1226,16 @@ async def dispatch_conversation_complete_event_job( """ from advanced_omi_backend.models.conversation import Conversation - logger.info(f"šŸ“Œ Dispatching conversation.complete event for conversation {conversation_id}") + logger.info( + f"šŸ“Œ Dispatching conversation.complete event for conversation {conversation_id}" + ) start_time = time.time() # Get the conversation to include in event data - conversation = await Conversation.find_one(Conversation.conversation_id == conversation_id) + conversation = await Conversation.find_one( + Conversation.conversation_id == conversation_id + ) if not conversation: logger.error(f"Conversation {conversation_id} not found") return {"success": False, "error": "Conversation not found"} @@ -1222,9 +1312,13 @@ async def dispatch_conversation_complete_event_job( f"šŸ”Œ RESULT: conversation.complete dispatched to {len(plugin_results) if plugin_results else 0} plugins" ) if plugin_results: - logger.info(f"šŸ“Œ Triggered {len(plugin_results)} conversation-level plugins") + logger.info( + f"šŸ“Œ Triggered {len(plugin_results)} conversation-level plugins" + ) for result in plugin_results: - logger.info(f" Plugin result: success={result.success}, message={result.message}") + logger.info( + f" Plugin result: success={result.success}, message={result.message}" + ) if result.message: logger.info(f" Plugin result: {result.message}") diff --git a/backends/advanced/src/advanced_omi_backend/workers/memory_jobs.py b/backends/advanced/src/advanced_omi_backend/workers/memory_jobs.py index 53f0a103..492dc650 100644 --- a/backends/advanced/src/advanced_omi_backend/workers/memory_jobs.py +++ b/backends/advanced/src/advanced_omi_backend/workers/memory_jobs.py @@ -22,6 +22,10 @@ memory_queue, ) from advanced_omi_backend.models.job import JobPriority, async_job +from advanced_omi_backend.observability.otel_setup import ( + clear_otel_session, + set_otel_session, +) from advanced_omi_backend.plugins.events import PluginEvent from advanced_omi_backend.services.plugin_service import ensure_plugin_router @@ -110,7 +114,9 @@ def compute_speaker_diff( @async_job(redis=True, beanie=True) -async def process_memory_job(conversation_id: str, *, redis_client=None) -> Dict[str, Any]: +async def process_memory_job( + conversation_id: str, *, redis_client=None +) -> Dict[str, Any]: """ RQ job function for memory extraction and processing from conversations. @@ -135,6 +141,7 @@ async def process_memory_job(conversation_id: str, *, redis_client=None) -> Dict from advanced_omi_backend.services.memory import get_memory_service from advanced_omi_backend.users import get_user_by_id + set_otel_session(conversation_id) start_time = time.time() logger.info(f"šŸ”„ Starting memory processing for conversation {conversation_id}") @@ -169,11 +176,13 @@ async def process_memory_job(conversation_id: str, *, redis_client=None) -> Dict for segment in segments: text = segment.text.strip() speaker = segment.speaker - seg_type = getattr(segment, 'segment_type', 'speech') + seg_type = getattr(segment, "segment_type", "speech") if text: if seg_type == "event": # Non-speech event: include as context marker without speaker prefix - dialogue_lines.append(f"[{text}]" if not text.startswith("[") else text) + dialogue_lines.append( + f"[{text}]" if not text.startswith("[") else text + ) elif seg_type == "note": # User-inserted note: include as distinct context dialogue_lines.append(f"[Note: {text}]") @@ -197,14 +206,20 @@ async def process_memory_job(conversation_id: str, *, redis_client=None) -> Dict full_conversation = conversation_model.transcript if len(full_conversation) < MIN_CONVERSATION_LENGTH: - logger.warning(f"Conversation too short for memory processing: {conversation_id}") + logger.warning( + f"Conversation too short for memory processing: {conversation_id}" + ) return {"success": False, "error": "Conversation too short"} # Check primary speakers filter (reuse `user` from above ā€” no duplicate DB call) if user and user.primary_speakers: - primary_speaker_names = {ps["name"].strip().lower() for ps in user.primary_speakers} + primary_speaker_names = { + ps["name"].strip().lower() for ps in user.primary_speakers + } - if transcript_speakers and not transcript_speakers.intersection(primary_speaker_names): + if transcript_speakers and not transcript_speakers.intersection( + primary_speaker_names + ): logger.info( f"Skipping memory - no primary speakers found in conversation {conversation_id}" ) @@ -301,11 +316,18 @@ async def process_memory_job(conversation_id: str, *, redis_client=None) -> Dict # Fetch memory details to display in UI memory_details = [] try: - for memory_id in created_memory_ids[:5]: # Limit to first 5 for display - memory_entry = await memory_service.get_memory(memory_id, user_id) + for memory_id in created_memory_ids[ + :5 + ]: # Limit to first 5 for display + memory_entry = await memory_service.get_memory( + memory_id, user_id + ) if memory_entry: memory_details.append( - {"memory_id": memory_id, "text": memory_entry.content[:200]} + { + "memory_id": memory_id, + "text": memory_entry.content[:200], + } ) except Exception as e: logger.warning(f"Failed to fetch memory details for UI: {e}") @@ -335,7 +357,9 @@ async def process_memory_job(conversation_id: str, *, redis_client=None) -> Dict config = get_config() kg_enabled = ( - config.get("memory", {}).get("knowledge_graph", {}).get("enabled", False) + config.get("memory", {}) + .get("knowledge_graph", {}) + .get("enabled", False) ) if kg_enabled: @@ -379,7 +403,9 @@ async def process_memory_job(conversation_id: str, *, redis_client=None) -> Dict "user_id": user_id, "user_email": user_email, }, - "memory_count": len(created_memory_ids) if created_memory_ids else 0, + "memory_count": ( + len(created_memory_ids) if created_memory_ids else 0 + ), "conversation_id": conversation_id, } @@ -403,7 +429,9 @@ async def process_memory_job(conversation_id: str, *, redis_client=None) -> Dict ) if plugin_results: - logger.info(f"šŸ“Œ Triggered {len(plugin_results)} memory-level plugins") + logger.info( + f"šŸ“Œ Triggered {len(plugin_results)} memory-level plugins" + ) for result in plugin_results: if result.message: logger.info(f" Plugin result: {result.message}") @@ -413,7 +441,9 @@ async def process_memory_job(conversation_id: str, *, redis_client=None) -> Dict return { "success": True, - "memories_created": len(created_memory_ids) if created_memory_ids else 0, + "memories_created": ( + len(created_memory_ids) if created_memory_ids else 0 + ), "processing_time": processing_time, } else: @@ -461,7 +491,11 @@ async def _process_speaker_reprocess( f"falling back to normal extraction" ) return await memory_service.add_memory( - full_conversation, client_id, conversation_id, user_id, user_email, + full_conversation, + client_id, + conversation_id, + user_id, + user_email, allow_update=True, ) @@ -474,7 +508,11 @@ async def _process_speaker_reprocess( f"for {conversation_id}, falling back to normal extraction" ) return await memory_service.add_memory( - full_conversation, client_id, conversation_id, user_id, user_email, + full_conversation, + client_id, + conversation_id, + user_id, + user_email, allow_update=True, ) @@ -491,7 +529,11 @@ async def _process_speaker_reprocess( f"for {conversation_id}, falling back to normal extraction" ) return await memory_service.add_memory( - full_conversation, client_id, conversation_id, user_id, user_email, + full_conversation, + client_id, + conversation_id, + user_id, + user_email, allow_update=True, ) @@ -507,7 +549,11 @@ async def _process_speaker_reprocess( f"for {conversation_id}, falling back to normal extraction" ) return await memory_service.add_memory( - full_conversation, client_id, conversation_id, user_id, user_email, + full_conversation, + client_id, + conversation_id, + user_id, + user_email, allow_update=True, ) @@ -564,5 +610,7 @@ def enqueue_memory_processing( description=f"Process memory for conversation {conversation_id[:8]}", ) - logger.info(f"šŸ“„ RQ: Enqueued memory job {job.id} for conversation {conversation_id}") + logger.info( + f"šŸ“„ RQ: Enqueued memory job {job.id} for conversation {conversation_id}" + ) return job diff --git a/backends/advanced/src/advanced_omi_backend/workers/rq_worker_entry.py b/backends/advanced/src/advanced_omi_backend/workers/rq_worker_entry.py index d9da1c6a..a3c07fbc 100755 --- a/backends/advanced/src/advanced_omi_backend/workers/rq_worker_entry.py +++ b/backends/advanced/src/advanced_omi_backend/workers/rq_worker_entry.py @@ -14,7 +14,7 @@ logging.basicConfig( level=logging.INFO, format="%(asctime)s [%(levelname)s] %(name)s: %(message)s", - stream=sys.stdout + stream=sys.stdout, ) logger = logging.getLogger(__name__) @@ -22,14 +22,24 @@ def main(): """Start RQ worker with proper logging configuration.""" + # Initialize OTEL/Galileo if configured (patches OpenAI before any job imports) + try: + from advanced_omi_backend.observability.otel_setup import init_otel + + init_otel() + except Exception: + pass # Optional ā€” don't block workers + from redis import Redis from rq import Worker # Get Redis URL from environment - redis_url = os.getenv('REDIS_URL', 'redis://localhost:6379/0') + redis_url = os.getenv("REDIS_URL", "redis://localhost:6379/0") # Get queue names from command line arguments - queue_names = sys.argv[1:] if len(sys.argv) > 1 else ['transcription', 'memory', 'default'] + queue_names = ( + sys.argv[1:] if len(sys.argv) > 1 else ["transcription", "memory", "default"] + ) logger.info(f"šŸš€ Starting RQ worker for queues: {', '.join(queue_names)}") logger.info(f"šŸ“” Redis URL: {redis_url}") @@ -38,16 +48,12 @@ def main(): redis_conn = Redis.from_url(redis_url) # Create and start worker - worker = Worker( - queue_names, - connection=redis_conn, - log_job_description=True - ) + worker = Worker(queue_names, connection=redis_conn, log_job_description=True) logger.info("āœ… RQ worker ready") # This blocks until worker is stopped - worker.work(logging_level='INFO') + worker.work(logging_level="INFO") if __name__ == "__main__": diff --git a/backends/advanced/src/advanced_omi_backend/workers/transcription_jobs.py b/backends/advanced/src/advanced_omi_backend/workers/transcription_jobs.py index 5d6e592e..fb644ec1 100644 --- a/backends/advanced/src/advanced_omi_backend/workers/transcription_jobs.py +++ b/backends/advanced/src/advanced_omi_backend/workers/transcription_jobs.py @@ -21,7 +21,10 @@ from rq.exceptions import NoSuchJobError from rq.job import Job -from advanced_omi_backend.config import get_backend_config, get_transcription_job_timeout +from advanced_omi_backend.config import ( + get_backend_config, + get_transcription_job_timeout, +) from advanced_omi_backend.controllers.queue_controller import ( JOB_RESULT_TTL, REDIS_URL, @@ -32,8 +35,8 @@ from advanced_omi_backend.models.audio_chunk import AudioChunkDocument from advanced_omi_backend.models.conversation import Conversation from advanced_omi_backend.models.job import BaseRQJob, JobPriority, async_job -from advanced_omi_backend.services.audio_stream import TranscriptionResultsAggregator from advanced_omi_backend.plugins.events import PluginEvent +from advanced_omi_backend.services.audio_stream import TranscriptionResultsAggregator from advanced_omi_backend.services.plugin_service import ensure_plugin_router from advanced_omi_backend.services.transcription import ( get_transcription_provider, @@ -82,7 +85,9 @@ async def apply_speaker_recognition( speaker_client = SpeakerRecognitionClient() if not speaker_client.enabled: - logger.info(f"šŸŽ¤ Speaker recognition disabled, using original speaker labels") + logger.info( + f"šŸŽ¤ Speaker recognition disabled, using original speaker labels" + ) return segments logger.info( @@ -123,7 +128,9 @@ def get_speaker_at_time(timestamp: float, speaker_segments: list) -> str: updated_count = 0 for seg in segments: seg_mid = (seg.start + seg.end) / 2.0 - identified_speaker = get_speaker_at_time(seg_mid, speaker_identified_segments) + identified_speaker = get_speaker_at_time( + seg_mid, speaker_identified_segments + ) if identified_speaker and identified_speaker != "Unknown": original_speaker = seg.speaker @@ -186,7 +193,9 @@ async def transcribe_full_audio_job( start_time = time.time() # Get the conversation - conversation = await Conversation.find_one(Conversation.conversation_id == conversation_id) + conversation = await Conversation.find_one( + Conversation.conversation_id == conversation_id + ) if not conversation: raise ValueError(f"Conversation {conversation_id} not found") @@ -203,18 +212,23 @@ async def transcribe_full_audio_job( logger.info(f"Using transcription provider: {provider_name}") # Reconstruct audio from MongoDB chunks - logger.info(f"šŸ“¦ Reconstructing audio from MongoDB chunks for conversation {conversation_id}") + logger.info( + f"šŸ“¦ Reconstructing audio from MongoDB chunks for conversation {conversation_id}" + ) try: # Reconstruct WAV from MongoDB chunks (already in memory as bytes) wav_data = await reconstruct_wav_from_conversation(conversation_id) logger.info( - f"šŸ“¦ Reconstructed audio from MongoDB chunks: " f"{len(wav_data) / 1024 / 1024:.2f} MB" + f"šŸ“¦ Reconstructed audio from MongoDB chunks: " + f"{len(wav_data) / 1024 / 1024:.2f} MB" ) except ValueError as e: # No chunks found for conversation - raise FileNotFoundError(f"No audio chunks found for conversation {conversation_id}: {e}") + raise FileNotFoundError( + f"No audio chunks found for conversation {conversation_id}: {e}" + ) except Exception as e: logger.error(f"Failed to reconstruct audio from MongoDB: {e}", exc_info=True) raise RuntimeError(f"Audio reconstruction failed: {e}") @@ -236,11 +250,27 @@ async def transcribe_full_audio_job( actual_sample_rate = 16000 try: + # Progress callback: writes batch progress to RQ job.meta so the + # queue API and UI can show "Transcribing segment X of Y". + def _on_batch_progress(event: dict) -> None: + job = get_current_job() + if job: + current = event.get("current", 0) + total = event.get("total", 0) + job.meta["batch_progress"] = { + "current": current, + "total": total, + "percent": int(current / total * 100) if total else 0, + "message": f"Transcribing segment {current} of {total}", + } + job.save_meta() + # Transcribe the audio directly from memory (no disk I/O needed) transcribe_kwargs: Dict[str, Any] = { "audio_data": wav_data, "sample_rate": actual_sample_rate, "diarize": True, + "progress_callback": _on_batch_progress, } if context_info: transcribe_kwargs["context_info"] = context_info @@ -308,7 +338,9 @@ async def transcribe_full_audio_job( if result.message: logger.info(f" Plugin: {result.message}") except Exception as e: - logger.exception(f"āš ļø Error triggering transcript plugins in batch mode: {e}") + logger.exception( + f"āš ļø Error triggering transcript plugins in batch mode: {e}" + ) logger.info(f"šŸ” DEBUG: Plugin processing complete, moving to speech validation") @@ -363,7 +395,9 @@ async def transcribe_full_audio_job( f"Job {job_id} hash not found (likely already completed or expired)" ) else: - logger.debug(f"Job {job_id} not found or already completed: {e}") + logger.debug( + f"Job {job_id} not found or already completed: {e}" + ) if cancelled_jobs: logger.info( @@ -586,7 +620,10 @@ async def create_audio_only_conversation( placeholder_conversation = await Conversation.find_one( Conversation.client_id == session_id, Conversation.always_persist == True, - In(Conversation.processing_status, ["pending_transcription", "transcription_failed"]), + In( + Conversation.processing_status, + ["pending_transcription", "transcription_failed"], + ), ) if placeholder_conversation: @@ -597,7 +634,9 @@ async def create_audio_only_conversation( # Update status to show batch transcription is starting placeholder_conversation.processing_status = "batch_transcription" placeholder_conversation.title = "Audio Recording (Batch Transcription...)" - placeholder_conversation.summary = "Processing audio with offline transcription..." + placeholder_conversation.summary = ( + "Processing audio with offline transcription..." + ) await placeholder_conversation.save() # Audio chunks are already linked to this conversation_id @@ -624,13 +663,20 @@ async def create_audio_only_conversation( ) await conversation.insert() - logger.info(f"āœ… Created batch transcription conversation {session_id[:12]} for fallback") + logger.info( + f"āœ… Created batch transcription conversation {session_id[:12]} for fallback" + ) return conversation @async_job(redis=True, beanie=True) async def transcription_fallback_check_job( - session_id: str, user_id: str, client_id: str, timeout_seconds: int = None, *, redis_client=None + session_id: str, + user_id: str, + client_id: str, + timeout_seconds: int = None, + *, + redis_client=None, ) -> Dict[str, Any]: """ Check if streaming transcription succeeded, fallback to batch if needed. @@ -765,14 +811,18 @@ async def transcription_fallback_check_job( sample_rate, channels, sample_width = 16000, 1, 2 session_key = f"audio:session:{session_id}" try: - audio_format_raw = await redis_client.hget(session_key, "audio_format") + audio_format_raw = await redis_client.hget( + session_key, "audio_format" + ) if audio_format_raw: audio_format = json.loads(audio_format_raw) sample_rate = int(audio_format.get("rate", 16000)) channels = int(audio_format.get("channels", 1)) sample_width = int(audio_format.get("width", 2)) except Exception as e: - logger.warning(f"Failed to read audio_format from Redis for {session_id}: {e}") + logger.warning( + f"Failed to read audio_format from Redis for {session_id}: {e}" + ) bytes_per_second = sample_rate * channels * sample_width logger.info( @@ -781,7 +831,9 @@ async def transcription_fallback_check_job( ) # Create conversation placeholder - conversation = await create_audio_only_conversation(session_id, user_id, client_id) + conversation = await create_audio_only_conversation( + session_id, user_id, client_id + ) # Save audio to MongoDB chunks for batch transcription num_chunks = await convert_audio_to_chunks( @@ -798,7 +850,9 @@ async def transcription_fallback_check_job( ) except Exception as e: - logger.error(f"āŒ Failed to extract audio from Redis stream: {e}", exc_info=True) + logger.error( + f"āŒ Failed to extract audio from Redis stream: {e}", exc_info=True + ) raise else: logger.info( @@ -807,7 +861,9 @@ async def transcription_fallback_check_job( ) # Create conversation placeholder for batch transcription - conversation = await create_audio_only_conversation(session_id, user_id, client_id) + conversation = await create_audio_only_conversation( + session_id, user_id, client_id + ) # Enqueue batch transcription job version_id = f"batch_fallback_{session_id[:12]}" @@ -903,10 +959,14 @@ async def stream_speech_detection_job( # Get conversation count conversation_count_key = f"session:conversation_count:{session_id}" conversation_count_bytes = await redis_client.get(conversation_count_key) - conversation_count = int(conversation_count_bytes) if conversation_count_bytes else 0 + conversation_count = ( + int(conversation_count_bytes) if conversation_count_bytes else 0 + ) # Check if speaker filtering is enabled - speaker_filter_enabled = os.getenv("RECORD_ONLY_ENROLLED_SPEAKERS", "false").lower() == "true" + speaker_filter_enabled = ( + os.getenv("RECORD_ONLY_ENROLLED_SPEAKERS", "false").lower() == "true" + ) logger.info( f"šŸ“Š Conversation #{conversation_count + 1}, Speaker filter: {'enabled' if speaker_filter_enabled else 'disabled'}" ) @@ -942,7 +1002,10 @@ async def stream_speech_detection_job( # Check if session has closed session_status = await redis_client.hget(session_key, "status") - session_closed = session_status and session_status.decode() in ["finalizing", "finished"] + session_closed = session_status and session_status.decode() in [ + "finalizing", + "finished", + ] if session_closed and session_closed_at is None: # Session just closed - start grace period for final transcription @@ -952,10 +1015,30 @@ async def stream_speech_detection_job( ) # Exit if grace period expired without speech - if session_closed_at and (time.time() - session_closed_at) > final_check_grace_period: + if ( + session_closed_at + and (time.time() - session_closed_at) > final_check_grace_period + ): logger.info(f"āœ… Session ended without speech (grace period expired)") break + # Consume any stale conversation close request (defensive ā€” shouldn't normally + # appear since services.py gates on conversation:current, but handles race conditions) + close_reason = await redis_client.hget( + session_key, "conversation_close_requested" + ) + if close_reason: + await redis_client.hdel(session_key, "conversation_close_requested") + close_reason_str = ( + close_reason.decode() + if isinstance(close_reason, bytes) + else close_reason + ) + logger.info( + f"šŸ”’ Conversation close requested ({close_reason_str}) during speech detection ā€” " + f"no open conversation to close, flag consumed" + ) + if time.time() - start_time > max_runtime: logger.warning(f"ā±ļø Max runtime reached, exiting") break @@ -966,11 +1049,15 @@ async def stream_speech_detection_job( # Health check: detect transcription errors early during grace period if session_closed_at: # Check for streaming consumer errors in session metadata - error_status = await redis_client.hget(session_key, "transcription_error") + error_status = await redis_client.hget( + session_key, "transcription_error" + ) if error_status: error_msg = error_status.decode() logger.error(f"āŒ Transcription service error: {error_msg}") - logger.error(f"āŒ Session failed - transcription service unavailable") + logger.error( + f"āŒ Session failed - transcription service unavailable" + ) break # Check if we've been waiting too long with no results at all @@ -980,7 +1067,9 @@ async def stream_speech_detection_job( logger.error( f"āŒ No transcription activity after {grace_elapsed:.1f}s - possible API key or connectivity issue" ) - logger.error(f"āŒ Session failed - check transcription service configuration") + logger.error( + f"āŒ Session failed - check transcription service configuration" + ) break await asyncio.sleep(2) @@ -1016,9 +1105,13 @@ async def stream_speech_detection_job( from datetime import datetime await redis_client.hset( - session_key, "last_event", f"speech_detected:{datetime.utcnow().isoformat()}" + session_key, + "last_event", + f"speech_detected:{datetime.utcnow().isoformat()}", + ) + await redis_client.hset( + session_key, "speech_detected_at", datetime.utcnow().isoformat() ) - await redis_client.hset(session_key, "speech_detected_at", datetime.utcnow().isoformat()) # Step 2: If speaker filter enabled, check for enrolled speakers identified_speakers = [] @@ -1028,7 +1121,9 @@ async def stream_speech_detection_job( # Add session event for speaker check starting await redis_client.hset( - session_key, "last_event", f"speaker_check_starting:{datetime.utcnow().isoformat()}" + session_key, + "last_event", + f"speaker_check_starting:{datetime.utcnow().isoformat()}", ) await redis_client.hset(session_key, "speaker_check_status", "checking") from .speaker_jobs import check_enrolled_speakers_job @@ -1068,7 +1163,9 @@ async def stream_speech_detection_job( result = speaker_check_job.result enrolled_present = result.get("enrolled_present", False) identified_speakers = result.get("identified_speakers", []) - logger.info(f"āœ… Speaker check completed: enrolled={enrolled_present}") + logger.info( + f"āœ… Speaker check completed: enrolled={enrolled_present}" + ) # Update session event for speaker check complete await redis_client.hset( @@ -1083,7 +1180,9 @@ async def stream_speech_detection_job( ) if identified_speakers: await redis_client.hset( - session_key, "identified_speakers", ",".join(identified_speakers) + session_key, + "identified_speakers", + ",".join(identified_speakers), ) break elif speaker_check_job.is_failed: @@ -1095,7 +1194,9 @@ async def stream_speech_detection_job( "last_event", f"speaker_check_failed:{datetime.utcnow().isoformat()}", ) - await redis_client.hset(session_key, "speaker_check_status", "failed") + await redis_client.hset( + session_key, "speaker_check_status", "failed" + ) break await asyncio.sleep(poll_interval) waited += poll_interval @@ -1148,7 +1249,9 @@ async def stream_speech_detection_job( ) # Track the job - await redis_client.set(open_job_key, open_job.id, ex=10800) # 3 hours to match job timeout + await redis_client.set( + open_job_key, open_job.id, ex=10800 + ) # 3 hours to match job timeout # Store metadata in speech detection job if current_job: @@ -1161,23 +1264,31 @@ async def stream_speech_detection_job( current_job.meta.update( { "conversation_job_id": open_job.id, - "speaker_check_job_id": speaker_check_job.id if speaker_check_job else None, + "speaker_check_job_id": ( + speaker_check_job.id if speaker_check_job else None + ), "detected_speakers": identified_speakers, - "speech_detected_at": datetime.fromtimestamp(speech_detected_at).isoformat(), + "speech_detected_at": datetime.fromtimestamp( + speech_detected_at + ).isoformat(), "session_id": session_id, "client_id": client_id, # For job grouping } ) current_job.save_meta() - logger.info(f"āœ… Started conversation job {open_job.id}, exiting speech detection") + logger.info( + f"āœ… Started conversation job {open_job.id}, exiting speech detection" + ) return { "session_id": session_id, "user_id": user_id, "client_id": client_id, "conversation_job_id": open_job.id, - "speech_detected_at": datetime.fromtimestamp(speech_detected_at).isoformat(), + "speech_detected_at": datetime.fromtimestamp( + speech_detected_at + ).isoformat(), "runtime_seconds": time.time() - start_time, } @@ -1205,7 +1316,9 @@ async def stream_speech_detection_job( # Check if this is an always_persist conversation that needs to be marked as failed # NOTE: We check MongoDB directly because the conversation:current Redis key might have been # deleted by the audio persistence job cleanup (which runs in parallel). - logger.info(f"šŸ” Checking MongoDB for always_persist conversation with client_id: {client_id}") + logger.info( + f"šŸ” Checking MongoDB for always_persist conversation with client_id: {client_id}" + ) # Find conversation by client_id that matches this session # session_id == client_id for streaming sessions (set in _initialize_streaming_session) diff --git a/config/config.yml.template b/config/config.yml.template index c4aaa5ae..e292a7f3 100644 --- a/config/config.yml.template +++ b/config/config.yml.template @@ -174,6 +174,11 @@ models: method: POST path: /transcribe content_type: multipart/form-data + timeout: 300 + # Per-read timeout for NDJSON batch progress responses. + # Each batch window can take minutes; this timeout covers + # the gap between successive progress lines. + read_timeout: 600 response: type: json extract: @@ -319,12 +324,15 @@ models: api_family: websocket model_url: wss://api.deepgram.com/v1/listen api_key: ${oc.env:DEEPGRAM_API_KEY,''} + capabilities: + - diarization operations: query: model: nova-3 language: multi smart_format: 'true' punctuate: 'true' + diarize: 'true' encoding: linear16 sample_rate: 16000 channels: '1' diff --git a/config/defaults.yml b/config/defaults.yml index a631f486..b32ec1c1 100644 --- a/config/defaults.yml +++ b/config/defaults.yml @@ -207,12 +207,15 @@ models: api_family: websocket model_url: wss://api.deepgram.com/v1/listen api_key: ${oc.env:DEEPGRAM_API_KEY,''} + capabilities: + - diarization operations: query: model: nova-3 language: multi smart_format: 'true' punctuate: 'true' + diarize: 'true' encoding: linear16 sample_rate: 16000 channels: '1' @@ -298,6 +301,7 @@ models: sample_rate: 16000 word_timestamps: 'true' diarize: 'true' + sentence_timestamps: 'true' end: message: type: finalize diff --git a/config/plugins.yml.template b/config/plugins.yml.template index 789cd9ed..2eec468d 100644 --- a/config/plugins.yml.template +++ b/config/plugins.yml.template @@ -27,10 +27,9 @@ plugins: # - transcript.batch # Uncomment to also handle batch transcription # - conversation.complete # Uncomment to handle completed conversations condition: - type: wake_word - wake_words: # Support multiple wake words - - hey vivi # Example: "hey vivi, turn off the lights" - - hey jarvis # Example: "hey jarvis, what's the temperature" + type: keyword_anywhere # Trigger when keyword appears anywhere in transcript + keywords: # Support multiple keywords + - vivi # Example: "turn off the lights, vivi" ha_url: http://host.docker.internal:8123 # Your Home Assistant URL ha_token: ${HA_TOKEN} # ALWAYS use env var - never paste actual token here! # To get a long-lived token: diff --git a/extras/asr-services/tests/test_qwen3_asr_parsing.py b/extras/asr-services/tests/test_qwen3_asr_parsing.py new file mode 100644 index 00000000..1b8f4c2f --- /dev/null +++ b/extras/asr-services/tests/test_qwen3_asr_parsing.py @@ -0,0 +1,125 @@ +""" +Tests for Qwen3-ASR output parsing and repetition detection. + +Pure function tests ā€” no GPU, no vLLM, no network required. + +Run: + cd extras/asr-services + uv run pytest tests/test_qwen3_asr_parsing.py -v +""" + +import sys +from pathlib import Path + +import pytest + +sys.path.insert(0, str(Path(__file__).resolve().parent.parent)) + +from providers.qwen3_asr.transcriber import ( + _parse_qwen3_output, + detect_and_fix_repetitions, +) + +# --------------------------------------------------------------------------- +# _parse_qwen3_output tests +# --------------------------------------------------------------------------- + + +class TestParseQwen3Output: + """Tests for _parse_qwen3_output(raw) ā†’ (language, text).""" + + def test_standard_english(self): + lang, text = _parse_qwen3_output( + "language Englishhello world" + ) + assert lang == "English" + assert text == "hello world" + + def test_standard_chinese(self): + lang, text = _parse_qwen3_output( + "language Chinese你儽äø–ē•Œ" + ) + assert lang == "Chinese" + assert text == "你儽äø–ē•Œ" + + def test_silent_audio_language_none(self): + lang, text = _parse_qwen3_output("language None") + assert lang == "" + assert text == "" + + def test_silent_with_unexpected_text(self): + lang, text = _parse_qwen3_output("language Nonehmm") + assert lang == "" + assert text == "hmm" + + def test_plain_text_no_tags(self): + lang, text = _parse_qwen3_output("just plain text") + assert lang == "" + assert text == "just plain text" + + def test_empty_string(self): + lang, text = _parse_qwen3_output("") + assert lang == "" + assert text == "" + + def test_none_input(self): + lang, text = _parse_qwen3_output(None) + assert lang == "" + assert text == "" + + def test_whitespace_only(self): + lang, text = _parse_qwen3_output(" ") + assert lang == "" + assert text == "" + + def test_missing_closing_tag(self): + lang, text = _parse_qwen3_output("language Englishhello world") + assert lang == "English" + assert text == "hello world" + + def test_multiline_metadata(self): + raw = "language English\nsome extra\ntext here" + lang, text = _parse_qwen3_output(raw) + assert lang == "English" + assert text == "text here" + + def test_whitespace_around_text(self): + lang, text = _parse_qwen3_output( + "language English hello " + ) + assert lang == "English" + assert text == "hello" + + +# --------------------------------------------------------------------------- +# detect_and_fix_repetitions tests +# --------------------------------------------------------------------------- + + +class TestDetectAndFixRepetitions: + """Tests for detect_and_fix_repetitions(text, threshold).""" + + def test_normal_text_unchanged(self): + text = "Hello, how are you?" + assert detect_and_fix_repetitions(text) == text + + def test_single_char_repeated_above_threshold(self): + result = detect_and_fix_repetitions("a" * 50) + assert result == "a" + + def test_single_char_repeated_below_threshold(self): + text = "a" * 10 + assert detect_and_fix_repetitions(text) == text + + def test_pattern_repeated_above_threshold(self): + result = detect_and_fix_repetitions("ha" * 30) + assert result == "ha" + + def test_short_text_unchanged(self): + assert detect_and_fix_repetitions("hi") == "hi" + + def test_mixed_content_with_repeating_tail(self): + result = detect_and_fix_repetitions("Hello " + "x" * 50) + assert result.startswith("Hello ") + # The long run of x's should be collapsed + assert len(result) < len("Hello " + "x" * 50) diff --git a/extras/friend-lite-sdk/friend_lite/button.py b/extras/friend-lite-sdk/friend_lite/button.py index 421a87b1..11e6ecd6 100644 --- a/extras/friend-lite-sdk/friend_lite/button.py +++ b/extras/friend-lite-sdk/friend_lite/button.py @@ -6,8 +6,8 @@ class ButtonState(IntEnum): IDLE = 0 - SINGLE_TAP = 1 - DOUBLE_TAP = 2 + SINGLE_PRESS = 1 + DOUBLE_PRESS = 2 LONG_PRESS = 3 PRESS = 4 RELEASE = 5 diff --git a/plugins/homeassistant/command_parser.py b/plugins/homeassistant/command_parser.py index cc73626d..de56ce97 100644 --- a/plugins/homeassistant/command_parser.py +++ b/plugins/homeassistant/command_parser.py @@ -51,16 +51,16 @@ class ParsedCommand: - set_color: Set color TARGET_TYPE (choose one): -- area: Targeting all entities of a type in an area (e.g., "study lights") -- all_in_area: Targeting ALL entities in an area (e.g., "everything in study") -- entity: Targeting a specific entity by name (e.g., "desk lamp") +- area: Target entities in a specific area OR label. Labels are groups of areas (e.g., "hall" label might cover dining_room + living_room). Both areas and labels are valid targets. +- all: Target entities across ALL areas (e.g., "all lights", "every light") +- entity: Target a specific entity by name (e.g., "desk lamp") ENTITY_TYPE (optional, use null if not specified): - light: Light entities - switch: Switch entities - fan: Fan entities - cover: Covers/blinds -- null: All entity types (when target_type is "all_in_area") +- null: All entity types PARAMETERS (optional, empty dict if none): - brightness_pct: Brightness percentage (0-100) @@ -71,8 +71,8 @@ class ParsedCommand: Command: "turn off study lights" Response: {"action": "turn_off", "target_type": "area", "target": "study", "entity_type": "light", "parameters": {}} -Command: "turn off everything in study" -Response: {"action": "turn_off", "target_type": "all_in_area", "target": "study", "entity_type": null, "parameters": {}} +Command: "turn off hall lights" +Response: {"action": "turn_off", "target_type": "area", "target": "hall", "entity_type": "light", "parameters": {}} Command: "turn on desk lamp" Response: {"action": "turn_on", "target_type": "entity", "target": "desk lamp", "entity_type": null, "parameters": {}} @@ -80,11 +80,11 @@ class ParsedCommand: Command: "set study lights to 50%" Response: {"action": "set_brightness", "target_type": "area", "target": "study", "entity_type": "light", "parameters": {"brightness_pct": 50}} -Command: "turn on living room fan" -Response: {"action": "turn_on", "target_type": "area", "target": "living room", "entity_type": "fan", "parameters": {}} - Command: "turn off all lights" -Response: {"action": "turn_off", "target_type": "entity", "target": "all", "entity_type": "light", "parameters": {}} +Response: {"action": "turn_off", "target_type": "all", "target": "all", "entity_type": "light", "parameters": {}} + +Command: "turn off everything" +Response: {"action": "turn_off", "target_type": "all", "target": "all", "entity_type": null, "parameters": {}} Command: "toggle hallway light" Response: {"action": "toggle", "target_type": "entity", "target": "hallway light", "entity_type": null, "parameters": {}} @@ -94,4 +94,6 @@ class ParsedCommand: 2. Use lowercase for action, target_type, target, entity_type 3. Use null (not "null" string) for missing entity_type 4. Always include all 5 fields: action, target_type, target, entity_type, parameters +5. The "target" for target_type "area" MUST be an area name or label name from the provided context +6. Use target_type "all" when the user says "all lights", "every light", "all the lights", etc. """ diff --git a/plugins/homeassistant/plugin.py b/plugins/homeassistant/plugin.py index b94f8ae4..f6e2d973 100644 --- a/plugins/homeassistant/plugin.py +++ b/plugins/homeassistant/plugin.py @@ -2,7 +2,7 @@ Home Assistant plugin for Chronicle. Enables control of Home Assistant devices through natural language commands -triggered by a wake word. +triggered by a keyword anywhere in the transcript. """ import json @@ -10,6 +10,7 @@ from typing import Any, Dict, List, Optional from advanced_omi_backend.plugins.base import BasePlugin, PluginContext, PluginResult + from .entity_cache import EntityCache from .mcp_client import HAMCPClient, MCPError @@ -18,20 +19,20 @@ class HomeAssistantPlugin(BasePlugin): """ - Plugin for controlling Home Assistant devices via wake word commands. + Plugin for controlling Home Assistant devices via keyword commands. Example: - User says: "Vivi, turn off the hall lights" - -> Wake word "vivi" detected by router - -> Command "turn off the hall lights" passed to on_transcript() - -> Plugin parses command and calls HA MCP to execute + User says: "Turn off the hall lights, VV" + -> Keyword "vv" detected anywhere in transcript by router + -> Command "Turn off the hall lights" passed to on_transcript() + -> Plugin parses command and calls HA to execute -> Returns: PluginResult with "I've turned off the hall light" """ SUPPORTED_ACCESS_LEVELS: List[str] = ["transcript", "button"] name = "Home Assistant" - description = "Wake word device control with Home Assistant integration" + description = "Keyword-triggered device control with Home Assistant integration" def __init__(self, config: Dict[str, Any]): """ @@ -155,7 +156,9 @@ async def on_transcript(self, context: PluginContext) -> Optional[PluginResult]: command = context.data.get("command", "") if not command: - return PluginResult(success=False, message="No command provided", should_continue=True) + return PluginResult( + success=False, message="No command provided", should_continue=True + ) if not self.mcp_client: logger.error("MCP client not initialized") @@ -166,9 +169,21 @@ async def on_transcript(self, context: PluginContext) -> Optional[PluginResult]: ) try: + conversation_id = context.data.get("conversation_id") + + # Step 0: Extract just the HA command from mixed transcript + extracted = await self._extract_ha_command( + command, conversation_id=conversation_id + ) + if extracted: + logger.info(f"Extracted HA command: '{extracted}' (from: '{command}')") + command = extracted + # Step 1: Parse command using hybrid LLM + fallback parsing logger.info(f"Processing HA command: '{command}'") - parsed = await self._parse_command_hybrid(command) + parsed = await self._parse_command_hybrid( + command, conversation_id=conversation_id + ) if not parsed: return PluginResult( @@ -199,15 +214,25 @@ async def on_transcript(self, context: PluginContext) -> Optional[PluginResult]: service = service_map.get(parsed.action, "turn_on") # Step 4: Call Home Assistant service - logger.info(f"Calling {domain}.{service} for {len(entity_ids)} entities: {entity_ids}") + logger.info( + f"Calling {domain}.{service} for {len(entity_ids)} entities: {entity_ids}" + ) result = await self.mcp_client.call_service( - domain=domain, service=service, entity_ids=entity_ids, **parsed.parameters + domain=domain, + service=service, + entity_ids=entity_ids, + **parsed.parameters, ) # Step 5: Format user-friendly response entity_type_name = parsed.entity_type or domain - if parsed.target_type == "area": + if parsed.target_type == "all": + message = ( + f"I've {parsed.action.replace('_', ' ')} {len(entity_ids)} " + f"{entity_type_name}{'s' if len(entity_ids) != 1 else ''} everywhere" + ) + elif parsed.target_type == "area": message = ( f"I've {parsed.action.replace('_', ' ')} {len(entity_ids)} " f"{entity_type_name}{'s' if len(entity_ids) != 1 else ''} " @@ -339,9 +364,10 @@ async def on_button_event(self, context: PluginContext) -> Optional[PluginResult using the button_actions config. Reuses the same entity resolution and service call logic as on_plugin_action(). """ - from .command_parser import ParsedCommand from advanced_omi_backend.plugins.events import PluginEvent + from .command_parser import ParsedCommand + # Map event to config key if context.event == PluginEvent.BUTTON_DOUBLE_PRESS: action_key = "double_press" @@ -369,7 +395,9 @@ async def on_button_event(self, context: PluginContext) -> Optional[PluginResult entity_type = action_config.get("entity_type", "light") if not target: - return PluginResult(success=False, message="No target in button_actions config") + return PluginResult( + success=False, message="No target in button_actions config" + ) parsed = ParsedCommand( action=service, @@ -430,7 +458,11 @@ async def health_check(self) -> dict: latency_ms = int((time.time() - start) * 1000) if str(result).strip() == "2": return {"ok": True, "message": "Connected", "latency_ms": latency_ms} - return {"ok": False, "message": f"Unexpected result: {result}", "latency_ms": latency_ms} + return { + "ok": False, + "message": f"Unexpected result: {result}", + "latency_ms": latency_ms, + } except Exception as e: return {"ok": False, "message": str(e)} @@ -503,14 +535,96 @@ async def _refresh_cache(self): ) logger.info( - f"Entity cache refreshed: {len(areas)} areas, " f"{len(entity_details)} entities" + f"Entity cache refreshed: {len(areas)} areas, " + f"{len(entity_details)} entities" ) except Exception as e: logger.error(f"Failed to refresh entity cache: {e}", exc_info=True) raise - async def _parse_command_with_llm(self, command: str) -> Optional["ParsedCommand"]: + async def _extract_ha_command( + self, transcript: str, *, conversation_id: Optional[str] = None + ) -> Optional[str]: + """ + Use a lightweight LLM call to extract only the Home Assistant command + from a transcript that may contain mixed conversation. + + When the keyword (e.g. "vivi") is detected anywhere in a long transcript, + the surrounding text often includes unrelated speech. This method asks the + LLM to return just the smart-home command portion. + + Args: + transcript: Transcript text with keyword already stripped. + conversation_id: Optional conversation ID for Langfuse session grouping. + + Returns: + Extracted command string, or None to fall back to the raw text. + """ + # Short transcripts are likely already just the command + if len(transcript.split()) <= 8: + return None + + try: + from advanced_omi_backend.llm_client import get_llm_client + from advanced_omi_backend.openai_factory import is_langfuse_enabled + + llm_client = get_llm_client() + + system_prompt = ( + "Extract ONLY the smart home / home assistant command from the " + "transcript below. The transcript is from a conversation and may " + "contain unrelated speech mixed in. Return ONLY the command text, " + "nothing else. If no smart home command is found, return NONE.\n\n" + "Examples:\n" + 'Input: "so anyway I was saying turn off the hall lights and then we went to dinner"\n' + "Output: turn off the hall lights\n\n" + 'Input: "turn on bedroom lights"\n' + "Output: turn on bedroom lights\n\n" + 'Input: "yeah the meeting was great oh and set living room brightness ' + 'to 50 percent and also the deadline is tomorrow"\n' + "Output: set living room brightness to 50 percent\n\n" + 'Input: "so I told him about the project and then toggle the kitchen fan ' + 'and after that we discussed lunch plans"\n' + "Output: toggle the kitchen fan" + ) + + params = { + "model": llm_client.model, + "messages": [ + {"role": "system", "content": system_prompt}, + {"role": "user", "content": transcript}, + ], + "temperature": 0.0, + "max_tokens": 100, + } + if is_langfuse_enabled(): + params["name"] = "ha-command-extraction" + params["metadata"] = { + "plugin": "homeassistant", + "step": "extract_command", + } + if conversation_id: + params["langfuse_session_id"] = conversation_id + + response = llm_client.client.chat.completions.create(**params) + + result = response.choices[0].message.content.strip() + + if not result or result.upper() == "NONE": + logger.info("LLM extraction found no HA command in transcript") + return None + + logger.info(f"LLM extracted HA command: '{result}'") + return result + + except Exception as e: + logger.warning(f"LLM command extraction failed: {e}, using raw text") + return None + + async def _parse_command_with_llm( + self, command: str, *, conversation_id: Optional[str] = None + ) -> Optional["ParsedCommand"]: """ Parse command using LLM with structured system prompt. @@ -532,26 +646,60 @@ async def _parse_command_with_llm(self, command: str) -> Optional["ParsedCommand """ try: from advanced_omi_backend.llm_client import get_llm_client + from advanced_omi_backend.openai_factory import is_langfuse_enabled from advanced_omi_backend.prompt_registry import get_prompt_registry from .command_parser import ParsedCommand llm_client = get_llm_client() registry = get_prompt_registry() - system_prompt = await registry.get_prompt("plugin.homeassistant.command_parser") + system_prompt = await registry.get_prompt( + "plugin.homeassistant.command_parser" + ) logger.debug(f"Parsing command with LLM: '{command}'") + # Build context from entity cache so the LLM knows valid targets + entity_context = "" + await self._ensure_cache_initialized() + if self.entity_cache: + areas = ( + ", ".join(self.entity_cache.areas) + if self.entity_cache.areas + else "none" + ) + labels = "" + if self.entity_cache.label_areas: + label_parts = [ + f"{lbl} (covers: {', '.join(areas_list)})" + for lbl, areas_list in self.entity_cache.label_areas.items() + ] + labels = "\nAvailable labels: " + ", ".join(label_parts) + entity_context = f"\n\nAvailable areas: {areas}{labels}\nUse target_type 'area' with an area/label name above, or target_type 'all' for everything." + # Use OpenAI chat format with system + user messages - response = llm_client.client.chat.completions.create( - model=llm_client.model, - messages=[ + params = { + "model": llm_client.model, + "messages": [ {"role": "system", "content": system_prompt}, - {"role": "user", "content": f'Command: "{command}"\n\nReturn JSON only.'}, + { + "role": "user", + "content": f'Command: "{command}"{entity_context}\n\nReturn JSON only.', + }, ], - temperature=0.1, - max_tokens=150, - ) + "temperature": 0.1, + "max_tokens": 150, + } + if is_langfuse_enabled(): + params["name"] = "ha-command-parser" + params["metadata"] = { + "plugin": "homeassistant", + "step": "parse_command", + } + if conversation_id: + params["langfuse_session_id"] = conversation_id + + response = llm_client.client.chat.completions.create(**params) result_text = response.choices[0].message.content.strip() logger.debug(f"LLM response: {result_text}") @@ -588,7 +736,9 @@ async def _parse_command_with_llm(self, command: str) -> Optional["ParsedCommand return parsed except json.JSONDecodeError as e: - logger.error(f"Failed to parse LLM JSON response: {e}\nResponse: {result_text}") + logger.error( + f"Failed to parse LLM JSON response: {e}\nResponse: {result_text}" + ) return None except Exception as e: logger.error(f"LLM command parsing failed: {e}", exc_info=True) @@ -631,8 +781,12 @@ async def _resolve_entities(self, parsed: "ParsedCommand") -> List[str]: ) if not entities: - entity_desc = f"{parsed.entity_type}s" if parsed.entity_type else "entities" - available = list(self.entity_cache.areas) + list(self.entity_cache.label_areas.keys()) + entity_desc = ( + f"{parsed.entity_type}s" if parsed.entity_type else "entities" + ) + available = list(self.entity_cache.areas) + list( + self.entity_cache.label_areas.keys() + ) raise ValueError( f"No {entity_desc} found in area/label '{parsed.target}'. " f"Available: {', '.join(available)}" @@ -644,9 +798,33 @@ async def _resolve_entities(self, parsed: "ParsedCommand") -> List[str]: ) return entities + elif parsed.target_type == "all": + # Get entities across ALL areas, optionally filtered by type + entities = [] + for area in self.entity_cache.areas: + entities.extend( + self.entity_cache.get_entities_in_area( + area=area, entity_type=parsed.entity_type + ) + ) + + if not entities: + entity_desc = ( + f"{parsed.entity_type}s" if parsed.entity_type else "entities" + ) + raise ValueError(f"No {entity_desc} found in any area") + + logger.info( + f"Resolved 'all' to {len(entities)} " + f"{parsed.entity_type or 'entity'}(s) across {len(self.entity_cache.areas)} areas" + ) + return entities + elif parsed.target_type == "all_in_area": # Get ALL entities in area (no filter) - entities = self.entity_cache.get_entities_in_area(area=parsed.target, entity_type=None) + entities = self.entity_cache.get_entities_in_area( + area=parsed.target, entity_type=None + ) if not entities: raise ValueError( @@ -654,7 +832,9 @@ async def _resolve_entities(self, parsed: "ParsedCommand") -> List[str]: f"Available areas: {', '.join(self.entity_cache.areas)}" ) - logger.info(f"Resolved 'all in {parsed.target}' to {len(entities)} entities") + logger.info( + f"Resolved 'all in {parsed.target}' to {len(entities)} entities" + ) return entities elif parsed.target_type == "entity": @@ -726,7 +906,9 @@ async def _parse_command_fallback(self, command: str) -> Optional[Dict[str, Any] "action_desc": action_desc, } - async def _parse_command_hybrid(self, command: str) -> Optional["ParsedCommand"]: + async def _parse_command_hybrid( + self, command: str, *, conversation_id: Optional[str] = None + ) -> Optional["ParsedCommand"]: """ Hybrid command parser: Try LLM first, fallback to keywords. @@ -736,6 +918,7 @@ async def _parse_command_hybrid(self, command: str) -> Optional["ParsedCommand"] Args: command: Natural language command + conversation_id: Optional conversation ID for Langfuse session grouping. Returns: ParsedCommand if successful, None otherwise @@ -751,7 +934,10 @@ async def _parse_command_hybrid(self, command: str) -> Optional["ParsedCommand"] # Try LLM parsing with timeout try: logger.debug("Attempting LLM-based command parsing...") - parsed = await asyncio.wait_for(self._parse_command_with_llm(command), timeout=5.0) + parsed = await asyncio.wait_for( + self._parse_command_with_llm(command, conversation_id=conversation_id), + timeout=5.0, + ) if parsed: logger.info("LLM parsing succeeded") @@ -823,7 +1009,9 @@ async def test_connection(config: Dict[str, Any]) -> Dict[str, Any]: try: # Validate required config fields required_fields = ["ha_url", "ha_token"] - missing_fields = [field for field in required_fields if not config.get(field)] + missing_fields = [ + field for field in required_fields if not config.get(field) + ] if missing_fields: return { diff --git a/restart.sh b/restart.sh index 019518c4..9ac7c54d 100755 --- a/restart.sh +++ b/restart.sh @@ -1,2 +1,3 @@ #!/bin/bash +source "$(dirname "$0")/scripts/check_uv.sh" uv run --with-requirements setup-requirements.txt python services.py restart --all diff --git a/scripts/check_uv.sh b/scripts/check_uv.sh new file mode 100755 index 00000000..8325bfa3 --- /dev/null +++ b/scripts/check_uv.sh @@ -0,0 +1,15 @@ +#!/bin/bash +# Shared prerequisite check for Chronicle scripts. +# Source this at the top of any shell script that needs uv. + +if ! command -v uv &> /dev/null; then + echo "āŒ 'uv' is not installed." + echo "" + echo "Chronicle requires 'uv' (Python package manager) to run." + echo "Install it with:" + echo "" + echo " curl -LsSf https://astral.sh/uv/install.sh | sh" + echo "" + echo "Then restart your terminal and try again." + exit 1 +fi diff --git a/start.sh b/start.sh index b01ef87a..723fb20f 100755 --- a/start.sh +++ b/start.sh @@ -1 +1,3 @@ +#!/bin/bash +source "$(dirname "$0")/scripts/check_uv.sh" uv run --with-requirements setup-requirements.txt python services.py start --all "$@" diff --git a/status.sh b/status.sh index a66fe459..6ded4c94 100755 --- a/status.sh +++ b/status.sh @@ -1,2 +1,3 @@ #!/bin/bash +source "$(dirname "$0")/scripts/check_uv.sh" uv run --with-requirements setup-requirements.txt python status.py "$@" diff --git a/stop.sh b/stop.sh index 0f49add7..0fc033d4 100755 --- a/stop.sh +++ b/stop.sh @@ -1 +1,3 @@ +#!/bin/bash +source "$(dirname "$0")/scripts/check_uv.sh" uv run --with-requirements setup-requirements.txt python services.py stop --all diff --git a/tests/Makefile b/tests/Makefile index 2992c3a1..b4b3e1b2 100644 --- a/tests/Makefile +++ b/tests/Makefile @@ -5,7 +5,7 @@ containers-start containers-stop containers-restart containers-rebuild \ containers-start-rebuild containers-clean containers-status containers-logs \ start stop restart rebuild start-rebuild status logs \ - test test-quick test-slow test-sdk test-no-api test-with-api-keys test-all-with-slow-and-sdk clean-all \ + test test-quick test-custom test-slow test-sdk test-no-api test-with-api-keys test-all-with-slow-and-sdk clean-all \ test-asr test-asr-gpu \ results results-path results-detailed @@ -56,6 +56,11 @@ help: @echo " make test-asr - Run ASR protocol tests (no GPU required)" @echo " make test-asr-gpu - Run ASR GPU tests (requires NVIDIA GPU)" @echo "" + @echo "Running Specific Tests:" + @echo " make test-custom T=\"Test Name\" - Run a single test by name" + @echo " make test-custom TAG=audio-streaming - Run tests by tag" + @echo " make test-custom F=integration/foo.robot - Run a specific file" + @echo "" @echo "Special Test Tags:" @echo " make test-slow - Run ONLY slow tests (backend restarts)" @echo " make test-sdk - Run ONLY SDK tests (unreleased)" @@ -232,6 +237,14 @@ test: # Quick workflow: run tests on existing containers (ignores CONFIG changes) test-quick: all +# Run specific test by name, tag, or file +# Usage: +# make test-custom T="Chunk Count Increments In Redis Session" +# make test-custom TAG=audio-streaming +# make test-custom F=integration/audio_streaming_integration_tests.robot +test-custom: + @./run-custom.sh $(if $(T),--test "$(T)") $(if $(TAG),--tag $(TAG)) $(if $(F),$(F)) + # Run ONLY slow tests (backend restarts, long timeouts) test-slow: @echo "Running slow tests only..." diff --git a/tests/config/plugins.test.yml b/tests/config/plugins.test.yml index 89772a56..7ba31157 100644 --- a/tests/config/plugins.test.yml +++ b/tests/config/plugins.test.yml @@ -12,3 +12,14 @@ plugins: condition: type: always # Capture all events without filtering db_path: /app/debug/test_plugin_events.db + + test_button_actions: + enabled: true + events: + - button.single_press + - button.double_press + condition: + type: always + actions: + single_press: + type: close_conversation diff --git a/tests/integration/audio_streaming_integration_tests.robot b/tests/integration/audio_streaming_integration_tests.robot index f34b7984..3ae29331 100644 --- a/tests/integration/audio_streaming_integration_tests.robot +++ b/tests/integration/audio_streaming_integration_tests.robot @@ -65,29 +65,33 @@ Redis Session Schema Contains All Required Fields Chunk Count Increments In Redis Session - [Documentation] Verify chunk count is tracked in Redis (not ClientState) + [Documentation] Verify chunk count is tracked in Redis (not ClientState). + ... Note: The producer re-chunks client audio into 250ms fixed-size chunks + ... (8000 bytes at 16kHz/16-bit/mono). Client sends 100ms chunks (3200 bytes). + ... So N client chunks produce floor(N * 3200 / 8000) published chunks. [Tags] infra audio-streaming ${device_name}= Set Variable chunk-count-test ${stream_id}= Open Audio Stream device_name=${device_name} ${client_id}= Get Client ID From Device Name ${device_name} - # Send chunks and verify count increases - Send Audio Chunks To Stream ${stream_id} ${TEST_AUDIO_FILE} num_chunks=3 + # Send first batch: 10 client chunks (10 * 3200 = 32000 bytes ā†’ 4 published chunks) + Send Audio Chunks To Stream ${stream_id} ${TEST_AUDIO_FILE} num_chunks=10 Sleep 1s # Allow chunk counter to update ${session1}= Get Redis Session Data ${client_id} ${count1}= Convert To Integer ${session1}[chunks_published] + Should Be True ${count1} > 0 First batch should produce at least 1 published chunk - Send Audio Chunks To Stream ${stream_id} ${TEST_AUDIO_FILE} num_chunks=5 + # Send second batch: 10 more client chunks + Send Audio Chunks To Stream ${stream_id} ${TEST_AUDIO_FILE} num_chunks=10 Sleep 1s # Allow chunk counter to update ${session2}= Get Redis Session Data ${client_id} ${count2}= Convert To Integer ${session2}[chunks_published] - # Verify count increased (should be at least 8) - Should Be True ${count2} > ${count1} - Should Be True ${count2} >= 8 + # Verify count increased between batches + Should Be True ${count2} > ${count1} Chunk count should increase after sending more audio (${count1} ā†’ ${count2}) - Log āœ… Chunk count tracked in Redis: ${count1} ā†’ ${count2} + Log Chunk count tracked in Redis: ${count1} ā†’ ${count2} # Close stream after test completes ${total_chunks}= Close Audio Stream ${stream_id} diff --git a/tests/integration/websocket_streaming_tests.robot b/tests/integration/websocket_streaming_tests.robot index 63baadf8..5161b816 100644 --- a/tests/integration/websocket_streaming_tests.robot +++ b/tests/integration/websocket_streaming_tests.robot @@ -34,12 +34,12 @@ Streaming jobs created on stream start Sleep 2s # Check speech detection job ${jobs}= Get Jobs By Type speech_detection - Should Not Be Empty ${jobs} + Should Not Be Empty ${jobs} ${speech_job}= Find Job For Client ${jobs} ${device_name} Should Not Be Equal ${speech_job} ${None} Speech detection job not created # Check audio persistence job - ${persist_job}= Find Job For Client ${jobs} ${device_name} + ${persist_job}= Find Job For Client ${jobs} ${device_name} Should Not Be Equal ${persist_job} ${None} Audio persistence job not created Log Both jobs active during streaming @@ -102,6 +102,40 @@ Conversation Job Created After Speech Detection Log Closed stream, sent ${total_chunks} total chunks +Button Press Should Close Active Conversation + [Documentation] Verify that a button single press during an active conversation + ... closes it with end_reason=close_requested and triggers post-processing + [Tags] audio-streaming conversation + [Timeout] 120s + + # Arrange: Open stream and get enough speech to start a conversation + ${device_name}= Set Variable ws-button-close + ${stream_id}= Open Audio Stream device_name=${device_name} + ${client_id}= Get Client ID From Device Name ${device_name} + + # Get baseline conversation jobs + ${baseline_jobs}= Get Jobs By Type And Client open_conversation ${client_id} + ${baseline_count}= Get Length ${baseline_jobs} + + # Send audio with speech (realtime pacing for Deepgram to finalize segments) + Send Audio Chunks To Stream ${stream_id} ${TEST_AUDIO_FILE} num_chunks=200 realtime_pacing=True + + # Wait for open_conversation_job to start (speech detected -> conversation opened) + ${jobs}= Wait Until Keyword Succeeds 60s 3s + ... Wait For New Job To Appear open_conversation ${client_id} ${baseline_count} + ${conversation_id}= Evaluate $jobs[0]['meta'].get('conversation_id', '') + Should Not Be Empty ${conversation_id} msg=Conversation ID not found in job meta + + # Act: Send button press to close the conversation + Send Button Event To Stream ${stream_id} SINGLE_PRESS + + # Assert: Conversation should close with end_reason=close_requested + Wait Until Keyword Succeeds 30s 2s + ... Conversation Should Have End Reason ${conversation_id} close_requested + + # Cleanup + [Teardown] Run Keyword And Ignore Error Close Audio Stream ${stream_id} + Conversation Closes On Inactivity Timeout And Restarts Speech Detection [Documentation] Verify that after SPEECH_INACTIVITY_THRESHOLD_SECONDS of silence (audio time), ... the open_conversation job closes with timeout_triggered=True, @@ -168,6 +202,3 @@ Conversation Closes On Inactivity Timeout And Restarts Speech Detection # Memory extraction job should be created ${memory_jobs}= Get Jobs By Type And Conversation process_memory_job ${conversation_id} Log To Console Memory jobs found: ${memory_jobs.__len__()} - - - diff --git a/tests/integration/websocket_transcription_e2e_test.robot b/tests/integration/websocket_transcription_e2e_test.robot index 3711ac54..951f3aa7 100644 --- a/tests/integration/websocket_transcription_e2e_test.robot +++ b/tests/integration/websocket_transcription_e2e_test.robot @@ -53,16 +53,13 @@ WebSocket Stream Produces Final Transcripts In Redis Log Closing stream - should trigger: end_marker ā†’ CloseStream ā†’ final results Close Audio Stream ${stream_id} - # Allow time for streaming consumer to process end_marker and get final results - Sleep 5s - - # Verify Redis stream transcription:results:{client_id} has entries + # Wait for streaming consumer to process end_marker and write final results to Redis + # Use retry loop instead of fixed sleep - consumer processing time varies ${stream_name}= Set Variable transcription:results:${client_id} - ${stream_length}= Redis Command XLEN ${stream_name} - - Should Be True ${stream_length} > 0 - ... Redis stream ${stream_name} is empty - no final transcripts received! This means end_marker was not sent or CloseStream failed. + Wait Until Keyword Succeeds 30s 2s + ... Redis Stream Should Not Be Empty ${stream_name} + ${stream_length}= Redis Command XLEN ${stream_name} Log āœ… Redis stream has ${stream_length} final transcript(s) @@ -387,3 +384,13 @@ Streaming Completion Signal Is Set Before Transcript Read Log āœ… Completion signal ${completion_key} = ${signal_value} (consumer completed before job reads) +*** Keywords *** + +Redis Stream Should Not Be Empty + [Documentation] Assert that a Redis stream has at least one entry. + ... Used with Wait Until Keyword Succeeds for retry-based checks. + [Arguments] ${stream_name} + + ${stream_length}= Redis Command XLEN ${stream_name} + Should Be True ${stream_length} > 0 + ... Redis stream ${stream_name} is empty - no final transcripts received! This means end_marker was not sent or CloseStream failed. diff --git a/tests/libs/audio_stream_library.py b/tests/libs/audio_stream_library.py index e14a174e..f1651dc1 100644 --- a/tests/libs/audio_stream_library.py +++ b/tests/libs/audio_stream_library.py @@ -27,7 +27,10 @@ sys.path.insert(0, str(backend_src)) from advanced_omi_backend.clients import AudioStreamClient -from advanced_omi_backend.clients.audio_stream_client import StreamManager, stream_audio_file as _stream_audio_file +from advanced_omi_backend.clients.audio_stream_client import StreamManager +from advanced_omi_backend.clients.audio_stream_client import ( + stream_audio_file as _stream_audio_file, +) # Module-level manager for non-blocking streams _manager = StreamManager() @@ -37,6 +40,7 @@ # Blocking Mode (simple, streams entire file) # ============================================================================= + def stream_audio_file( base_url: str, token: str, @@ -60,6 +64,7 @@ def stream_audio_file( # Non-blocking Mode (for testing during stream) # ============================================================================= + def start_audio_stream( base_url: str, token: str, @@ -145,6 +150,16 @@ def close_audio_stream_without_stop(stream_id: str) -> int: return _manager.close_stream_without_stop(stream_id) +def send_button_event(stream_id: str, button_state: str = "SINGLE_PRESS") -> None: + """Send a button event to an open stream. + + Args: + stream_id: Stream session ID + button_state: Button state ("SINGLE_PRESS" or "DOUBLE_PRESS") + """ + _manager.send_button_event(stream_id, button_state) + + def cleanup_all_streams(): """Stop all active streams.""" _manager.cleanup_all() @@ -154,6 +169,7 @@ def cleanup_all_streams(): # Advanced Usage # ============================================================================= + def get_audio_stream_client( base_url: str, token: str, diff --git a/tests/resources/websocket_keywords.robot b/tests/resources/websocket_keywords.robot index 2eb9381b..58014de5 100644 --- a/tests/resources/websocket_keywords.robot +++ b/tests/resources/websocket_keywords.robot @@ -158,6 +158,12 @@ Close Audio Stream Log Stopped stream ${stream_id}, total chunks: ${total_chunks} RETURN ${total_chunks} +Send Button Event To Stream + [Documentation] Send a button event (SINGLE_PRESS, DOUBLE_PRESS) to an open stream + [Arguments] ${stream_id} ${button_state}=SINGLE_PRESS + Send Button Event ${stream_id} ${button_state} + Log Sent button event ${button_state} to stream ${stream_id} + Close Audio Stream Without Stop Event [Documentation] Close WebSocket connection without sending audio-stop event. ... This simulates abrupt disconnection (network failure, client crash) diff --git a/wizard.py b/wizard.py index b04f028c..e4d52395 100755 --- a/wizard.py +++ b/wizard.py @@ -26,38 +26,67 @@ console = Console() SERVICES = { - 'backend': { - 'advanced': { - 'path': 'backends/advanced', - 'cmd': ['uv', 'run', '--with-requirements', '../../setup-requirements.txt', 'python', 'init.py'], - 'description': 'Advanced AI backend with full feature set', - 'required': True + "backend": { + "advanced": { + "path": "backends/advanced", + "cmd": [ + "uv", + "run", + "--with-requirements", + "../../setup-requirements.txt", + "python", + "init.py", + ], + "description": "Advanced AI backend with full feature set", + "required": True, } }, - 'extras': { - 'speaker-recognition': { - 'path': 'extras/speaker-recognition', - 'cmd': ['uv', 'run', '--with-requirements', '../../setup-requirements.txt', 'python', 'init.py'], - 'description': 'Speaker identification and enrollment' + "extras": { + "speaker-recognition": { + "path": "extras/speaker-recognition", + "cmd": [ + "uv", + "run", + "--with-requirements", + "../../setup-requirements.txt", + "python", + "init.py", + ], + "description": "Speaker identification and enrollment", }, - 'asr-services': { - 'path': 'extras/asr-services', - 'cmd': ['uv', 'run', '--with-requirements', '../../setup-requirements.txt', 'python', 'init.py'], - 'description': 'Offline speech-to-text' + "asr-services": { + "path": "extras/asr-services", + "cmd": [ + "uv", + "run", + "--with-requirements", + "../../setup-requirements.txt", + "python", + "init.py", + ], + "description": "Offline speech-to-text", }, - 'openmemory-mcp': { - 'path': 'extras/openmemory-mcp', - 'cmd': ['./setup.sh'], - 'description': 'OpenMemory MCP server' + "openmemory-mcp": { + "path": "extras/openmemory-mcp", + "cmd": ["./setup.sh"], + "description": "OpenMemory MCP server", }, - 'langfuse': { - 'path': 'extras/langfuse', - 'cmd': ['uv', 'run', '--with-requirements', '../../setup-requirements.txt', 'python', 'init.py'], - 'description': 'LLM observability and prompt management (local)' - } - } + "langfuse": { + "path": "extras/langfuse", + "cmd": [ + "uv", + "run", + "--with-requirements", + "../../setup-requirements.txt", + "python", + "init.py", + ], + "description": "LLM observability and prompt management (local)", + }, + }, } + def discover_available_plugins(): """ Discover plugins by scanning plugins directory. @@ -75,11 +104,13 @@ def discover_available_plugins(): plugins_dir = Path("backends/advanced/src/advanced_omi_backend/plugins") if not plugins_dir.exists(): - console.print(f"[yellow]Warning: Plugins directory not found: {plugins_dir}[/yellow]") + console.print( + f"[yellow]Warning: Plugins directory not found: {plugins_dir}[/yellow]" + ) return {} discovered = {} - skip_dirs = {'__pycache__', '__init__.py', 'base.py', 'router.py'} + skip_dirs = {"__pycache__", "__init__.py", "base.py", "router.py"} for plugin_dir in plugins_dir.iterdir(): if not plugin_dir.is_dir() or plugin_dir.name in skip_dirs: @@ -89,32 +120,37 @@ def discover_available_plugins(): setup_script = plugin_dir / "setup.py" discovered[plugin_id] = { - 'has_setup': setup_script.exists(), - 'setup_path': setup_script if setup_script.exists() else None, - 'dir': plugin_dir + "has_setup": setup_script.exists(), + "setup_path": setup_script if setup_script.exists() else None, + "dir": plugin_dir, } return discovered + def check_service_exists(service_name, service_config): """Check if service directory and script exist""" - service_path = Path(service_config['path']) + service_path = Path(service_config["path"]) if not service_path.exists(): return False, f"Directory {service_path} does not exist" # For services with Python init scripts, check if init.py exists - if service_name in ['advanced', 'speaker-recognition', 'asr-services', 'langfuse']: - script_path = service_path / 'init.py' + if service_name in ["advanced", "speaker-recognition", "asr-services", "langfuse"]: + script_path = service_path / "init.py" if not script_path.exists(): return False, f"Script {script_path} does not exist" else: # For other extras, check if setup.sh exists - script_path = service_path / 'setup.sh' + script_path = service_path / "setup.sh" if not script_path.exists(): - return False, f"Script {script_path} does not exist (will be created in Phase 2)" + return ( + False, + f"Script {script_path} does not exist (will be created in Phase 2)", + ) return True, "OK" + def select_services(transcription_provider=None): """Let user select which services to setup""" console.print("šŸš€ [bold cyan]Chronicle Service Setup[/bold cyan]") @@ -125,24 +161,30 @@ def select_services(transcription_provider=None): # Backend is required console.print("šŸ“± [bold]Backend (Required):[/bold]") console.print(" āœ… Advanced Backend - Full AI features") - selected.append('advanced') + selected.append("advanced") # Services that will be auto-added based on transcription provider choice auto_added = set() if transcription_provider in ("parakeet", "vibevoice", "qwen3-asr"): - auto_added.add('asr-services') + auto_added.add("asr-services") # Optional extras console.print("\nšŸ”§ [bold]Optional Services:[/bold]") - for service_name, service_config in SERVICES['extras'].items(): + for service_name, service_config in SERVICES["extras"].items(): # Skip services that will be auto-added based on earlier choices if service_name in auto_added: - provider_label = {"vibevoice": "VibeVoice", "parakeet": "Parakeet", "qwen3-asr": "Qwen3-ASR"}.get(transcription_provider, transcription_provider) - console.print(f" āœ… {service_config['description']} ({provider_label}) [dim](auto-selected)[/dim]") + provider_label = { + "vibevoice": "VibeVoice", + "parakeet": "Parakeet", + "qwen3-asr": "Qwen3-ASR", + }.get(transcription_provider, transcription_provider) + console.print( + f" āœ… {service_config['description']} ({provider_label}) [dim](auto-selected)[/dim]" + ) continue # LangFuse is handled separately via setup_langfuse_choice() - if service_name == 'langfuse': + if service_name == "langfuse": continue # Check if service exists @@ -152,10 +194,12 @@ def select_services(transcription_provider=None): continue # Speaker recognition is recommended by default - default_enable = service_name == 'speaker-recognition' + default_enable = service_name == "speaker-recognition" try: - enable_service = Confirm.ask(f" Setup {service_config['description']}?", default=default_enable) + enable_service = Confirm.ask( + f" Setup {service_config['description']}?", default=default_enable + ) except EOFError: console.print(f"Using default: {'Yes' if default_enable else 'No'}") enable_service = default_enable @@ -165,213 +209,268 @@ def select_services(transcription_provider=None): return selected + def cleanup_unselected_services(selected_services): """Backup and remove .env files from services that weren't selected""" - - all_services = list(SERVICES['backend'].keys()) + list(SERVICES['extras'].keys()) - + + all_services = list(SERVICES["backend"].keys()) + list(SERVICES["extras"].keys()) + for service_name in all_services: if service_name not in selected_services: - if service_name == 'advanced': - service_path = Path(SERVICES['backend'][service_name]['path']) + if service_name == "advanced": + service_path = Path(SERVICES["backend"][service_name]["path"]) else: - service_path = Path(SERVICES['extras'][service_name]['path']) - - env_file = service_path / '.env' + service_path = Path(SERVICES["extras"][service_name]["path"]) + + env_file = service_path / ".env" if env_file.exists(): # Create backup with timestamp timestamp = datetime.now().strftime("%Y%m%d_%H%M%S") - backup_file = service_path / f'.env.backup.{timestamp}.unselected' + backup_file = service_path / f".env.backup.{timestamp}.unselected" env_file.rename(backup_file) - console.print(f"šŸ§¹ [dim]Backed up {service_name} configuration to {backup_file.name} (service not selected)[/dim]") - -def run_service_setup(service_name, selected_services, https_enabled=False, server_ip=None, - obsidian_enabled=False, neo4j_password=None, hf_token=None, - transcription_provider='deepgram', admin_email=None, admin_password=None, - langfuse_public_key=None, langfuse_secret_key=None, langfuse_host=None, - streaming_provider=None): + console.print( + f"šŸ§¹ [dim]Backed up {service_name} configuration to {backup_file.name} (service not selected)[/dim]" + ) + + +def run_service_setup( + service_name, + selected_services, + https_enabled=False, + server_ip=None, + obsidian_enabled=False, + neo4j_password=None, + hf_token=None, + transcription_provider="deepgram", + admin_email=None, + admin_password=None, + langfuse_public_key=None, + langfuse_secret_key=None, + langfuse_host=None, + streaming_provider=None, +): """Execute individual service setup script""" - if service_name == 'advanced': - service = SERVICES['backend'][service_name] + if service_name == "advanced": + service = SERVICES["backend"][service_name] # For advanced backend, pass URLs of other selected services and HTTPS config - cmd = service['cmd'].copy() - if 'speaker-recognition' in selected_services: - cmd.extend(['--speaker-service-url', 'http://speaker-service:8085']) - if 'asr-services' in selected_services: - cmd.extend(['--parakeet-asr-url', 'http://host.docker.internal:8767']) + cmd = service["cmd"].copy() + if "speaker-recognition" in selected_services: + cmd.extend(["--speaker-service-url", "http://speaker-service:8085"]) + if "asr-services" in selected_services: + cmd.extend(["--parakeet-asr-url", "http://host.docker.internal:8767"]) # Pass transcription provider choice from wizard if transcription_provider: - cmd.extend(['--transcription-provider', transcription_provider]) + cmd.extend(["--transcription-provider", transcription_provider]) # Pass streaming provider (different from batch) for re-transcription setup if streaming_provider: - cmd.extend(['--streaming-provider', streaming_provider]) + cmd.extend(["--streaming-provider", streaming_provider]) # Add HTTPS configuration if https_enabled and server_ip: - cmd.extend(['--enable-https', '--server-ip', server_ip]) + cmd.extend(["--enable-https", "--server-ip", server_ip]) # Always pass Neo4j password (neo4j is a required service) if neo4j_password: - cmd.extend(['--neo4j-password', neo4j_password]) + cmd.extend(["--neo4j-password", neo4j_password]) # Add Obsidian configuration if obsidian_enabled: - cmd.extend(['--enable-obsidian']) + cmd.extend(["--enable-obsidian"]) # Pass LangFuse keys from langfuse init or external config if langfuse_public_key and langfuse_secret_key: - cmd.extend(['--langfuse-public-key', langfuse_public_key]) - cmd.extend(['--langfuse-secret-key', langfuse_secret_key]) + cmd.extend(["--langfuse-public-key", langfuse_public_key]) + cmd.extend(["--langfuse-secret-key", langfuse_secret_key]) if langfuse_host: - cmd.extend(['--langfuse-host', langfuse_host]) + cmd.extend(["--langfuse-host", langfuse_host]) else: - service = SERVICES['extras'][service_name] - cmd = service['cmd'].copy() - + service = SERVICES["extras"][service_name] + cmd = service["cmd"].copy() + # Add HTTPS configuration for services that support it - if service_name == 'speaker-recognition' and https_enabled and server_ip: - cmd.extend(['--enable-https', '--server-ip', server_ip]) + if service_name == "speaker-recognition" and https_enabled and server_ip: + cmd.extend(["--enable-https", "--server-ip", server_ip]) # For speaker-recognition, pass HF_TOKEN from centralized configuration - if service_name == 'speaker-recognition': + if service_name == "speaker-recognition": # Define the speaker env path - speaker_env_path = 'extras/speaker-recognition/.env' + speaker_env_path = "extras/speaker-recognition/.env" # HF Token should have been provided via setup_hf_token_if_needed() if hf_token: - cmd.extend(['--hf-token', hf_token]) + cmd.extend(["--hf-token", hf_token]) else: - console.print("[yellow][WARNING][/yellow] No HF_TOKEN provided - speaker recognition may fail to download models") + console.print( + "[yellow][WARNING][/yellow] No HF_TOKEN provided - speaker recognition may fail to download models" + ) # Pass Deepgram API key from backend if available - backend_env_path = 'backends/advanced/.env' - deepgram_key = read_env_value(backend_env_path, 'DEEPGRAM_API_KEY') - if deepgram_key and not is_placeholder(deepgram_key, 'your_deepgram_api_key_here', 'your-deepgram-api-key-here'): - cmd.extend(['--deepgram-api-key', deepgram_key]) - console.print("[blue][INFO][/blue] Found existing DEEPGRAM_API_KEY from backend config, reusing") + backend_env_path = "backends/advanced/.env" + deepgram_key = read_env_value(backend_env_path, "DEEPGRAM_API_KEY") + if deepgram_key and not is_placeholder( + deepgram_key, "your_deepgram_api_key_here", "your-deepgram-api-key-here" + ): + cmd.extend(["--deepgram-api-key", deepgram_key]) + console.print( + "[blue][INFO][/blue] Found existing DEEPGRAM_API_KEY from backend config, reusing" + ) # Pass compute mode from existing .env if available - compute_mode = read_env_value(speaker_env_path, 'COMPUTE_MODE') - if compute_mode in ['cpu', 'gpu']: - cmd.extend(['--compute-mode', compute_mode]) - console.print(f"[blue][INFO][/blue] Found existing COMPUTE_MODE ({compute_mode}), reusing") - + compute_mode = read_env_value(speaker_env_path, "COMPUTE_MODE") + if compute_mode in ["cpu", "gpu"]: + cmd.extend(["--compute-mode", compute_mode]) + console.print( + f"[blue][INFO][/blue] Found existing COMPUTE_MODE ({compute_mode}), reusing" + ) + # For asr-services, pass provider from wizard's transcription choice and reuse CUDA version - if service_name == 'asr-services': + if service_name == "asr-services": # Map wizard transcription provider to asr-services provider name wizard_to_asr_provider = { - 'vibevoice': 'vibevoice', - 'parakeet': 'nemo', - 'qwen3-asr': 'qwen3-asr', + "vibevoice": "vibevoice", + "parakeet": "nemo", + "qwen3-asr": "qwen3-asr", } asr_provider = wizard_to_asr_provider.get(transcription_provider) if asr_provider: - cmd.extend(['--provider', asr_provider]) - console.print(f"[blue][INFO][/blue] Pre-selecting ASR provider: {asr_provider} (from wizard choice: {transcription_provider})") - - speaker_env_path = 'extras/speaker-recognition/.env' - cuda_version = read_env_value(speaker_env_path, 'PYTORCH_CUDA_VERSION') - if cuda_version and cuda_version in ['cu121', 'cu126', 'cu128']: - cmd.extend(['--pytorch-cuda-version', cuda_version]) - console.print(f"[blue][INFO][/blue] Found existing PYTORCH_CUDA_VERSION ({cuda_version}) from speaker-recognition, reusing") + cmd.extend(["--provider", asr_provider]) + console.print( + f"[blue][INFO][/blue] Pre-selecting ASR provider: {asr_provider} (from wizard choice: {transcription_provider})" + ) + + speaker_env_path = "extras/speaker-recognition/.env" + cuda_version = read_env_value(speaker_env_path, "PYTORCH_CUDA_VERSION") + if cuda_version and cuda_version in ["cu121", "cu126", "cu128"]: + cmd.extend(["--pytorch-cuda-version", cuda_version]) + console.print( + f"[blue][INFO][/blue] Found existing PYTORCH_CUDA_VERSION ({cuda_version}) from speaker-recognition, reusing" + ) # For langfuse, pass admin credentials from backend - if service_name == 'langfuse': + if service_name == "langfuse": if admin_email: - cmd.extend(['--admin-email', admin_email]) + cmd.extend(["--admin-email", admin_email]) if admin_password: - cmd.extend(['--admin-password', admin_password]) + cmd.extend(["--admin-password", admin_password]) # For openmemory-mcp, try to pass OpenAI API key from backend if available - if service_name == 'openmemory-mcp': - backend_env_path = 'backends/advanced/.env' - openai_key = read_env_value(backend_env_path, 'OPENAI_API_KEY') - if openai_key and not is_placeholder(openai_key, 'your_openai_api_key_here', 'your-openai-api-key-here', 'your_openai_key_here', 'your-openai-key-here'): - cmd.extend(['--openai-api-key', openai_key]) - console.print("[blue][INFO][/blue] Found existing OPENAI_API_KEY from backend config, reusing") - + if service_name == "openmemory-mcp": + backend_env_path = "backends/advanced/.env" + openai_key = read_env_value(backend_env_path, "OPENAI_API_KEY") + if openai_key and not is_placeholder( + openai_key, + "your_openai_api_key_here", + "your-openai-api-key-here", + "your_openai_key_here", + "your-openai-key-here", + ): + cmd.extend(["--openai-api-key", openai_key]) + console.print( + "[blue][INFO][/blue] Found existing OPENAI_API_KEY from backend config, reusing" + ) + console.print(f"\nšŸ”§ [bold]Setting up {service_name}...[/bold]") - + # Check if service exists before running exists, msg = check_service_exists(service_name, service) if not exists: console.print(f"āŒ {service_name} setup failed: {msg}") return False - + try: result = subprocess.run( - cmd, - cwd=service['path'], + cmd, + cwd=service["path"], check=True, - timeout=300 # 5 minute timeout for service setup + timeout=300, # 5 minute timeout for service setup ) - + console.print(f"āœ… {service_name} setup completed") return True - + except FileNotFoundError as e: console.print(f"āŒ {service_name} setup failed: {e}") - console.print(f"[yellow] Check that the service directory exists: {service['path']}[/yellow]") - console.print(f"[yellow] And that 'uv' is installed and on your PATH[/yellow]") + console.print( + f"[yellow] Check that the service directory exists: {service['path']}[/yellow]" + ) + console.print( + f"[yellow] And that 'uv' is installed and on your PATH[/yellow]" + ) return False except subprocess.TimeoutExpired as e: console.print(f"āŒ {service_name} setup timed out after {e.timeout}s") console.print(f"[yellow] Configuration may be partially written.[/yellow]") console.print(f"[yellow] To retry just this service:[/yellow]") - console.print(f"[yellow] cd {service['path']} && {' '.join(service['cmd'])}[/yellow]") + console.print( + f"[yellow] cd {service['path']} && {' '.join(service['cmd'])}[/yellow]" + ) return False except subprocess.CalledProcessError as e: console.print(f"āŒ {service_name} setup failed with exit code {e.returncode}") console.print(f"[yellow] Check the error output above for details.[/yellow]") console.print(f"[yellow] To retry just this service:[/yellow]") - console.print(f"[yellow] cd {service['path']} && {' '.join(service['cmd'])}[/yellow]") + console.print( + f"[yellow] cd {service['path']} && {' '.join(service['cmd'])}[/yellow]" + ) return False except Exception as e: console.print(f"āŒ {service_name} setup failed: {e}") return False + def show_service_status(): """Show which services are available""" console.print("\nšŸ“‹ [bold]Service Status:[/bold]") - + # Check backend - exists, msg = check_service_exists('advanced', SERVICES['backend']['advanced']) + exists, msg = check_service_exists("advanced", SERVICES["backend"]["advanced"]) status = "āœ…" if exists else "āŒ" console.print(f" {status} Advanced Backend - {msg}") - + # Check extras - for service_name, service_config in SERVICES['extras'].items(): + for service_name, service_config in SERVICES["extras"].items(): exists, msg = check_service_exists(service_name, service_config) status = "āœ…" if exists else "āøļø" console.print(f" {status} {service_config['description']} - {msg}") + def run_plugin_setup(plugin_id, plugin_info): """Run a plugin's setup.py script""" - setup_path = plugin_info['setup_path'] + setup_path = plugin_info["setup_path"] try: # Run plugin setup script interactively (don't capture output) # This allows the plugin to prompt for user input result = subprocess.run( - ['uv', 'run', '--with-requirements', 'setup-requirements.txt', 'python', str(setup_path)], - cwd=str(Path.cwd()) + [ + "uv", + "run", + "--with-requirements", + "setup-requirements.txt", + "python", + str(setup_path), + ], + cwd=str(Path.cwd()), ) if result.returncode == 0: console.print(f"\n[green]āœ… {plugin_id} configured successfully[/green]") return True else: - console.print(f"\n[red]āŒ {plugin_id} setup failed with exit code {result.returncode}[/red]") + console.print( + f"\n[red]āŒ {plugin_id} setup failed with exit code {result.returncode}[/red]" + ) return False except Exception as e: console.print(f"[red]āŒ Error running {plugin_id} setup: {e}[/red]") return False + def setup_plugins(): """Discover and setup plugins via delegation""" console.print("\nšŸ”Œ [bold cyan]Plugin Configuration[/bold cyan]") @@ -386,10 +485,7 @@ def setup_plugins(): # Ask about enabling community plugins try: - enable_plugins = Confirm.ask( - "Enable community plugins?", - default=True - ) + enable_plugins = Confirm.ask("Enable community plugins?", default=True) except EOFError: console.print("Using default: Yes") enable_plugins = True @@ -401,16 +497,15 @@ def setup_plugins(): # For each plugin with setup script configured_count = 0 for plugin_id, plugin_info in available_plugins.items(): - if not plugin_info['has_setup']: - console.print(f"[dim] {plugin_id}: No setup wizard available (configure manually)[/dim]") + if not plugin_info["has_setup"]: + console.print( + f"[dim] {plugin_id}: No setup wizard available (configure manually)[/dim]" + ) continue # Ask if user wants to configure this plugin try: - configure = Confirm.ask( - f" Configure {plugin_id} plugin?", - default=False - ) + configure = Confirm.ask(f" Configure {plugin_id} plugin?", default=False) except EOFError: configure = False @@ -423,36 +518,53 @@ def setup_plugins(): console.print(f"\n[green]āœ… Configured {configured_count} plugin(s)[/green]") + def setup_git_hooks(): """Setup pre-commit hooks for development""" console.print("\nšŸ”§ [bold]Setting up development environment...[/bold]") + # Check if git is available + if not shutil.which("git"): + console.print( + "āš ļø [yellow]git not found, skipping git hooks setup (optional)[/yellow]" + ) + return + try: - # Install pre-commit if not already installed - subprocess.run(['pip', 'install', 'pre-commit'], - stdout=subprocess.DEVNULL, - stderr=subprocess.DEVNULL, - check=False) + # Install pre-commit via uv tool (uv is our package manager) + subprocess.run( + ["uv", "tool", "install", "pre-commit"], + stdout=subprocess.DEVNULL, + stderr=subprocess.DEVNULL, + check=False, + ) # Install git hooks - result = subprocess.run(['pre-commit', 'install', '--hook-type', 'pre-push'], - capture_output=True, - text=True) + result = subprocess.run( + ["pre-commit", "install", "--hook-type", "pre-push"], + capture_output=True, + text=True, + ) if result.returncode == 0: - console.print("āœ… [green]Git hooks installed (tests will run before push)[/green]") + console.print( + "āœ… [green]Git hooks installed (tests will run before push)[/green]" + ) else: console.print("āš ļø [yellow]Could not install git hooks (optional)[/yellow]") # Also install pre-commit hook - subprocess.run(['pre-commit', 'install', '--hook-type', 'pre-commit'], - stdout=subprocess.DEVNULL, - stderr=subprocess.DEVNULL, - check=False) + subprocess.run( + ["pre-commit", "install", "--hook-type", "pre-commit"], + stdout=subprocess.DEVNULL, + stderr=subprocess.DEVNULL, + check=False, + ) except Exception as e: console.print(f"āš ļø [yellow]Could not setup git hooks: {e} (optional)[/yellow]") + def setup_hf_token_if_needed(selected_services): """Prompt for Hugging Face token if needed by selected services. @@ -463,40 +575,54 @@ def setup_hf_token_if_needed(selected_services): HF_TOKEN string if provided, None otherwise """ # Check if any selected services need HF_TOKEN - needs_hf_token = 'speaker-recognition' in selected_services + needs_hf_token = "speaker-recognition" in selected_services if not needs_hf_token: return None console.print("\nšŸ¤— [bold cyan]Hugging Face Token Configuration[/bold cyan]") console.print("Required for speaker recognition (PyAnnote models)") - console.print("\n[blue][INFO][/blue] Get your token from: https://huggingface.co/settings/tokens") + console.print( + "\n[blue][INFO][/blue] Get your token from: https://huggingface.co/settings/tokens" + ) console.print() - console.print("[yellow]āš ļø You must also accept the model agreements for these gated models:[/yellow]") + console.print( + "[yellow]āš ļø You must also accept the model agreements for these gated models:[/yellow]" + ) console.print(" 1. [cyan]Speaker Diarization[/cyan]") - console.print(" https://huggingface.co/pyannote/speaker-diarization-community-1") + console.print( + " https://huggingface.co/pyannote/speaker-diarization-community-1" + ) console.print(" 2. [cyan]Segmentation Model[/cyan]") console.print(" https://huggingface.co/pyannote/segmentation-3.0") console.print(" 3. [cyan]Segmentation Model[/cyan]") console.print(" https://huggingface.co/pyannote/segmentation-3.1") console.print(" 4. [cyan]Embedding Model[/cyan]") - console.print(" https://huggingface.co/pyannote/wespeaker-voxceleb-resnet34-LM") + console.print( + " https://huggingface.co/pyannote/wespeaker-voxceleb-resnet34-LM" + ) console.print() - console.print("[yellow]ā†’[/yellow] Open each link and click 'Agree and access repository'") + console.print( + "[yellow]ā†’[/yellow] Open each link and click 'Agree and access repository'" + ) console.print("[yellow]ā†’[/yellow] Use the same Hugging Face account as your token") console.print() # Check for existing token from speaker-recognition service - speaker_env_path = 'extras/speaker-recognition/.env' - existing_token = read_env_value(speaker_env_path, 'HF_TOKEN') + speaker_env_path = "extras/speaker-recognition/.env" + existing_token = read_env_value(speaker_env_path, "HF_TOKEN") # Use the masked prompt function hf_token = prompt_with_existing_masked( prompt_text="Hugging Face Token", existing_value=existing_token, - placeholders=['your_huggingface_token_here', 'your-huggingface-token-here', 'hf_xxxxx'], + placeholders=[ + "your_huggingface_token_here", + "your-huggingface-token-here", + "hf_xxxxx", + ], is_password=True, - default="" + default="", ) if hf_token: @@ -504,9 +630,12 @@ def setup_hf_token_if_needed(selected_services): console.print(f"[green]āœ… HF_TOKEN configured: {masked}[/green]\n") return hf_token else: - console.print("[yellow]āš ļø No HF_TOKEN provided - speaker recognition may fail[/yellow]\n") + console.print( + "[yellow]āš ļø No HF_TOKEN provided - speaker recognition may fail[/yellow]\n" + ) return None + def setup_config_file(): """Setup config/config.yml from template if it doesn't exist""" config_file = Path("config/config.yml") @@ -519,9 +648,14 @@ def setup_config_file(): shutil.copy(config_template, config_file) console.print("āœ… [green]Created config/config.yml from template[/green]") else: - console.print("āš ļø [yellow]config/config.yml.template not found, skipping config setup[/yellow]") + console.print( + "āš ļø [yellow]config/config.yml.template not found, skipping config setup[/yellow]" + ) else: - console.print("ā„¹ļø [blue]config/config.yml already exists, keeping existing configuration[/blue]") + console.print( + "ā„¹ļø [blue]config/config.yml already exists, keeping existing configuration[/blue]" + ) + # Providers that support real-time streaming STREAMING_CAPABLE = {"deepgram", "smallest", "qwen3-asr"} @@ -530,8 +664,12 @@ def setup_config_file(): def select_transcription_provider(): """Ask user which transcription provider they want (batch/primary).""" console.print("\nšŸŽ¤ [bold cyan]Transcription Provider[/bold cyan]") - console.print("Choose your speech-to-text provider (used for [bold]batch[/bold]/high-quality transcription):") - console.print("[dim]If it also supports streaming, it will be used for real-time too by default.[/dim]") + console.print( + "Choose your speech-to-text provider (used for [bold]batch[/bold]/high-quality transcription):" + ) + console.print( + "[dim]If it also supports streaming, it will be used for real-time too by default.[/dim]" + ) console.print() choices = { @@ -540,7 +678,7 @@ def select_transcription_provider(): "3": "VibeVoice ASR (offline, batch only, built-in diarization, GPU)", "4": "Qwen3-ASR (offline, streaming + batch, 52 languages, GPU)", "5": "Smallest.ai Pulse (cloud, streaming + batch)", - "6": "None (skip transcription setup)" + "6": "None (skip transcription setup)", } for key, desc in choices.items(): @@ -563,7 +701,9 @@ def select_transcription_provider(): return "smallest" elif choice == "6": return "none" - console.print(f"[red]Invalid choice. Please select from {list(choices.keys())}[/red]") + console.print( + f"[red]Invalid choice. Please select from {list(choices.keys())}[/red]" + ) except EOFError: console.print("Using default: Deepgram") return "deepgram" @@ -586,7 +726,9 @@ def select_streaming_provider(batch_provider): console.print(f"\nšŸ”Š [bold cyan]Streaming[/bold cyan]") console.print(f"{batch_provider} supports both batch and streaming.") try: - use_different = Confirm.ask("Use a different provider for real-time streaming?", default=False) + use_different = Confirm.ask( + "Use a different provider for real-time streaming?", default=False + ) except EOFError: return None if not use_different: @@ -594,7 +736,9 @@ def select_streaming_provider(batch_provider): else: # Batch-only provider ā€” need to pick a streaming provider console.print(f"\nšŸ”Š [bold cyan]Streaming[/bold cyan]") - console.print(f"{batch_provider} is batch-only. Pick a streaming provider for real-time transcription:") + console.print( + f"{batch_provider} is batch-only. Pick a streaming provider for real-time transcription:" + ) # Show streaming-capable providers (excluding the batch provider) streaming_choices = {} @@ -625,9 +769,13 @@ def select_streaming_provider(batch_provider): if choice in streaming_choices: result = provider_map[choice] if result: - console.print(f"[green]āœ…[/green] Streaming: {result}, Batch: {batch_provider}") + console.print( + f"[green]āœ…[/green] Streaming: {result}, Batch: {batch_provider}" + ) return result - console.print(f"[red]Invalid choice. Please select from {list(streaming_choices.keys())}[/red]") + console.print( + f"[red]Invalid choice. Please select from {list(streaming_choices.keys())}[/red]" + ) except EOFError: return None @@ -648,31 +796,36 @@ def setup_langfuse_choice(): console.print() try: - has_existing = Confirm.ask("Use an existing external LangFuse instance instead of local?", default=False) + has_existing = Confirm.ask( + "Use an existing external LangFuse instance instead of local?", + default=False, + ) except EOFError: console.print("Using default: No (will set up locally)") has_existing = False if not has_existing: # Check if the local langfuse directory exists - exists, msg = check_service_exists('langfuse', SERVICES['extras']['langfuse']) + exists, msg = check_service_exists("langfuse", SERVICES["extras"]["langfuse"]) if exists: console.print("[green]āœ…[/green] Will set up local LangFuse instance") - return 'local', {} + return "local", {} else: console.print(f"[yellow]āš ļø Local LangFuse not available: {msg}[/yellow]") - console.print("[yellow] Will proceed without LangFuse ā€” add it later when available[/yellow]") - return 'local', {} + console.print( + "[yellow] Will proceed without LangFuse ā€” add it later when available[/yellow]" + ) + return "local", {} # External LangFuse ā€” collect connection details console.print() console.print("[bold]Enter your external LangFuse connection details:[/bold]") - backend_env_path = 'backends/advanced/.env' + backend_env_path = "backends/advanced/.env" - existing_host = read_env_value(backend_env_path, 'LANGFUSE_HOST') + existing_host = read_env_value(backend_env_path, "LANGFUSE_HOST") # Don't treat the local docker host as an existing external value - if existing_host and 'langfuse-web' in existing_host: + if existing_host and "langfuse-web" in existing_host: existing_host = None host = prompt_with_existing_masked( @@ -680,36 +833,38 @@ def setup_langfuse_choice(): existing_value=existing_host, placeholders=[""], is_password=False, - default="https://cloud.langfuse.com" + default="https://cloud.langfuse.com", ) - existing_pub = read_env_value(backend_env_path, 'LANGFUSE_PUBLIC_KEY') + existing_pub = read_env_value(backend_env_path, "LANGFUSE_PUBLIC_KEY") public_key = prompt_with_existing_masked( prompt_text="LangFuse public key", existing_value=existing_pub, placeholders=[""], is_password=False, - default="" + default="", ) - existing_sec = read_env_value(backend_env_path, 'LANGFUSE_SECRET_KEY') + existing_sec = read_env_value(backend_env_path, "LANGFUSE_SECRET_KEY") secret_key = prompt_with_existing_masked( prompt_text="LangFuse secret key", existing_value=existing_sec, placeholders=[""], is_password=True, - default="" + default="", ) if not (host and public_key and secret_key): - console.print("[yellow]āš ļø Incomplete LangFuse configuration ā€” skipping[/yellow]") + console.print( + "[yellow]āš ļø Incomplete LangFuse configuration ā€” skipping[/yellow]" + ) return None, {} console.print(f"[green]āœ…[/green] External LangFuse configured: {host}") - return 'external', { - 'host': host, - 'public_key': public_key, - 'secret_key': secret_key, + return "external", { + "host": host, + "public_key": public_key, + "secret_key": secret_key, } @@ -717,8 +872,12 @@ def main(): """Main orchestration logic""" console.print("šŸŽ‰ [bold green]Welcome to Chronicle![/bold green]\n") console.print("[dim]This wizard is safe to run as many times as you like.[/dim]") - console.print("[dim]It backs up your existing config and preserves previously entered values.[/dim]") - console.print("[dim]When unsure, just press Enter ā€” the defaults will work.[/dim]\n") + console.print( + "[dim]It backs up your existing config and preserves previously entered values.[/dim]" + ) + console.print( + "[dim]When unsure, just press Enter ā€” the defaults will work.[/dim]\n" + ) # Setup config file from template setup_config_file() @@ -740,11 +899,19 @@ def main(): # Auto-add asr-services if any local ASR was chosen (batch or streaming) local_asr_providers = ("parakeet", "vibevoice", "qwen3-asr") - needs_asr = transcription_provider in local_asr_providers or (streaming_provider and streaming_provider in local_asr_providers) - if needs_asr and 'asr-services' not in selected_services: - reason = transcription_provider if transcription_provider in local_asr_providers else streaming_provider - console.print(f"[blue][INFO][/blue] Auto-adding ASR services for {reason} transcription") - selected_services.append('asr-services') + needs_asr = transcription_provider in local_asr_providers or ( + streaming_provider and streaming_provider in local_asr_providers + ) + if needs_asr and "asr-services" not in selected_services: + reason = ( + transcription_provider + if transcription_provider in local_asr_providers + else streaming_provider + ) + console.print( + f"[blue][INFO][/blue] Auto-adding ASR services for {reason} transcription" + ) + selected_services.append("asr-services") if not selected_services: console.print("\n[yellow]No services selected. Exiting.[/yellow]") @@ -752,8 +919,8 @@ def main(): # LangFuse Configuration (before service setup so keys can be passed to backend) langfuse_mode, langfuse_external = setup_langfuse_choice() - if langfuse_mode == 'local' and 'langfuse' not in selected_services: - selected_services.append('langfuse') + if langfuse_mode == "local" and "langfuse" not in selected_services: + selected_services.append("langfuse") # HF Token Configuration (if services require it) hf_token = setup_hf_token_if_needed(selected_services) @@ -761,17 +928,24 @@ def main(): # HTTPS Configuration (for services that need it) https_enabled = False server_ip = None - + # Check if we have services that benefit from HTTPS - https_services = {'advanced', 'speaker-recognition'} # advanced will always need https then + https_services = { + "advanced", + "speaker-recognition", + } # advanced will always need https then needs_https = bool(https_services.intersection(selected_services)) - + if needs_https: console.print("\nšŸ”’ [bold cyan]HTTPS Configuration[/bold cyan]") - console.print("HTTPS enables microphone access in browsers and secure connections") + console.print( + "HTTPS enables microphone access in browsers and secure connections" + ) try: - https_enabled = Confirm.ask("Enable HTTPS for selected services?", default=False) + https_enabled = Confirm.ask( + "Enable HTTPS for selected services?", default=False + ) except EOFError: console.print("Using default: No") https_enabled = False @@ -781,24 +955,30 @@ def main(): ts_dns, ts_ip = detect_tailscale_info() if ts_dns: - console.print(f"\n[green][AUTO-DETECTED][/green] Tailscale DNS: {ts_dns}") + console.print( + f"\n[green][AUTO-DETECTED][/green] Tailscale DNS: {ts_dns}" + ) if ts_ip: - console.print(f"[green][AUTO-DETECTED][/green] Tailscale IP: {ts_ip}") + console.print( + f"[green][AUTO-DETECTED][/green] Tailscale IP: {ts_ip}" + ) default_address = ts_dns elif ts_ip: console.print(f"\n[green][AUTO-DETECTED][/green] Tailscale IP: {ts_ip}") default_address = ts_ip else: console.print("\n[blue][INFO][/blue] Tailscale not detected") - console.print("[blue][INFO][/blue] To find your Tailscale address: tailscale status --json | jq -r '.Self.DNSName'") + console.print( + "[blue][INFO][/blue] To find your Tailscale address: tailscale status --json | jq -r '.Self.DNSName'" + ) default_address = None console.print("[blue][INFO][/blue] For local-only access, use 'localhost'") console.print("Examples: localhost, myhost.tail1234.ts.net, 100.64.1.2") # Check for existing SERVER_IP from backend .env - backend_env_path = 'backends/advanced/.env' - existing_ip = read_env_value(backend_env_path, 'SERVER_IP') + backend_env_path = "backends/advanced/.env" + existing_ip = read_env_value(backend_env_path, "SERVER_IP") # Use existing value, or auto-detected address, or localhost as default effective_default = default_address or "localhost" @@ -806,9 +986,9 @@ def main(): server_ip = prompt_with_existing_masked( prompt_text="Server IP/Domain for SSL certificates", existing_value=existing_ip, - placeholders=['localhost', 'your-server-ip-here'], + placeholders=["localhost", "your-server-ip-here"], is_password=False, - default=effective_default + default=effective_default, ) console.print(f"[green]āœ…[/green] HTTPS configured for: {server_ip}") @@ -817,14 +997,18 @@ def main(): neo4j_password = None obsidian_enabled = False - if 'advanced' in selected_services: + if "advanced" in selected_services: console.print("\nšŸ—„ļø [bold cyan]Neo4j Configuration[/bold cyan]") - console.print("Neo4j is used for Knowledge Graph (entity/relationship extraction from conversations)") + console.print( + "Neo4j is used for Knowledge Graph (entity/relationship extraction from conversations)" + ) console.print() # Always prompt for Neo4j password (masked input) try: - console.print("Neo4j password (min 8 chars) [leave empty for default: neo4jpassword]") + console.print( + "Neo4j password (min 8 chars) [leave empty for default: neo4jpassword]" + ) neo4j_password = prompt_password("Neo4j password", min_length=8) except (EOFError, KeyboardInterrupt): neo4j_password = "neo4jpassword" @@ -836,11 +1020,15 @@ def main(): # Obsidian is optional (graph-based knowledge management for vault notes) console.print("\nšŸ—‚ļø [bold cyan]Obsidian Integration (Optional)[/bold cyan]") - console.print("Enable graph-based knowledge management for Obsidian vault notes") + console.print( + "Enable graph-based knowledge management for Obsidian vault notes" + ) console.print() try: - obsidian_enabled = Confirm.ask("Enable Obsidian integration?", default=False) + obsidian_enabled = Confirm.ask( + "Enable Obsidian integration?", default=False + ) except EOFError: console.print("Using default: No") obsidian_enabled = False @@ -858,40 +1046,59 @@ def main(): failed_services = [] # Pre-populate langfuse keys from external config (if user chose external mode) - langfuse_public_key = langfuse_external.get('public_key') - langfuse_secret_key = langfuse_external.get('secret_key') - langfuse_host = langfuse_external.get('host') # None for local (backend defaults to langfuse-web) + langfuse_public_key = langfuse_external.get("public_key") + langfuse_secret_key = langfuse_external.get("secret_key") + langfuse_host = langfuse_external.get( + "host" + ) # None for local (backend defaults to langfuse-web) # Determine setup order: langfuse first (to get API keys), then backend (with langfuse keys), then others setup_order = [] - if 'langfuse' in selected_services: - setup_order.append('langfuse') - if 'advanced' in selected_services: - setup_order.append('advanced') + if "langfuse" in selected_services: + setup_order.append("langfuse") + if "advanced" in selected_services: + setup_order.append("advanced") for service in selected_services: if service not in setup_order: setup_order.append(service) # Read admin credentials from existing backend .env (for langfuse init reuse) - backend_env_path = 'backends/advanced/.env' - wizard_admin_email = read_env_value(backend_env_path, 'ADMIN_EMAIL') - wizard_admin_password = read_env_value(backend_env_path, 'ADMIN_PASSWORD') + backend_env_path = "backends/advanced/.env" + wizard_admin_email = read_env_value(backend_env_path, "ADMIN_EMAIL") + wizard_admin_password = read_env_value(backend_env_path, "ADMIN_PASSWORD") for service in setup_order: - if run_service_setup(service, selected_services, https_enabled, server_ip, - obsidian_enabled, neo4j_password, hf_token, transcription_provider, - admin_email=wizard_admin_email, admin_password=wizard_admin_password, - langfuse_public_key=langfuse_public_key, langfuse_secret_key=langfuse_secret_key, - langfuse_host=langfuse_host, streaming_provider=streaming_provider): + if run_service_setup( + service, + selected_services, + https_enabled, + server_ip, + obsidian_enabled, + neo4j_password, + hf_token, + transcription_provider, + admin_email=wizard_admin_email, + admin_password=wizard_admin_password, + langfuse_public_key=langfuse_public_key, + langfuse_secret_key=langfuse_secret_key, + langfuse_host=langfuse_host, + streaming_provider=streaming_provider, + ): success_count += 1 # After local langfuse setup, read generated API keys for backend - if service == 'langfuse': - langfuse_env_path = 'extras/langfuse/.env' - langfuse_public_key = read_env_value(langfuse_env_path, 'LANGFUSE_INIT_PROJECT_PUBLIC_KEY') - langfuse_secret_key = read_env_value(langfuse_env_path, 'LANGFUSE_INIT_PROJECT_SECRET_KEY') + if service == "langfuse": + langfuse_env_path = "extras/langfuse/.env" + langfuse_public_key = read_env_value( + langfuse_env_path, "LANGFUSE_INIT_PROJECT_PUBLIC_KEY" + ) + langfuse_secret_key = read_env_value( + langfuse_env_path, "LANGFUSE_INIT_PROJECT_SECRET_KEY" + ) if langfuse_public_key and langfuse_secret_key: - console.print("[blue][INFO][/blue] LangFuse API keys will be passed to backend configuration") + console.print( + "[blue][INFO][/blue] LangFuse API keys will be passed to backend configuration" + ) else: failed_services.append(service) @@ -902,11 +1109,13 @@ def main(): # Final Summary console.print(f"\nšŸŽŠ [bold green]Setup Complete![/bold green]") - console.print(f"āœ… {success_count}/{len(selected_services)} services configured successfully") + console.print( + f"āœ… {success_count}/{len(selected_services)} services configured successfully" + ) if failed_services: console.print(f"āŒ Failed services: {', '.join(failed_services)}") - + # Next Steps console.print("\nšŸ“– [bold]Next Steps:[/bold]") @@ -914,68 +1123,101 @@ def main(): console.print("") console.print("šŸ“ [bold cyan]Configuration Files Updated:[/bold cyan]") console.print(" ā€¢ [green].env files[/green] - API keys and service URLs") - console.print(" ā€¢ [green]config.yml[/green] - Model definitions and memory provider settings") + console.print( + " ā€¢ [green]config.yml[/green] - Model definitions and memory provider settings" + ) console.print("") # Development Environment Setup console.print("1. Setup development environment (git hooks, testing):") console.print(" [cyan]make setup-dev[/cyan]") - console.print(" [dim]This installs pre-commit hooks to run tests before pushing[/dim]") + console.print( + " [dim]This installs pre-commit hooks to run tests before pushing[/dim]" + ) console.print("") # Service Management Commands console.print("2. Start all configured services:") console.print(" [cyan]./start.sh[/cyan]") - console.print(" [dim]Or: uv run --with-requirements setup-requirements.txt python services.py start --all --build[/dim]") + console.print( + " [dim]Or: uv run --with-requirements setup-requirements.txt python services.py start --all --build[/dim]" + ) console.print("") console.print("3. Or start individual services:") - + configured_services = [] - if 'advanced' in selected_services and 'advanced' not in failed_services: + if "advanced" in selected_services and "advanced" not in failed_services: configured_services.append("backend") - if 'speaker-recognition' in selected_services and 'speaker-recognition' not in failed_services: - configured_services.append("speaker-recognition") - if 'asr-services' in selected_services and 'asr-services' not in failed_services: + if ( + "speaker-recognition" in selected_services + and "speaker-recognition" not in failed_services + ): + configured_services.append("speaker-recognition") + if "asr-services" in selected_services and "asr-services" not in failed_services: configured_services.append("asr-services") - if 'openmemory-mcp' in selected_services and 'openmemory-mcp' not in failed_services: + if ( + "openmemory-mcp" in selected_services + and "openmemory-mcp" not in failed_services + ): configured_services.append("openmemory-mcp") - if 'langfuse' in selected_services and 'langfuse' not in failed_services: + if "langfuse" in selected_services and "langfuse" not in failed_services: configured_services.append("langfuse") # LangFuse prompt management info - if langfuse_mode == 'local' and 'langfuse' not in failed_services: + if langfuse_mode == "local" and "langfuse" not in failed_services: console.print("") - console.print("[bold cyan]Prompt Management:[/bold cyan] Once services are running, edit AI prompts at:") + console.print( + "[bold cyan]Prompt Management:[/bold cyan] Once services are running, edit AI prompts at:" + ) if https_enabled and server_ip: - console.print(f" [link=https://{server_ip}:3443/project/chronicle/prompts]https://{server_ip}:3443/project/chronicle/prompts[/link]") + console.print( + f" [link=https://{server_ip}:3443/project/chronicle/prompts]https://{server_ip}:3443/project/chronicle/prompts[/link]" + ) else: - console.print(" [link=http://localhost:3002/project/chronicle/prompts]http://localhost:3002/project/chronicle/prompts[/link]") - elif langfuse_mode == 'external' and langfuse_host: + console.print( + " [link=http://localhost:3002/project/chronicle/prompts]http://localhost:3002/project/chronicle/prompts[/link]" + ) + elif langfuse_mode == "external" and langfuse_host: console.print("") - console.print(f"[bold cyan]Prompt Management:[/bold cyan] Edit AI prompts at your LangFuse instance:") + console.print( + f"[bold cyan]Prompt Management:[/bold cyan] Edit AI prompts at your LangFuse instance:" + ) console.print(f" {langfuse_host}") if configured_services: service_list = " ".join(configured_services) - console.print(f" [cyan]uv run --with-requirements setup-requirements.txt python services.py start {service_list}[/cyan]") - + console.print( + f" [cyan]uv run --with-requirements setup-requirements.txt python services.py start {service_list}[/cyan]" + ) + console.print("") console.print("3. Check service status:") console.print(" [cyan]./status.sh[/cyan]") - console.print(" [dim]Or: uv run --with-requirements setup-requirements.txt python services.py status[/dim]") + console.print( + " [dim]Or: uv run --with-requirements setup-requirements.txt python services.py status[/dim]" + ) console.print("") console.print("4. Stop services when done:") console.print(" [cyan]./stop.sh[/cyan]") - console.print(" [dim]Or: uv run --with-requirements setup-requirements.txt python services.py stop --all[/dim]") - + console.print( + " [dim]Or: uv run --with-requirements setup-requirements.txt python services.py stop --all[/dim]" + ) + console.print(f"\nšŸš€ [bold]Enjoy Chronicle![/bold]") - + # Show individual service usage console.print(f"\nšŸ’” [dim]Tip: You can also setup services individually:[/dim]") - console.print(f"[dim] cd backends/advanced && uv run --with-requirements ../../setup-requirements.txt python init.py[/dim]") - console.print(f"[dim] cd extras/speaker-recognition && uv run --with-requirements ../../setup-requirements.txt python init.py[/dim]") - console.print(f"[dim] cd extras/asr-services && uv run --with-requirements ../../setup-requirements.txt python init.py[/dim]") + console.print( + f"[dim] cd backends/advanced && uv run --with-requirements ../../setup-requirements.txt python init.py[/dim]" + ) + console.print( + f"[dim] cd extras/speaker-recognition && uv run --with-requirements ../../setup-requirements.txt python init.py[/dim]" + ) + console.print( + f"[dim] cd extras/asr-services && uv run --with-requirements ../../setup-requirements.txt python init.py[/dim]" + ) + if __name__ == "__main__": - main() \ No newline at end of file + main() diff --git a/wizard.sh b/wizard.sh index 02942349..fb749554 100755 --- a/wizard.sh +++ b/wizard.sh @@ -1 +1,3 @@ +#!/bin/bash +source "$(dirname "$0")/scripts/check_uv.sh" uv run --with-requirements setup-requirements.txt wizard.py From d59a788ca4e3c1243f1defaa496313f8bec18d1d Mon Sep 17 00:00:00 2001 From: Ankush Malaker <43288948+AnkushMalaker@users.noreply.github.com> Date: Sun, 22 Feb 2026 11:37:30 +0000 Subject: [PATCH 06/11] Refactor test workflows to utilize Makefile targets and enhance documentation - Replaced the `run-no-api-tests.sh` script with a Makefile target `make test-no-api` for executing tests without API keys, streamlining the testing process. - Updated GitHub Actions workflows and README documentation to reflect the new Makefile usage, improving clarity for contributors. - Removed the deprecated `run-no-api-tests.sh` script to reduce redundancy and simplify the codebase. --- .github/workflows/README.md | 4 +- .github/workflows/robot-tests.yml | 6 +- tests/TESTING_GUIDELINES.md | 6 +- tests/run-no-api-tests.sh | 176 ------------------------------ 4 files changed, 7 insertions(+), 185 deletions(-) delete mode 100755 tests/run-no-api-tests.sh diff --git a/.github/workflows/README.md b/.github/workflows/README.md index 0b8987c5..bacc73ab 100644 --- a/.github/workflows/README.md +++ b/.github/workflows/README.md @@ -32,7 +32,7 @@ on: - **No secrets required** - Works for external contributors - **Excludes**: Tests tagged with `requires-api-keys` - **Config**: `tests/configs/mock-services.yml` -- **Test Script**: `./run-no-api-tests.sh` +- **Makefile Target**: `make test-no-api OUTPUTDIR=results-no-api` - **Results**: `results-no-api/` - **Time**: ~10-15 minutes - **Coverage**: ~70% of test suite @@ -312,7 +312,7 @@ runs-on: ubuntu-latest CLEANUP_CONTAINERS: "false" # Handled by workflow # API keys if needed run: | - ./run-{no-api|robot}-tests.sh + make test-no-api # or ./run-robot-tests.sh for full suite TEST_EXIT_CODE=$? echo "test_exit_code=$TEST_EXIT_CODE" >> $GITHUB_ENV exit 0 # Don't fail yet diff --git a/.github/workflows/robot-tests.yml b/.github/workflows/robot-tests.yml index 35e4dffa..603025b8 100644 --- a/.github/workflows/robot-tests.yml +++ b/.github/workflows/robot-tests.yml @@ -76,11 +76,9 @@ jobs: - name: Run Robot Framework tests (No API Keys) working-directory: tests - env: - CLEANUP_CONTAINERS: "false" # Don't cleanup in CI - handled by workflow run: | - # Use the no-API test script (excludes tests tagged with requires-api-keys) - ./run-no-api-tests.sh + # Use Makefile target (starts containers with mock config, excludes api-keys/slow/sdk/gpu tests) + make test-no-api OUTPUTDIR=results-no-api TEST_EXIT_CODE=$? echo "test_exit_code=$TEST_EXIT_CODE" >> $GITHUB_ENV exit 0 # Don't fail here, we'll fail at the end after uploading artifacts diff --git a/tests/TESTING_GUIDELINES.md b/tests/TESTING_GUIDELINES.md index 481b8017..f8c18e43 100644 --- a/tests/TESTING_GUIDELINES.md +++ b/tests/TESTING_GUIDELINES.md @@ -266,7 +266,7 @@ Full Pipeline Integration Test ```bash # Excludes tests tagged with requires-api-keys cd tests -./run-no-api-tests.sh +make test-no-api ``` - Uses `configs/mock-services.yml` - No external API calls @@ -375,7 +375,7 @@ Audio Upload Produces Quality Transcript **Running Tests Locally Without API Keys**: ```bash cd tests -./run-no-api-tests.sh +make test-no-api ``` - Works without any API key configuration - Fast feedback for most development @@ -532,4 +532,4 @@ As we develop more conventions and encounter new patterns, we will add them to t - Continuous integration considerations - Test reporting and metrics - Parallel test execution patterns -- Test data isolation strategies \ No newline at end of file +- Test data isolation strategies diff --git a/tests/run-no-api-tests.sh b/tests/run-no-api-tests.sh deleted file mode 100755 index b5c5a505..00000000 --- a/tests/run-no-api-tests.sh +++ /dev/null @@ -1,176 +0,0 @@ -#!/bin/bash - -# Robot Framework Test Runner (No API Keys Required) -# Runs tests that don't require external API services (Deepgram, OpenAI) -# Excludes tests tagged with 'requires-api-keys' - -set -e - -# Colors for output -RED='\033[0;31m' -GREEN='\033[0;32m' -YELLOW='\033[1;33m' -BLUE='\033[0;34m' -NC='\033[0m' # No Color - -print_info() { - echo -e "${BLUE}[INFO]${NC} $1" -} - -print_success() { - echo -e "${GREEN}[SUCCESS]${NC} $1" -} - -print_warning() { - echo -e "${YELLOW}[WARNING]${NC} $1" -} - -print_error() { - echo -e "${RED}[ERROR]${NC} $1" -} - -# Check if we're in the right directory -if [ ! -f "Makefile" ] || [ ! -d "endpoints" ]; then - print_error "Please run this script from the tests/ directory" - exit 1 -fi - -# Set absolute paths for consistent directory references -TESTS_DIR="$(pwd)" -BACKEND_DIR="$(cd ../backends/advanced && pwd)" - -print_info "Robot Framework Test Runner (No API Keys)" -print_info "==========================================" -print_info "This runner executes tests that don't require external API services" -print_info "Tests tagged with 'requires-api-keys', 'slow', and 'sdk' are excluded" - -# Configuration -CLEANUP_CONTAINERS="${CLEANUP_CONTAINERS:-false}" -OUTPUTDIR="${OUTPUTDIR:-results-no-api}" - -# Use mock services config (no API keys needed) -# Set TEST_CONFIG_FILE to point to mock-services.yml inside the container -export TEST_CONFIG_FILE="/app/test-configs/mock-services.yml" - -print_info "Using config file: ${TEST_CONFIG_FILE}" -print_warning "Memory extraction and transcription are disabled in this mode" - -# Load environment variables if available (but don't require them) -if [ -f "setup/.env.test" ]; then - print_info "Loading environment variables from setup/.env.test..." - set -a - source setup/.env.test - set +a -fi - -# Create test environment file if it doesn't exist (without API keys) -if [ ! -f "setup/.env.test" ]; then - print_info "Creating test environment file..." - mkdir -p setup - - cat > setup/.env.test << EOF -# API URLs -API_URL=http://localhost:8001 -BACKEND_URL=http://localhost:8001 -FRONTEND_URL=http://localhost:3001 - -# Test Admin Credentials -ADMIN_EMAIL=test-admin@example.com -ADMIN_PASSWORD=test-admin-password-123 - -# Test Configuration -TEST_TIMEOUT=120 -TEST_DEVICE_NAME=robot-test - -# Note: No API keys required for this test mode -# OPENAI_API_KEY and DEEPGRAM_API_KEY are not needed -EOF - print_success "Created setup/.env.test" -fi - -# Start test containers using dedicated startup script -FRESH_BUILD=true "$TESTS_DIR/setup-test-containers.sh" - -# Run Robot Framework tests via Makefile with tag exclusion -# Exclude tests that require API keys, slow tests, and SDK tests -print_info "Running Robot Framework tests (excluding requires-api-keys, slow, sdk tags)..." -print_info "Output directory: $OUTPUTDIR" - -# Run tests with tag exclusion -if timeout 30m uv run --with-requirements test-requirements.txt \ - robot --exclude requires-api-keys \ - --exclude slow \ - --exclude sdk \ - --outputdir "$OUTPUTDIR" \ - --loglevel INFO \ - --consolecolors on \ - --consolemarkers on \ - .; then - TEST_EXIT_CODE=0 -else - TEST_EXIT_CODE=$? -fi - -# Show service logs if tests failed -if [ $TEST_EXIT_CODE -ne 0 ]; then - print_info "Showing service logs..." - cd "$BACKEND_DIR" - echo "=== Backend Logs (last 50 lines) ===" - docker compose -f docker-compose-test.yml logs --tail=50 chronicle-backend-test - echo "" - echo "=== Worker Logs (last 50 lines) ===" - docker compose -f docker-compose-test.yml logs --tail=50 workers-test - cd "$TESTS_DIR" -fi - -# Display test results summary -if [ -f "$OUTPUTDIR/output.xml" ]; then - print_info "Test Results Summary:" - uv run python3 << 'PYTHON_SCRIPT' -import xml.etree.ElementTree as ET -import os - -output_file = os.getenv('OUTPUTDIR', 'results-no-api') + '/output.xml' -try: - tree = ET.parse(output_file) - root = tree.getroot() - - # Get overall stats - stats = root.find('.//total/stat') - if stats is not None: - passed = stats.get("pass", "0") - failed = stats.get("fail", "0") - print(f'āœ… Passed: {passed}') - print(f'āŒ Failed: {failed}') - print(f'šŸ“Š Total: {int(passed) + int(failed)}') - - # Show failed tests if any - if int(failed) > 0: - print('\nāŒ Failed Tests:') - failed_tests = root.findall('.//test') - for test in failed_tests: - status = test.find('status') - if status is not None and status.get('status') == 'FAIL': - test_name = test.get('name', 'Unknown') - print(f' - {test_name}') -except Exception as e: - print(f'Error parsing results: {e}') -PYTHON_SCRIPT -fi - -# Cleanup containers if requested -if [ "$CLEANUP_CONTAINERS" = "true" ]; then - print_info "Cleaning up test containers..." - cd "$BACKEND_DIR" - docker compose -f docker-compose-test.yml down -v --remove-orphans - cd "$TESTS_DIR" - print_success "Cleanup completed" -fi - -# Final status -if [ $TEST_EXIT_CODE -eq 0 ]; then - print_success "All tests passed! āœ…" -else - print_error "Some tests failed āŒ" - exit $TEST_EXIT_CODE -fi From 98b5dbfd4e853cd7268fa194e12809e7e69a3dcb Mon Sep 17 00:00:00 2001 From: Ankush Malaker <43288948+AnkushMalaker@users.noreply.github.com> Date: Sun, 22 Feb 2026 14:52:24 +0000 Subject: [PATCH 07/11] Update user loop with keyboard shortcuts - Added new service configuration for LangFuse in `services.py`, including path, compose file, description, and ports. - Refactored the backend initialization process to improve observability and prompt management with LangFuse. - Updated the `load_config_yml` function to ensure proper loading of LangFuse settings. - Enhanced error handling and logging during service startup to provide better visibility into configuration issues. --- .../src/advanced_omi_backend/app_factory.py | 373 +++++++----- .../controllers/conversation_controller.py | 238 +++----- .../controllers/memory_controller.py | 93 +-- .../advanced_omi_backend/models/annotation.py | 40 +- .../advanced_omi_backend/prompt_defaults.py | 36 ++ .../routers/modules/annotation_routes.py | 192 ++++++- .../routers/modules/finetuning_routes.py | 197 ++++--- .../services/plugin_service.py | 97 +++- .../advanced_omi_backend/utils/job_utils.py | 39 +- .../workers/annotation_jobs.py | 247 +++++--- .../workers/conversation_jobs.py | 108 ++-- .../workers/memory_jobs.py | 51 +- .../workers/speaker_jobs.py | 232 ++++---- .../workers/transcription_jobs.py | 89 +-- backends/advanced/webui/package-lock.json | 44 +- backends/advanced/webui/package.json | 1 + .../webui/src/components/UserLoopModal.tsx | 533 ++++++++++++++++++ .../webui/src/components/layout/Layout.tsx | 8 +- .../advanced/webui/src/pages/Finetuning.tsx | 70 ++- config/defaults.yml | 4 + .../providers/vibevoice/transcriber.py | 88 ++- extras/ml-training/README.md | 39 ++ extras/ml-training/event-detection/README.md | 476 ++++++++++++++++ .../event-detection/export_from_mongo.py | 147 +++++ .../event-detection/manage_data.py | 167 ++++++ .../event-detection/requirements.txt | 13 + extras/ml-training/event-detection/train.py | 258 +++++++++ .../whisper-adapter-finetuning/README.md | 202 +++++++ .../evaluate_sneeze_model.py | 123 ++++ .../prepare_sneeze_data.py | 80 +++ .../sneeze_data.jsonl | 23 + .../train_sneeze.py | 295 ++++++++++ services.py | 462 ++++++++------- .../annotation_suggestions_tests.robot | 65 +++ 34 files changed, 4060 insertions(+), 1070 deletions(-) create mode 100644 backends/advanced/webui/src/components/UserLoopModal.tsx create mode 100644 extras/ml-training/README.md create mode 100644 extras/ml-training/event-detection/README.md create mode 100755 extras/ml-training/event-detection/export_from_mongo.py create mode 100644 extras/ml-training/event-detection/manage_data.py create mode 100644 extras/ml-training/event-detection/requirements.txt create mode 100644 extras/ml-training/event-detection/train.py create mode 100644 extras/ml-training/whisper-adapter-finetuning/README.md create mode 100644 extras/ml-training/whisper-adapter-finetuning/evaluate_sneeze_model.py create mode 100644 extras/ml-training/whisper-adapter-finetuning/prepare_sneeze_data.py create mode 100644 extras/ml-training/whisper-adapter-finetuning/sneeze_data.jsonl create mode 100644 extras/ml-training/whisper-adapter-finetuning/train_sneeze.py create mode 100644 tests/endpoints/annotation_suggestions_tests.robot diff --git a/backends/advanced/src/advanced_omi_backend/app_factory.py b/backends/advanced/src/advanced_omi_backend/app_factory.py index 6083de97..c1d56fed 100644 --- a/backends/advanced/src/advanced_omi_backend/app_factory.py +++ b/backends/advanced/src/advanced_omi_backend/app_factory.py @@ -7,6 +7,7 @@ import asyncio import logging +import time from contextlib import asynccontextmanager from pathlib import Path @@ -122,10 +123,14 @@ async def initialize_openmemory_user() -> None: async def lifespan(app: FastAPI): """Manage application lifespan events.""" config = get_app_config() + startup_start = time.monotonic() # Startup application_logger.info("Starting application...") + # ā”€ā”€ Phase 1 (sequential ā€” dependencies) ā”€ā”€ā”€ā”€ā”€ā”€ā”€ā”€ā”€ā”€ā”€ā”€ā”€ā”€ā”€ā”€ā”€ā”€ā”€ā”€ā”€ā”€ā”€ā”€ā”€ā”€ + phase_start = time.monotonic() + # Initialize Beanie for all document models try: from beanie import init_beanie @@ -151,200 +156,258 @@ async def lifespan(app: FastAPI): application_logger.error(f"Failed to initialize Beanie: {e}") raise - # Create admin user if needed + # Create admin user if needed (requires Beanie) try: await create_admin_user_if_needed() except Exception as e: application_logger.error(f"Failed to create admin user: {e}") - # Don't raise here as this is not critical for startup - # Initialize Redis connection for RQ - try: - from advanced_omi_backend.controllers.queue_controller import redis_conn + application_logger.info( + f"Phase 1 (Beanie + admin) completed in {time.monotonic() - phase_start:.2f}s" + ) - redis_conn.ping() - application_logger.info("Redis connection established for RQ") - application_logger.info( - "RQ workers can be started with: rq worker transcription memory default" - ) - except Exception as e: - application_logger.error(f"Failed to connect to Redis for RQ: {e}") - application_logger.warning( - "RQ queue system will not be available - check Redis connection" - ) + # ā”€ā”€ Phase 2 (parallel ā€” all independent) ā”€ā”€ā”€ā”€ā”€ā”€ā”€ā”€ā”€ā”€ā”€ā”€ā”€ā”€ā”€ā”€ā”€ā”€ā”€ā”€ā”€ā”€ā”€ā”€ā”€ + phase_start = time.monotonic() - # Initialize BackgroundTaskManager (must happen before any code path uses it) - try: - task_manager = init_task_manager() - await task_manager.start() - application_logger.info("BackgroundTaskManager initialized and started") - except Exception as e: - application_logger.error(f"Failed to initialize task manager: {e}") - raise # Task manager is essential + async def _init_redis_rq(): + try: + from advanced_omi_backend.controllers.queue_controller import redis_conn - # Initialize ClientManager eagerly (prevents lazy race on first WebSocket connect) - get_client_manager() - application_logger.info("ClientManager initialized") + redis_conn.ping() + application_logger.info("Redis connection established for RQ") + except Exception as e: + application_logger.error(f"Failed to connect to Redis for RQ: {e}") + application_logger.warning( + "RQ queue system will not be available - check Redis connection" + ) - # Initialize OTEL/Galileo if configured (before LLM client so instrumentor patches OpenAI first) - try: - from advanced_omi_backend.observability.otel_setup import init_otel + async def _init_task_manager(): + try: + tm = init_task_manager() + await tm.start() + application_logger.info("BackgroundTaskManager initialized and started") + except Exception as e: + application_logger.error(f"Failed to initialize task manager: {e}") + raise # Task manager is essential - init_otel() - except Exception as e: - application_logger.warning(f"OTEL initialization skipped: {e}") + async def _init_client_manager(): + get_client_manager() + application_logger.info("ClientManager initialized") - # Initialize prompt registry with defaults; seed into LangFuse in background - try: - from advanced_omi_backend.prompt_defaults import register_all_defaults - from advanced_omi_backend.prompt_registry import get_prompt_registry + async def _init_otel(): + try: + from advanced_omi_backend.observability.otel_setup import init_otel - prompt_registry = get_prompt_registry() - register_all_defaults(prompt_registry) - application_logger.info( - f"Prompt registry initialized with {len(prompt_registry._defaults)} defaults" - ) + init_otel() + except Exception as e: + application_logger.warning(f"OTEL initialization skipped: {e}") - # Seed prompts in background ā€” Langfuse may not be ready at startup - async def _deferred_seed(): - await asyncio.sleep(10) - await prompt_registry.seed_prompts() + async def _init_prompt_registry(): + try: + from advanced_omi_backend.prompt_defaults import register_all_defaults + from advanced_omi_backend.prompt_registry import get_prompt_registry - asyncio.create_task(_deferred_seed()) - except Exception as e: - application_logger.warning(f"Prompt registry initialization failed: {e}") + registry = get_prompt_registry() + register_all_defaults(registry) + application_logger.info( + f"Prompt registry initialized with {len(registry._defaults)} defaults" + ) + except Exception as e: + application_logger.warning(f"Prompt registry initialization failed: {e}") + + await asyncio.gather( + _init_redis_rq(), + _init_task_manager(), + _init_client_manager(), + _init_otel(), + _init_prompt_registry(), + ) - # Initialize LLM client eagerly (catch config errors at startup, not on first request) - try: - from advanced_omi_backend.llm_client import get_llm_client + application_logger.info( + f"Phase 2 (Redis/TaskMgr/ClientMgr/OTEL/Prompts) completed in {time.monotonic() - phase_start:.2f}s" + ) - get_llm_client() - application_logger.info("LLM client initialized from config.yml") - except Exception as e: - application_logger.warning(f"LLM client initialization deferred: {e}") + # ā”€ā”€ Phase 3 (parallel ā€” OTEL done, safe for LLM patching) ā”€ā”€ā”€ā”€ā”€ā”€ā”€ā”€ + phase_start = time.monotonic() - # Initialize audio stream service for Redis Streams - try: - audio_service = get_audio_stream_service() - await audio_service.connect() - application_logger.info("Audio stream service connected to Redis Streams") - application_logger.info( - "Audio stream workers can be started with: python -m advanced_omi_backend.workers.audio_stream_worker" - ) - except Exception as e: - application_logger.error(f"Failed to connect audio stream service: {e}") - application_logger.warning( - "Redis Streams audio processing will not be available" - ) + async def _init_llm_client(): + try: + from advanced_omi_backend.llm_client import get_llm_client - # Initialize Redis client for audio streaming producer (used by WebSocket handlers) - try: - app.state.redis_audio_stream = await redis.from_url( - config.redis_url, encoding="utf-8", decode_responses=False - ) - from advanced_omi_backend.services.audio_stream import AudioStreamProducer + get_llm_client() + application_logger.info("LLM client initialized from config.yml") + except Exception as e: + application_logger.warning(f"LLM client initialization deferred: {e}") - app.state.audio_stream_producer = AudioStreamProducer( - app.state.redis_audio_stream - ) - application_logger.info( - "āœ… Redis client for audio streaming producer initialized" - ) + async def _init_audio_stream_service(): + try: + audio_service = get_audio_stream_service() + await audio_service.connect() + application_logger.info("Audio stream service connected to Redis Streams") + except Exception as e: + application_logger.error(f"Failed to connect audio stream service: {e}") + application_logger.warning( + "Redis Streams audio processing will not be available" + ) - # Initialize ClientManager Redis for cross-container clientā†’user mapping - from advanced_omi_backend.client_manager import ( - initialize_redis_for_client_manager, - ) + async def _init_redis_audio_producer(): + try: + app.state.redis_audio_stream = await redis.from_url( + config.redis_url, encoding="utf-8", decode_responses=False + ) + from advanced_omi_backend.services.audio_stream import AudioStreamProducer - initialize_redis_for_client_manager(config.redis_url) + app.state.audio_stream_producer = AudioStreamProducer( + app.state.redis_audio_stream + ) + application_logger.info( + "Redis client for audio streaming producer initialized" + ) - except Exception as e: - application_logger.error( - f"Failed to initialize Redis client for audio streaming: {e}", exc_info=True + from advanced_omi_backend.client_manager import ( + initialize_redis_for_client_manager, + ) + + initialize_redis_for_client_manager(config.redis_url) + except Exception as e: + application_logger.error( + f"Failed to initialize Redis client for audio streaming: {e}", + exc_info=True, + ) + application_logger.warning("Audio streaming producer will not be available") + + async def _deferred_prompt_seed(): + """Seed prompts into Langfuse with retry backoff.""" + try: + from advanced_omi_backend.prompt_registry import get_prompt_registry + + registry = get_prompt_registry() + except Exception: + return + + backoff_delays = [0, 2, 4, 8, 16, 32] + for delay in backoff_delays: + if delay: + await asyncio.sleep(delay) + try: + await registry.seed_prompts() + application_logger.info("Prompt seeding to Langfuse completed") + return + except Exception as e: + application_logger.debug( + f"Prompt seeding attempt failed (next retry in {delay}s): {e}" + ) + application_logger.warning( + "Prompt seeding to Langfuse failed after all retries" ) - application_logger.warning("Audio streaming producer will not be available") - # Skip memory service pre-initialization to avoid blocking FastAPI startup - # Memory service will be lazily initialized when first used + await asyncio.gather( + _init_llm_client(), + _init_audio_stream_service(), + _init_redis_audio_producer(), + ) + + # Launch deferred prompt seeding as a fire-and-forget background task + asyncio.create_task(_deferred_prompt_seed()) + application_logger.info( - "Memory service will be initialized on first use (lazy loading)" + f"Phase 3 (LLM/AudioStream/RedisProducer) completed in {time.monotonic() - phase_start:.2f}s" ) - # Register OpenMemory user if using openmemory_mcp provider - await initialize_openmemory_user() + # ā”€ā”€ Phase 4 (parallel ā€” all independent) ā”€ā”€ā”€ā”€ā”€ā”€ā”€ā”€ā”€ā”€ā”€ā”€ā”€ā”€ā”€ā”€ā”€ā”€ā”€ā”€ā”€ā”€ā”€ā”€ā”€ + phase_start = time.monotonic() - # Start cron scheduler (requires Redis to be available) - try: - from advanced_omi_backend.cron_scheduler import get_scheduler, register_cron_job - from advanced_omi_backend.workers.finetuning_jobs import ( - run_asr_finetuning_job, - run_asr_jargon_extraction_job, - run_speaker_finetuning_job, - ) - from advanced_omi_backend.workers.prompt_optimization_jobs import ( - run_prompt_optimization_job, - ) + application_logger.info( + "Memory service will be initialized on first use (lazy loading)" + ) - register_cron_job("speaker_finetuning", run_speaker_finetuning_job) - register_cron_job("asr_finetuning", run_asr_finetuning_job) - register_cron_job("asr_jargon_extraction", run_asr_jargon_extraction_job) - register_cron_job("prompt_optimization", run_prompt_optimization_job) + async def _init_openmemory(): + await initialize_openmemory_user() - scheduler = get_scheduler() - await scheduler.start() - application_logger.info("Cron scheduler started") - except Exception as e: - application_logger.warning(f"Cron scheduler failed to start: {e}") + async def _init_cron_scheduler(): + try: + from advanced_omi_backend.cron_scheduler import ( + get_scheduler, + register_cron_job, + ) + from advanced_omi_backend.workers.annotation_jobs import ( + surface_error_suggestions, + ) + from advanced_omi_backend.workers.finetuning_jobs import ( + run_asr_finetuning_job, + run_asr_jargon_extraction_job, + run_speaker_finetuning_job, + ) + from advanced_omi_backend.workers.prompt_optimization_jobs import ( + run_prompt_optimization_job, + ) - # SystemTracker is used for monitoring and debugging - application_logger.info("Using SystemTracker for monitoring and debugging") + register_cron_job("speaker_finetuning", run_speaker_finetuning_job) + register_cron_job("asr_finetuning", run_asr_finetuning_job) + register_cron_job("asr_jargon_extraction", run_asr_jargon_extraction_job) + register_cron_job("prompt_optimization", run_prompt_optimization_job) + register_cron_job("annotation_suggestions", surface_error_suggestions) - # Initialize plugins using plugin service - try: - from advanced_omi_backend.services.plugin_service import ( - init_plugin_router, - set_plugin_router, - ) + scheduler = get_scheduler() + await scheduler.start() + application_logger.info("Cron scheduler started") + except Exception as e: + application_logger.warning(f"Cron scheduler failed to start: {e}") - plugin_router = init_plugin_router() - - if plugin_router: - # Initialize async resources for each enabled plugin - for plugin_id, plugin in plugin_router.plugins.items(): - if plugin.enabled: - try: - await plugin.initialize() - plugin_router.mark_plugin_initialized(plugin_id) - application_logger.info(f"āœ… Plugin '{plugin_id}' initialized") - except Exception as e: - plugin_router.mark_plugin_failed(plugin_id, str(e)) - application_logger.error( - f"Failed to initialize plugin '{plugin_id}': {e}", - exc_info=True, - ) - - health = plugin_router.get_health_summary() - application_logger.info( - f"Plugins initialized: {health['initialized']}/{health['total']} active" - + (f", {health['failed']} failed" if health["failed"] else "") + async def _init_plugins(): + try: + from advanced_omi_backend.services.plugin_service import ( + init_plugin_router, + set_plugin_router, ) - # Store in app state for API access - app.state.plugin_router = plugin_router - # Register with plugin service for worker access - set_plugin_router(plugin_router) - else: - application_logger.info("No plugins configured") + plugin_router = init_plugin_router() + + if plugin_router: + for plugin_id, plugin in plugin_router.plugins.items(): + if plugin.enabled: + try: + await plugin.initialize() + plugin_router.mark_plugin_initialized(plugin_id) + application_logger.info(f"Plugin '{plugin_id}' initialized") + except Exception as e: + plugin_router.mark_plugin_failed(plugin_id, str(e)) + application_logger.error( + f"Failed to initialize plugin '{plugin_id}': {e}", + exc_info=True, + ) + + health = plugin_router.get_health_summary() + application_logger.info( + f"Plugins initialized: {health['initialized']}/{health['total']} active" + + (f", {health['failed']} failed" if health["failed"] else "") + ) + + app.state.plugin_router = plugin_router + set_plugin_router(plugin_router) + else: + application_logger.info("No plugins configured") + app.state.plugin_router = None + + except Exception as e: + application_logger.error( + f"Failed to initialize plugin system: {e}", exc_info=True + ) app.state.plugin_router = None - except Exception as e: - application_logger.error( - f"Failed to initialize plugin system: {e}", exc_info=True - ) - app.state.plugin_router = None + await asyncio.gather( + _init_openmemory(), + _init_cron_scheduler(), + _init_plugins(), + ) + + application_logger.info( + f"Phase 4 (OpenMemory/Cron/Plugins) completed in {time.monotonic() - phase_start:.2f}s" + ) + total_startup = time.monotonic() - startup_start application_logger.info( - "Application ready - using application-level processing architecture." + f"Application ready in {total_startup:.2f}s - using application-level processing architecture." ) logger.info("App ready") diff --git a/backends/advanced/src/advanced_omi_backend/controllers/conversation_controller.py b/backends/advanced/src/advanced_omi_backend/controllers/conversation_controller.py index 1bf41dfc..fd27dd38 100644 --- a/backends/advanced/src/advanced_omi_backend/controllers/conversation_controller.py +++ b/backends/advanced/src/advanced_omi_backend/controllers/conversation_controller.py @@ -46,6 +46,25 @@ audio_logger = logging.getLogger("audio_processing") +async def _get_conversation_or_error(conversation_id: str, user: User): + """Fetch a conversation and validate user access. + + Returns (conversation, None) on success, or (None, error_response) on failure. + """ + conversation = await Conversation.find_one( + Conversation.conversation_id == conversation_id + ) + if not conversation: + return None, JSONResponse( + status_code=404, content={"error": "Conversation not found"} + ) + if not user.is_superuser and conversation.user_id != str(user.user_id): + return None, JSONResponse( + status_code=403, content={"error": "Access forbidden"} + ) + return conversation, None + + async def close_current_conversation(client_id: str, user: User): """Close the current conversation for a specific client. @@ -112,18 +131,9 @@ async def close_current_conversation(client_id: str, user: User): async def get_conversation(conversation_id: str, user: User): """Get a single conversation with full transcript details.""" try: - # Find the conversation using Beanie - conversation = await Conversation.find_one( - Conversation.conversation_id == conversation_id - ) - if not conversation: - return JSONResponse( - status_code=404, content={"error": "Conversation not found"} - ) - - # Check ownership for non-admin users - if not user.is_superuser and conversation.user_id != str(user.user_id): - return JSONResponse(status_code=403, content={"error": "Access forbidden"}) + conversation, error = await _get_conversation_or_error(conversation_id, user) + if error: + return error # Build response with explicit curated fields response = { @@ -184,16 +194,9 @@ async def get_conversation(conversation_id: str, user: User): async def get_conversation_memories(conversation_id: str, user: User, limit: int = 100): """Get memories extracted from a specific conversation.""" try: - conversation = await Conversation.find_one( - Conversation.conversation_id == conversation_id - ) - if not conversation: - return JSONResponse( - status_code=404, content={"error": "Conversation not found"} - ) - - if not user.is_superuser and conversation.user_id != str(user.user_id): - return JSONResponse(status_code=403, content={"error": "Access forbidden"}) + conversation, error = await _get_conversation_or_error(conversation_id, user) + if error: + return error memory_service = get_memory_service() memories = await memory_service.get_memories_by_source( @@ -671,29 +674,13 @@ async def delete_conversation( f"Attempting to {'permanently ' if permanent else ''}delete conversation: {masked_id}" ) - # Find the conversation using Beanie - conversation = await Conversation.find_one( - Conversation.conversation_id == conversation_id - ) - - if not conversation: - return JSONResponse( - status_code=404, - content={"error": f"Conversation '{conversation_id}' not found"}, - ) - - # Check ownership for non-admin users - if not user.is_superuser and conversation.user_id != str(user.user_id): - logger.warning( - f"User {user.user_id} attempted to delete conversation {conversation_id} without permission" - ) - return JSONResponse( - status_code=403, - content={ - "error": "Access forbidden. You can only delete your own conversations.", - "details": f"Conversation '{conversation_id}' does not belong to your account.", - }, - ) + conversation, error = await _get_conversation_or_error(conversation_id, user) + if error: + if error.status_code == 403: + logger.warning( + f"User {user.user_id} attempted to delete conversation {conversation_id} without permission" + ) + return error # Hard delete (admin only, permanent flag) if permanent and user.is_superuser: @@ -719,18 +706,9 @@ async def restore_conversation(conversation_id: str, user: User) -> JSONResponse user: Requesting user """ try: - conversation = await Conversation.find_one( - Conversation.conversation_id == conversation_id - ) - - if not conversation: - return JSONResponse( - status_code=404, content={"error": "Conversation not found"} - ) - - # Permission check - if not user.is_superuser and conversation.user_id != str(user.user_id): - return JSONResponse(status_code=403, content={"error": "Access denied"}) + conversation, error = await _get_conversation_or_error(conversation_id, user) + if error: + return error if not conversation.deleted: return JSONResponse( @@ -933,16 +911,9 @@ def _enqueue_speaker_reprocessing_chain( async def toggle_star(conversation_id: str, user: User): """Toggle the starred/favorite status of a conversation.""" try: - conversation = await Conversation.find_one( - Conversation.conversation_id == conversation_id - ) - if not conversation: - return JSONResponse( - status_code=404, content={"error": "Conversation not found"} - ) - - if not user.is_superuser and conversation.user_id != str(user.user_id): - return JSONResponse(status_code=403, content={"error": "Access forbidden"}) + conversation, error = await _get_conversation_or_error(conversation_id, user) + if error: + return error # Toggle conversation.starred = not conversation.starred @@ -993,17 +964,9 @@ async def toggle_star(conversation_id: str, user: User): async def reprocess_orphan(conversation_id: str, user: User): """Reprocess an orphan audio session - restore if deleted and enqueue full processing chain.""" try: - conversation = await Conversation.find_one( - Conversation.conversation_id == conversation_id - ) - if not conversation: - return JSONResponse( - status_code=404, content={"error": "Conversation not found"} - ) - - # Check ownership - if not user.is_superuser and conversation.user_id != str(user.user_id): - return JSONResponse(status_code=403, content={"error": "Access forbidden"}) + conversation, error = await _get_conversation_or_error(conversation_id, user) + if error: + return error # Verify audio chunks exist (check both deleted and non-deleted) total_chunks = await AudioChunkDocument.find( @@ -1068,23 +1031,11 @@ async def reprocess_orphan(conversation_id: str, user: User): async def reprocess_transcript(conversation_id: str, user: User): """Reprocess transcript for a conversation. Users can only reprocess their own conversations.""" try: - # Find the conversation using Beanie - conversation_model = await Conversation.find_one( - Conversation.conversation_id == conversation_id + conversation_model, error = await _get_conversation_or_error( + conversation_id, user ) - if not conversation_model: - return JSONResponse( - status_code=404, content={"error": "Conversation not found"} - ) - - # Check ownership for non-admin users - if not user.is_superuser and conversation_model.user_id != str(user.user_id): - return JSONResponse( - status_code=403, - content={ - "error": "Access forbidden. You can only reprocess your own conversations." - }, - ) + if error: + return error # Get audio_uuid from conversation # Validate audio chunks exist in MongoDB @@ -1137,23 +1088,11 @@ async def reprocess_memory( ): """Reprocess memory extraction for a specific transcript version. Users can only reprocess their own conversations.""" try: - # Find the conversation using Beanie - conversation_model = await Conversation.find_one( - Conversation.conversation_id == conversation_id + conversation_model, error = await _get_conversation_or_error( + conversation_id, user ) - if not conversation_model: - return JSONResponse( - status_code=404, content={"error": "Conversation not found"} - ) - - # Check ownership for non-admin users - if not user.is_superuser and conversation_model.user_id != str(user.user_id): - return JSONResponse( - status_code=403, - content={ - "error": "Access forbidden. You can only reprocess your own conversations." - }, - ) + if error: + return error # Resolve transcript version ID (handle "active" special case) error, transcript_version_id, transcript_version = _resolve_transcript_version( @@ -1205,22 +1144,11 @@ async def reprocess_speakers( """ try: # 1. Find conversation and validate ownership - conversation_model = await Conversation.find_one( - Conversation.conversation_id == conversation_id + conversation_model, error = await _get_conversation_or_error( + conversation_id, user ) - if not conversation_model: - return JSONResponse( - status_code=404, content={"error": "Conversation not found"} - ) - - # Check ownership for non-admin users - if not user.is_superuser and conversation_model.user_id != str(user.user_id): - return JSONResponse( - status_code=403, - content={ - "error": "Access forbidden. You can only reprocess your own conversations." - }, - ) + if error: + return error # 2-3. Resolve source transcript version ID and find version object error, source_version_id, source_version = _resolve_transcript_version( @@ -1349,23 +1277,11 @@ async def activate_transcript_version( ): """Activate a specific transcript version. Users can only modify their own conversations.""" try: - # Find the conversation using Beanie - conversation_model = await Conversation.find_one( - Conversation.conversation_id == conversation_id + conversation_model, error = await _get_conversation_or_error( + conversation_id, user ) - if not conversation_model: - return JSONResponse( - status_code=404, content={"error": "Conversation not found"} - ) - - # Check ownership for non-admin users - if not user.is_superuser and conversation_model.user_id != str(user.user_id): - return JSONResponse( - status_code=403, - content={ - "error": "Access forbidden. You can only modify your own conversations." - }, - ) + if error: + return error # Activate the transcript version using Beanie model method success = conversation_model.set_active_transcript_version(version_id) @@ -1401,23 +1317,11 @@ async def activate_transcript_version( async def activate_memory_version(conversation_id: str, version_id: str, user: User): """Activate a specific memory version. Users can only modify their own conversations.""" try: - # Find the conversation using Beanie - conversation_model = await Conversation.find_one( - Conversation.conversation_id == conversation_id + conversation_model, error = await _get_conversation_or_error( + conversation_id, user ) - if not conversation_model: - return JSONResponse( - status_code=404, content={"error": "Conversation not found"} - ) - - # Check ownership for non-admin users - if not user.is_superuser and conversation_model.user_id != str(user.user_id): - return JSONResponse( - status_code=403, - content={ - "error": "Access forbidden. You can only modify your own conversations." - }, - ) + if error: + return error # Activate the memory version using Beanie model method success = conversation_model.set_active_memory_version(version_id) @@ -1449,23 +1353,11 @@ async def activate_memory_version(conversation_id: str, version_id: str, user: U async def get_conversation_version_history(conversation_id: str, user: User): """Get version history for a conversation. Users can only access their own conversations.""" try: - # Find the conversation using Beanie to check ownership - conversation_model = await Conversation.find_one( - Conversation.conversation_id == conversation_id + conversation_model, error = await _get_conversation_or_error( + conversation_id, user ) - if not conversation_model: - return JSONResponse( - status_code=404, content={"error": "Conversation not found"} - ) - - # Check ownership for non-admin users - if not user.is_superuser and conversation_model.user_id != str(user.user_id): - return JSONResponse( - status_code=403, - content={ - "error": "Access forbidden. You can only access your own conversations." - }, - ) + if error: + return error # Get version history from model # Convert datetime objects to ISO strings for JSON serialization diff --git a/backends/advanced/src/advanced_omi_backend/controllers/memory_controller.py b/backends/advanced/src/advanced_omi_backend/controllers/memory_controller.py index fe4fca88..40c1ac51 100644 --- a/backends/advanced/src/advanced_omi_backend/controllers/memory_controller.py +++ b/backends/advanced/src/advanced_omi_backend/controllers/memory_controller.py @@ -17,15 +17,19 @@ audio_logger = logging.getLogger("audio_processing") +def _resolve_target_user(user: User, user_id: Optional[str] = None) -> str: + """Return the effective user ID: admins may override with user_id param.""" + if user.is_superuser and user_id: + return user_id + return user.user_id + + async def get_memories(user: User, limit: int, user_id: Optional[str] = None): """Get memories. Users see only their own memories, admins can see all or filter by user.""" try: memory_service = get_memory_service() - # Determine which user's memories to fetch - target_user_id = user.user_id - if user.is_superuser and user_id: - target_user_id = user_id + target_user_id = _resolve_target_user(user, user_id) # Execute memory retrieval directly (now async) memories = await memory_service.get_all_memories(target_user_id, limit) @@ -40,7 +44,7 @@ async def get_memories(user: User, limit: int, user_id: Optional[str] = None): "memories": memories_dicts, "count": len(memories), "total_count": total_count, - "user_id": target_user_id + "user_id": target_user_id, } except Exception as e: @@ -50,15 +54,14 @@ async def get_memories(user: User, limit: int, user_id: Optional[str] = None): ) -async def get_memories_with_transcripts(user: User, limit: int, user_id: Optional[str] = None): +async def get_memories_with_transcripts( + user: User, limit: int, user_id: Optional[str] = None +): """Get memories with their source transcripts. Users see only their own memories, admins can see all or filter by user.""" try: memory_service = get_memory_service() - # Determine which user's memories to fetch - target_user_id = user.user_id - if user.is_superuser and user_id: - target_user_id = user_id + target_user_id = _resolve_target_user(user, user_id) # Execute memory retrieval directly (now async) memories_with_transcripts = await memory_service.get_memories_with_transcripts( @@ -72,25 +75,32 @@ async def get_memories_with_transcripts(user: User, limit: int, user_id: Optiona } except Exception as e: - audio_logger.error(f"Error fetching memories with transcripts: {e}", exc_info=True) + audio_logger.error( + f"Error fetching memories with transcripts: {e}", exc_info=True + ) return JSONResponse( status_code=500, content={"message": f"Error fetching memories with transcripts: {str(e)}"}, ) -async def search_memories(query: str, user: User, limit: int, score_threshold: float = 0.0, user_id: Optional[str] = None): +async def search_memories( + query: str, + user: User, + limit: int, + score_threshold: float = 0.0, + user_id: Optional[str] = None, +): """Search memories by text query. Users can only search their own memories, admins can search all or filter by user.""" try: memory_service = get_memory_service() - # Determine which user's memories to search - target_user_id = user.user_id - if user.is_superuser and user_id: - target_user_id = user_id + target_user_id = _resolve_target_user(user, user_id) # Execute search directly (now async) - search_results = await memory_service.search_memories(query, target_user_id, limit, score_threshold) + search_results = await memory_service.search_memories( + query, target_user_id, limit, score_threshold + ) # Convert MemoryEntry objects to dicts for JSON serialization results_dicts = [result.to_dict() for result in search_results] @@ -122,16 +132,26 @@ async def delete_memory(memory_id: str, user: User): # MemoryEntry is a dataclass, access id attribute directly memory_ids = [str(mem.id) for mem in user_memories] if memory_id not in memory_ids: - return JSONResponse(status_code=404, content={"message": "Memory not found"}) + return JSONResponse( + status_code=404, content={"message": "Memory not found"} + ) # Delete the memory - audio_logger.info(f"Deleting memory {memory_id} for user_id={user.user_id}, email={user.email}") - success = await memory_service.delete_memory(memory_id, user_id=user.user_id, user_email=user.email) + audio_logger.info( + f"Deleting memory {memory_id} for user_id={user.user_id}, email={user.email}" + ) + success = await memory_service.delete_memory( + memory_id, user_id=user.user_id, user_email=user.email + ) if success: - return JSONResponse(content={"message": f"Memory {memory_id} deleted successfully"}) + return JSONResponse( + content={"message": f"Memory {memory_id} deleted successfully"} + ) else: - return JSONResponse(status_code=404, content={"message": "Memory not found"}) + return JSONResponse( + status_code=404, content={"message": "Memory not found"} + ) except Exception as e: audio_logger.error(f"Error deleting memory: {e}", exc_info=True) @@ -146,7 +166,9 @@ async def add_memory(content: str, user: User, source_id: Optional[str] = None): memory_service = get_memory_service() # Use source_id or generate a unique one - memory_source_id = source_id or f"manual_{user.user_id}_{int(asyncio.get_event_loop().time())}" + memory_source_id = ( + source_id or f"manual_{user.user_id}_{int(asyncio.get_event_loop().time())}" + ) # Extract memories from content success, memory_ids = await memory_service.add_memory( @@ -156,7 +178,7 @@ async def add_memory(content: str, user: User, source_id: Optional[str] = None): user_id=user.user_id, user_email=user.email, allow_update=False, - db_helper=None + db_helper=None, ) if success: @@ -165,18 +187,19 @@ async def add_memory(content: str, user: User, source_id: Optional[str] = None): "memory_ids": memory_ids, "count": len(memory_ids), "source_id": memory_source_id, - "message": f"Successfully created {len(memory_ids)} memory/memories" + "message": f"Successfully created {len(memory_ids)} memory/memories", } else: return JSONResponse( status_code=500, - content={"success": False, "message": "Failed to create memories"} + content={"success": False, "message": "Failed to create memories"}, ) except Exception as e: audio_logger.error(f"Error adding memory: {e}", exc_info=True) return JSONResponse( - status_code=500, content={"success": False, "message": f"Error adding memory: {str(e)}"} + status_code=500, + content={"success": False, "message": f"Error adding memory: {str(e)}"}, ) @@ -225,7 +248,8 @@ async def get_all_memories_admin(user: User, limit: int): except Exception as e: audio_logger.error(f"Error fetching admin memories: {e}", exc_info=True) return JSONResponse( - status_code=500, content={"message": f"Error fetching admin memories: {str(e)}"} + status_code=500, + content={"message": f"Error fetching admin memories: {str(e)}"}, ) @@ -234,10 +258,7 @@ async def get_memory_by_id(memory_id: str, user: User, user_id: Optional[str] = try: memory_service = get_memory_service() - # Determine which user's memory to fetch - target_user_id = user.user_id - if user.is_superuser and user_id: - target_user_id = user_id + target_user_id = _resolve_target_user(user, user_id) # Get the specific memory memory = await memory_service.get_memory(memory_id, target_user_id) @@ -265,11 +286,15 @@ async def get_memory_by_id(memory_id: str, user: User, user_id: Optional[str] = ), } except Exception as e: - logger.warning(f"Failed to fetch source conversation {source_id}: {e}") + logger.warning( + f"Failed to fetch source conversation {source_id}: {e}" + ) return {"memory": memory_dict} else: - return JSONResponse(status_code=404, content={"message": "Memory not found"}) + return JSONResponse( + status_code=404, content={"message": "Memory not found"} + ) except Exception as e: audio_logger.error(f"Error fetching memory {memory_id}: {e}", exc_info=True) diff --git a/backends/advanced/src/advanced_omi_backend/models/annotation.py b/backends/advanced/src/advanced_omi_backend/models/annotation.py index 451d84d1..99974532 100644 --- a/backends/advanced/src/advanced_omi_backend/models/annotation.py +++ b/backends/advanced/src/advanced_omi_backend/models/annotation.py @@ -16,22 +16,26 @@ class AnnotationType(str, Enum): """Type of content being annotated.""" + MEMORY = "memory" TRANSCRIPT = "transcript" DIARIZATION = "diarization" # Speaker identification corrections ENTITY = "entity" # Knowledge graph entity corrections (name/details edits) TITLE = "title" # Conversation title corrections INSERT = "insert" # Insert new segment between existing segments + SPEECH_SUGGESTION_CORRECTION = "speech_suggestion_correction" # User-refined model suggestion (training signal triple) class AnnotationSource(str, Enum): """Origin of the annotation.""" + USER = "user" # User-created edit MODEL_SUGGESTION = "model_suggestion" # AI-generated suggestion class AnnotationStatus(str, Enum): """Lifecycle status of annotation.""" + PENDING = "pending" # Waiting for user review (suggestions) ACCEPTED = "accepted" # Applied to content REJECTED = "rejected" # User dismissed suggestion @@ -79,6 +83,11 @@ class Annotation(Document): entity_id: Optional[str] = None # Neo4j entity ID entity_field: Optional[str] = None # Which field was changed ("name" or "details") + # For SPEECH_SUGGESTION_CORRECTION annotations: + model_suggested_text: Optional[str] = ( + None # What AI originally suggested before user edited + ) + # For INSERT annotations: insert_after_index: Optional[int] = None # -1 = before first segment insert_text: Optional[str] = None # e.g., "[laughter]" or "wife laughed" @@ -86,17 +95,17 @@ class Annotation(Document): insert_speaker: Optional[str] = None # Speaker label for "speech" type inserts # Processed tracking (applies to ALL annotation types) - processed: bool = Field(default=False) # Whether annotation has been applied/sent to training + processed: bool = Field( + default=False + ) # Whether annotation has been applied/sent to training processed_at: Optional[datetime] = None # When annotation was processed - processed_by: Optional[str] = None # What processed it (manual, cron, apply, training, etc.) + processed_by: Optional[str] = ( + None # What processed it (manual, cron, apply, training, etc.) + ) # Timestamps (Python 3.12+ compatible) - created_at: datetime = Field( - default_factory=lambda: datetime.now(timezone.utc) - ) - updated_at: datetime = Field( - default_factory=lambda: datetime.now(timezone.utc) - ) + created_at: datetime = Field(default_factory=lambda: datetime.now(timezone.utc)) + updated_at: datetime = Field(default_factory=lambda: datetime.now(timezone.utc)) class Settings: name = "annotations" @@ -132,6 +141,10 @@ def is_title_annotation(self) -> bool: """Check if this is a title annotation.""" return self.annotation_type == AnnotationType.TITLE + def is_speech_suggestion_correction(self) -> bool: + """Check if this is a user-refined model suggestion.""" + return self.annotation_type == AnnotationType.SPEECH_SUGGESTION_CORRECTION + def is_pending_suggestion(self) -> bool: """Check if this is a pending AI suggestion.""" return ( @@ -145,6 +158,7 @@ def is_pending_suggestion(self) -> bool: class AnnotationCreateBase(BaseModel): """Base model for annotation creation.""" + original_text: str = "" # Optional for diarization corrected_text: str = "" # Optional for diarization status: AnnotationStatus = AnnotationStatus.ACCEPTED @@ -152,6 +166,7 @@ class AnnotationCreateBase(BaseModel): class MemoryAnnotationCreate(AnnotationCreateBase): """Create memory annotation request.""" + memory_id: str original_text: str # Required for memory annotations corrected_text: str # Required for memory annotations @@ -159,6 +174,7 @@ class MemoryAnnotationCreate(AnnotationCreateBase): class TranscriptAnnotationCreate(AnnotationCreateBase): """Create transcript annotation request.""" + conversation_id: str segment_index: int original_text: str # Required for transcript annotations @@ -167,6 +183,7 @@ class TranscriptAnnotationCreate(AnnotationCreateBase): class DiarizationAnnotationCreate(BaseModel): """Create diarization annotation request.""" + conversation_id: str segment_index: int original_speaker: str @@ -181,6 +198,7 @@ class EntityAnnotationCreate(BaseModel): Dual purpose: feeds both the jargon pipeline (entity name corrections = domain vocabulary the ASR should know) and the entity extraction pipeline (corrections improve future accuracy). """ + entity_id: str entity_field: str # "name" or "details" original_text: str @@ -189,6 +207,7 @@ class EntityAnnotationCreate(BaseModel): class TitleAnnotationCreate(AnnotationCreateBase): """Create title annotation request.""" + conversation_id: str original_text: str corrected_text: str @@ -196,6 +215,7 @@ class TitleAnnotationCreate(AnnotationCreateBase): class InsertAnnotationCreate(BaseModel): """Create insert annotation request (new segment between existing segments).""" + conversation_id: str insert_after_index: int # -1 = before first segment insert_text: str @@ -205,8 +225,10 @@ class InsertAnnotationCreate(BaseModel): class AnnotationUpdate(BaseModel): """Update an existing unprocessed annotation.""" + corrected_text: Optional[str] = None corrected_speaker: Optional[str] = None + model_suggested_text: Optional[str] = None insert_text: Optional[str] = None insert_segment_type: Optional[str] = None insert_speaker: Optional[str] = None @@ -214,6 +236,7 @@ class AnnotationUpdate(BaseModel): class AnnotationResponse(BaseModel): """Annotation response for API.""" + id: str annotation_type: AnnotationType user_id: str @@ -227,6 +250,7 @@ class AnnotationResponse(BaseModel): segment_start_time: Optional[float] = None entity_id: Optional[str] = None entity_field: Optional[str] = None + model_suggested_text: Optional[str] = None insert_after_index: Optional[int] = None insert_text: Optional[str] = None insert_segment_type: Optional[str] = None diff --git a/backends/advanced/src/advanced_omi_backend/prompt_defaults.py b/backends/advanced/src/advanced_omi_backend/prompt_defaults.py index 94b03d31..7e847b41 100644 --- a/backends/advanced/src/advanced_omi_backend/prompt_defaults.py +++ b/backends/advanced/src/advanced_omi_backend/prompt_defaults.py @@ -607,6 +607,42 @@ def register_all_defaults(registry: PromptRegistry) -> None: is_dynamic=True, ) + # ------------------------------------------------------------------ + # annotation.transcript_error_detection + # ------------------------------------------------------------------ + registry.register_default( + "annotation.transcript_error_detection", + template="""\ +You are a transcript quality reviewer. Analyze the following transcript segments \ +from a conversation and identify potential transcription errors. + +Look for: +- Misheard words (homophones, phonetically similar substitutions) +- Nonsensical phrases that are likely ASR mistakes +- Obvious hallucinations or repeated/garbled text +- Missing or extra words that break sentence meaning + +Conversation title: {{title}} + +Segments (index: speaker - text): +{{segments_text}} + +Return a JSON array of issues found. Each issue should have: +- "segment_index": the index number of the problematic segment +- "original_text": the exact text from that segment +- "corrected_text": your suggested correction +- "reason": brief explanation (e.g. "misheard word", "garbled text", "hallucination") + +If no issues are found, return an empty array: [] + +Return ONLY the JSON array, no other text.""", + name="Transcript Error Detection", + description="Analyzes transcript segments for ASR errors, hallucinations, and misheard words.", + category="annotation", + variables=["title", "segments_text"], + is_dynamic=True, + ) + # ------------------------------------------------------------------ # prompt_optimization.title_optimizer # ------------------------------------------------------------------ diff --git a/backends/advanced/src/advanced_omi_backend/routers/modules/annotation_routes.py b/backends/advanced/src/advanced_omi_backend/routers/modules/annotation_routes.py index e04e6c76..43ffa212 100644 --- a/backends/advanced/src/advanced_omi_backend/routers/modules/annotation_routes.py +++ b/backends/advanced/src/advanced_omi_backend/routers/modules/annotation_routes.py @@ -9,13 +9,14 @@ from datetime import datetime, timezone from typing import List -from fastapi import APIRouter, Depends, HTTPException +from fastapi import APIRouter, Depends, HTTPException, Query from fastapi.responses import JSONResponse from advanced_omi_backend.auth import current_active_user from advanced_omi_backend.models.annotation import ( Annotation, AnnotationResponse, + AnnotationSource, AnnotationStatus, AnnotationType, AnnotationUpdate, @@ -36,6 +37,103 @@ router = APIRouter(prefix="/annotations", tags=["annotations"]) +@router.get("/suggestions") +async def get_pending_suggestions( + current_user: User = Depends(current_active_user), + limit: int = Query(20, ge=1, le=100), +): + """ + Get pending AI-generated suggestions for the current user. + + Returns MODEL_SUGGESTION annotations with PENDING status, + enriched with conversation context (title, transcript snippet, + audio path) for the swipe review UI. + """ + try: + annotations = ( + await Annotation.find( + Annotation.user_id == current_user.user_id, + Annotation.source == AnnotationSource.MODEL_SUGGESTION, + Annotation.status == AnnotationStatus.PENDING, + ) + .sort("-created_at") + .limit(limit) + .to_list() + ) + + if not annotations: + return [] + + # Batch-fetch conversations for context + conversation_ids = list( + {a.conversation_id for a in annotations if a.conversation_id} + ) + conversations = await Conversation.find( + {"conversation_id": {"$in": conversation_ids}}, + ).to_list() + conv_map = {c.conversation_id: c for c in conversations} + + results = [] + for a in annotations: + conv = conv_map.get(a.conversation_id) + + segment_start = None + segment_end = None + if conv and a.segment_index is not None: + transcript = conv.active_transcript + if ( + transcript + and transcript.segments + and a.segment_index < len(transcript.segments) + ): + seg = transcript.segments[a.segment_index] + segment_start = seg.start + segment_end = seg.end + + results.append( + { + "id": a.id, + "annotation_type": a.annotation_type, + "conversation_id": a.conversation_id, + "segment_index": a.segment_index, + "original_text": a.original_text, + "corrected_text": a.corrected_text, + "created_at": a.created_at.isoformat(), + "conversation_title": conv.title if conv else None, + "transcript_snippet": _get_segment_context(conv, a.segment_index), + "segment_start": segment_start, + "segment_end": segment_end, + } + ) + + return results + + except Exception as e: + logger.error(f"Error fetching suggestions: {e}", exc_info=True) + raise HTTPException( + status_code=500, detail=f"Failed to fetch suggestions: {str(e)}" + ) + + +def _get_segment_context( + conversation, segment_index: int | None, context_size: int = 1 +) -> str | None: + """Get a snippet of transcript around the flagged segment for context.""" + if not conversation or segment_index is None: + return None + transcript = conversation.active_transcript + if not transcript or not transcript.segments: + return None + start = max(0, segment_index - context_size) + end = min(len(transcript.segments), segment_index + context_size + 1) + lines = [] + for i in range(start, end): + seg = transcript.segments[i] + prefix = ">>> " if i == segment_index else " " + lines.append(f"{prefix}{seg.speaker}: {seg.text}") + return "\n".join(lines) + + @router.post("/memory", response_model=AnnotationResponse) async def create_memory_annotation( annotation_data: MemoryAnnotationCreate, @@ -85,11 +183,15 @@ async def create_memory_annotation( content=annotation_data.corrected_text, user_id=current_user.user_id, ) - logger.info(f"Updated memory {annotation_data.memory_id} with corrected text") + logger.info( + f"Updated memory {annotation_data.memory_id} with corrected text" + ) except Exception as e: logger.error(f"Error updating memory: {e}") # Annotation is saved, but memory update failed - log but don't fail the request - logger.warning(f"Memory annotation {annotation.id} saved but memory update failed") + logger.warning( + f"Memory annotation {annotation.id} saved but memory update failed" + ) return AnnotationResponse.model_validate(annotation) @@ -237,7 +339,22 @@ async def update_annotation_status( annotation.updated_at = datetime.now(timezone.utc) # If accepting a pending suggestion, apply the correction - if status == AnnotationStatus.ACCEPTED and old_status == AnnotationStatus.PENDING: + if ( + status == AnnotationStatus.ACCEPTED + and old_status == AnnotationStatus.PENDING + ): + # Promote to SPEECH_SUGGESTION_CORRECTION if user edited the AI suggestion + if ( + annotation.source == AnnotationSource.MODEL_SUGGESTION + and annotation.model_suggested_text is not None + and annotation.is_transcript_annotation() + ): + annotation.annotation_type = AnnotationType.SPEECH_SUGGESTION_CORRECTION + logger.info( + f"Promoted annotation {annotation_id} to SPEECH_SUGGESTION_CORRECTION " + f"(AI suggested: {annotation.model_suggested_text!r}, user decided: {annotation.corrected_text!r})" + ) + if annotation.is_memory_annotation(): # Update memory try: @@ -251,8 +368,11 @@ async def update_annotation_status( except Exception as e: logger.error(f"Error applying memory suggestion: {e}") # Don't fail the status update if memory update fails - elif annotation.is_transcript_annotation(): - # Update transcript segment + elif ( + annotation.is_transcript_annotation() + or annotation.is_speech_suggestion_correction() + ): + # Update transcript segment (same logic for both TRANSCRIPT and SPEECH_SUGGESTION_CORRECTION) try: conversation = await Conversation.find_one( Conversation.conversation_id == annotation.conversation_id, @@ -260,7 +380,9 @@ async def update_annotation_status( ) if conversation: transcript = conversation.active_transcript - if transcript and annotation.segment_index < len(transcript.segments): + if transcript and annotation.segment_index < len( + transcript.segments + ): transcript.segments[annotation.segment_index].text = ( annotation.corrected_text ) @@ -286,7 +408,9 @@ async def update_annotation_status( user_id=annotation.user_id, **update_kwargs, ) - logger.info(f"Applied entity suggestion to entity {annotation.entity_id}") + logger.info( + f"Applied entity suggestion to entity {annotation.entity_id}" + ) except Exception as e: logger.error(f"Error applying entity suggestion: {e}") # Don't fail the status update if entity update fails @@ -310,7 +434,11 @@ async def update_annotation_status( await annotation.save() logger.info(f"Updated annotation {annotation_id} status to {status}") - return {"status": "updated", "annotation_id": annotation_id, "new_status": status} + return { + "status": "updated", + "annotation_id": annotation_id, + "new_status": status, + } except HTTPException: raise @@ -345,7 +473,9 @@ async def delete_annotation( raise HTTPException(status_code=404, detail="Annotation not found") if annotation.processed: - raise HTTPException(status_code=400, detail="Cannot delete a processed annotation") + raise HTTPException( + status_code=400, detail="Cannot delete a processed annotation" + ) await annotation.delete() logger.info(f"Deleted annotation {annotation_id}") @@ -384,10 +514,22 @@ async def update_annotation( raise HTTPException(status_code=404, detail="Annotation not found") if annotation.processed: - raise HTTPException(status_code=400, detail="Cannot update a processed annotation") + raise HTTPException( + status_code=400, detail="Cannot update a processed annotation" + ) if update_data.corrected_text is not None: + # Auto-capture AI's original suggestion before user overwrites it + if ( + annotation.source == AnnotationSource.MODEL_SUGGESTION + and annotation.model_suggested_text is None + and annotation.corrected_text + and update_data.corrected_text != annotation.corrected_text + ): + annotation.model_suggested_text = annotation.corrected_text annotation.corrected_text = update_data.corrected_text + if update_data.model_suggested_text is not None: + annotation.model_suggested_text = update_data.model_suggested_text if update_data.corrected_speaker is not None: annotation.corrected_speaker = update_data.corrected_speaker if update_data.insert_text is not None: @@ -441,7 +583,10 @@ async def create_insert_annotation( raise HTTPException(status_code=400, detail="No active transcript found") segment_count = len(active_transcript.segments) - if annotation_data.insert_after_index < -1 or annotation_data.insert_after_index >= segment_count: + if ( + annotation_data.insert_after_index < -1 + or annotation_data.insert_after_index >= segment_count + ): raise HTTPException( status_code=400, detail=f"insert_after_index must be between -1 and {segment_count - 1}", @@ -572,7 +717,9 @@ async def create_entity_annotation( user_id=current_user.user_id, **update_kwargs, ) - logger.info(f"Applied entity correction to Neo4j for entity {annotation_data.entity_id}") + logger.info( + f"Applied entity correction to Neo4j for entity {annotation_data.entity_id}" + ) except Exception as e: logger.error(f"Error applying entity correction to Neo4j: {e}") # Annotation is saved but Neo4j update failed ā€” log but don't fail the request @@ -657,7 +804,9 @@ async def create_title_annotation( try: conversation.title = annotation_data.corrected_text await conversation.save() - logger.info(f"Updated title for conversation {annotation_data.conversation_id}") + logger.info( + f"Updated title for conversation {annotation_data.conversation_id}" + ) except Exception as e: logger.error(f"Error updating conversation title: {e}") # Annotation is saved but title update failed ā€” log but don't fail the request @@ -697,7 +846,6 @@ async def get_title_annotations( ) - # === Diarization Annotation Routes === @@ -817,7 +965,10 @@ async def apply_diarization_annotations( if not annotations: return JSONResponse( - content={"message": "No pending annotations to apply", "applied_count": 0} + content={ + "message": "No pending annotations to apply", + "applied_count": 0, + } ) # Get active transcript version @@ -839,7 +990,9 @@ async def apply_diarization_annotations( key=lambda a: a.updated_at, reverse=True, ) - annotation_for_segment = annotations_for_segment[0] if annotations_for_segment else None + annotation_for_segment = ( + annotations_for_segment[0] if annotations_for_segment else None + ) if annotation_for_segment: # Apply correction @@ -951,7 +1104,10 @@ async def apply_all_annotations( a for a in annotations if a.annotation_type == AnnotationType.DIARIZATION ] transcript_annotations = [ - a for a in annotations if a.annotation_type == AnnotationType.TRANSCRIPT + a + for a in annotations + if a.annotation_type + in (AnnotationType.TRANSCRIPT, AnnotationType.SPEECH_SUGGESTION_CORRECTION) ] insert_annotations = [ a for a in annotations if a.annotation_type == AnnotationType.INSERT diff --git a/backends/advanced/src/advanced_omi_backend/routers/modules/finetuning_routes.py b/backends/advanced/src/advanced_omi_backend/routers/modules/finetuning_routes.py index 7abb8bbd..2e338232 100644 --- a/backends/advanced/src/advanced_omi_backend/routers/modules/finetuning_routes.py +++ b/backends/advanced/src/advanced_omi_backend/routers/modules/finetuning_routes.py @@ -25,7 +25,9 @@ @router.post("/process-annotations") async def process_annotations_for_training( current_user: User = Depends(current_active_user), - annotation_type: Optional[str] = Query("diarization", description="Type of annotations to process"), + annotation_type: Optional[str] = Query( + "diarization", description="Type of annotations to process" + ), ): """ Send processed annotations to speaker recognition service for training. @@ -44,8 +46,7 @@ async def process_annotations_for_training( # Only admins can trigger training for now (can expand to per-user later) if not current_user.is_superuser: raise HTTPException( - status_code=403, - detail="Only administrators can trigger model training" + status_code=403, detail="Only administrators can trigger model training" ) # Find annotations ready for training @@ -57,15 +58,18 @@ async def process_annotations_for_training( # Filter out already trained annotations (processed_by contains "training") ready_for_training = [ - a for a in annotations + a + for a in annotations if not a.processed_by or "training" not in a.processed_by ] if not ready_for_training: - return JSONResponse(content={ - "message": "No annotations ready for training", - "processed_count": 0 - }) + return JSONResponse( + content={ + "message": "No annotations ready for training", + "processed_count": 0, + } + ) # Import required modules from advanced_omi_backend.models.conversation import Conversation @@ -78,13 +82,16 @@ async def process_annotations_for_training( # Initialize speaker client speaker_client = SpeakerRecognitionClient() - + if not speaker_client.enabled: - return JSONResponse(content={ - "message": "Speaker recognition service is not enabled", - "processed_count": 0, - "status": "error" - }, status_code=503) + return JSONResponse( + content={ + "message": "Speaker recognition service is not enabled", + "processed_count": 0, + "status": "error", + }, + status_code=503, + ) # Track processing statistics enrolled_count = 0 @@ -101,27 +108,33 @@ async def process_annotations_for_training( if not conversation or not conversation.active_transcript: failed_count += 1 - errors.append(f"Conversation {annotation.conversation_id[:8]} not found") + errors.append( + f"Conversation {annotation.conversation_id[:8]} not found" + ) continue # Validate segment index - if annotation.segment_index >= len(conversation.active_transcript.segments): + if annotation.segment_index >= len( + conversation.active_transcript.segments + ): failed_count += 1 errors.append(f"Invalid segment index {annotation.segment_index}") continue - segment = conversation.active_transcript.segments[annotation.segment_index] + segment = conversation.active_transcript.segments[ + annotation.segment_index + ] # 2. Extract audio segment from MongoDB logger.info( f"Extracting audio for conversation {annotation.conversation_id[:8]}... " f"segment {annotation.segment_index} ({segment.start:.2f}s - {segment.end:.2f}s)" ) - + wav_bytes = await reconstruct_audio_segment( conversation_id=annotation.conversation_id, start_time=segment.start, - end_time=segment.end + end_time=segment.end, ) if not wav_bytes: @@ -135,42 +148,49 @@ async def process_annotations_for_training( # 3. Check if speaker exists existing_speaker = await speaker_client.get_speaker_by_name( speaker_name=annotation.corrected_speaker, - user_id=1 # TODO: Map Chronicle user_id to speaker service user_id + user_id=1, # TODO: Map Chronicle user_id to speaker service user_id ) if existing_speaker: # APPEND to existing speaker - logger.info(f"Appending to existing speaker: {annotation.corrected_speaker}") + logger.info( + f"Appending to existing speaker: {annotation.corrected_speaker}" + ) result = await speaker_client.append_to_speaker( - speaker_id=existing_speaker["id"], - audio_data=wav_bytes + speaker_id=existing_speaker["id"], audio_data=wav_bytes ) - + if "error" in result: logger.error(f"Failed to append to speaker: {result}") failed_count += 1 errors.append(f"Append failed: {result.get('error')}") continue - + appended_count += 1 - logger.info(f"āœ… Successfully appended to speaker '{annotation.corrected_speaker}'") + logger.info( + f"āœ… Successfully appended to speaker '{annotation.corrected_speaker}'" + ) else: # ENROLL new speaker - logger.info(f"Enrolling new speaker: {annotation.corrected_speaker}") + logger.info( + f"Enrolling new speaker: {annotation.corrected_speaker}" + ) result = await speaker_client.enroll_new_speaker( speaker_name=annotation.corrected_speaker, audio_data=wav_bytes, - user_id=1 # TODO: Map Chronicle user_id to speaker service user_id + user_id=1, # TODO: Map Chronicle user_id to speaker service user_id ) - + if "error" in result: logger.error(f"Failed to enroll speaker: {result}") failed_count += 1 errors.append(f"Enroll failed: {result.get('error')}") continue - + enrolled_count += 1 - logger.info(f"āœ… Successfully enrolled new speaker '{annotation.corrected_speaker}'") + logger.info( + f"āœ… Successfully enrolled new speaker '{annotation.corrected_speaker}'" + ) # 4. Mark annotation as trained if annotation.processed_by: @@ -181,7 +201,9 @@ async def process_annotations_for_training( await annotation.save() except Exception as e: - logger.error(f"Error processing annotation {annotation.id}: {e}", exc_info=True) + logger.error( + f"Error processing annotation {annotation.id}: {e}", exc_info=True + ) failed_count += 1 errors.append(f"Exception: {str(e)[:50]}") continue @@ -192,15 +214,17 @@ async def process_annotations_for_training( f"({enrolled_count} new, {appended_count} appended, {failed_count} failed)" ) - return JSONResponse(content={ - "message": "Training complete", - "enrolled_new_speakers": enrolled_count, - "appended_to_existing": appended_count, - "total_processed": total_processed, - "failed_count": failed_count, - "errors": errors[:10] if errors else [], - "status": "success" if total_processed > 0 else "partial_failure" - }) + return JSONResponse( + content={ + "message": "Training complete", + "enrolled_new_speakers": enrolled_count, + "appended_to_existing": appended_count, + "total_processed": total_processed, + "failed_count": failed_count, + "errors": errors[:10] if errors else [], + "status": "success" if total_processed > 0 else "partial_failure", + } + ) except HTTPException: raise @@ -226,7 +250,9 @@ async def export_asr_dataset( Export job results with counts of conversations exported and annotations consumed. """ if not current_user.is_superuser: - raise HTTPException(status_code=403, detail="Only administrators can trigger ASR dataset export") + raise HTTPException( + status_code=403, detail="Only administrators can trigger ASR dataset export" + ) try: from advanced_omi_backend.workers.finetuning_jobs import run_asr_finetuning_job @@ -235,7 +261,9 @@ async def export_asr_dataset( return JSONResponse(content=result) except Exception as e: logger.error(f"ASR dataset export failed: {e}", exc_info=True) - raise HTTPException(status_code=500, detail=f"ASR dataset export failed: {str(e)}") + raise HTTPException( + status_code=500, detail=f"ASR dataset export failed: {str(e)}" + ) @router.get("/status") @@ -269,7 +297,11 @@ async def get_finetuning_status( ).to_list() # Batch-check which conversation_ids still exist - conv_annotation_types = {AnnotationType.DIARIZATION, AnnotationType.TRANSCRIPT} + conv_annotation_types = { + AnnotationType.DIARIZATION, + AnnotationType.TRANSCRIPT, + AnnotationType.SPEECH_SUGGESTION_CORRECTION, + } all_conv_ids: set[str] = set() for ann_type in conv_annotation_types: for a in all_annotations_by_type.get(ann_type, []): @@ -291,8 +323,12 @@ async def get_finetuning_status( # Identify orphaned annotations for conversation-based types if ann_type in conv_annotation_types: - orphaned = [a for a in annotations if a.conversation_id in orphaned_conv_ids] - non_orphaned = [a for a in annotations if a.conversation_id not in orphaned_conv_ids] + orphaned = [ + a for a in annotations if a.conversation_id in orphaned_conv_ids + ] + non_orphaned = [ + a for a in annotations if a.conversation_id not in orphaned_conv_ids + ] else: # Memory/entity orphan detection is placeholder for now orphaned = [] @@ -300,9 +336,12 @@ async def get_finetuning_status( pending = [a for a in non_orphaned if not a.processed] processed = [a for a in non_orphaned if a.processed] - trained = [a for a in processed if a.processed_by and "training" in a.processed_by] + trained = [ + a for a in processed if a.processed_by and "training" in a.processed_by + ] applied_not_trained = [ - a for a in processed + a + for a in processed if not a.processed_by or "training" not in a.processed_by ] @@ -333,9 +372,17 @@ async def get_finetuning_status( if trained_diarization_list: latest_trained = max( trained_diarization_list, - key=lambda a: a.updated_at if a.updated_at else datetime.min.replace(tzinfo=timezone.utc) + key=lambda a: ( + a.updated_at + if a.updated_at + else datetime.min.replace(tzinfo=timezone.utc) + ), + ) + last_training_run = ( + latest_trained.updated_at.isoformat() + if latest_trained.updated_at + else None ) - last_training_run = latest_trained.updated_at.isoformat() if latest_trained.updated_at else None # Get cron job status from scheduler try: @@ -344,7 +391,9 @@ async def get_finetuning_status( scheduler = get_scheduler() all_jobs = await scheduler.get_all_jobs_status() # Find speaker finetuning job for backward compat - speaker_job = next((j for j in all_jobs if j["job_id"] == "speaker_finetuning"), None) + speaker_job = next( + (j for j in all_jobs if j["job_id"] == "speaker_finetuning"), None + ) cron_status = { "enabled": speaker_job["enabled"] if speaker_job else False, "schedule": speaker_job["schedule"] if speaker_job else "0 2 * * *", @@ -359,15 +408,17 @@ async def get_finetuning_status( "next_run": None, } - return JSONResponse(content={ - "pending_annotation_count": pending_count, - "applied_annotation_count": applied_count, - "trained_annotation_count": trained_count, - "last_training_run": last_training_run, - "cron_status": cron_status, - "annotation_counts": annotation_counts, - "orphaned_annotation_count": total_orphaned, - }) + return JSONResponse( + content={ + "pending_annotation_count": pending_count, + "applied_annotation_count": applied_count, + "trained_annotation_count": trained_count, + "last_training_run": last_training_run, + "cron_status": cron_status, + "annotation_counts": annotation_counts, + "orphaned_annotation_count": total_orphaned, + } + ) except Exception as e: logger.error(f"Error fetching fine-tuning status: {e}", exc_info=True) @@ -385,7 +436,9 @@ async def get_finetuning_status( @router.delete("/orphaned-annotations") async def delete_orphaned_annotations( current_user: User = Depends(current_active_user), - annotation_type: Optional[str] = Query(None, description="Filter by annotation type (e.g. 'diarization')"), + annotation_type: Optional[str] = Query( + None, description="Filter by annotation type (e.g. 'diarization')" + ), ): """ Find and delete orphaned annotations whose referenced conversation no longer exists. @@ -404,9 +457,17 @@ async def delete_orphaned_annotations( try: requested_type = AnnotationType(annotation_type) except ValueError: - raise HTTPException(status_code=400, detail=f"Unknown annotation type: {annotation_type}") + raise HTTPException( + status_code=400, detail=f"Unknown annotation type: {annotation_type}" + ) if requested_type not in conv_annotation_types: - return JSONResponse(content={"deleted_count": 0, "by_type": {}, "message": "Orphan detection not supported for this type"}) + return JSONResponse( + content={ + "deleted_count": 0, + "by_type": {}, + "message": "Orphan detection not supported for this type", + } + ) types_to_check = {requested_type} else: types_to_check = conv_annotation_types @@ -448,10 +509,12 @@ async def delete_orphaned_annotations( total_deleted += len(orphaned) logger.info(f"Deleted {total_deleted} orphaned annotations: {deleted_by_type}") - return JSONResponse(content={ - "deleted_count": total_deleted, - "by_type": deleted_by_type, - }) + return JSONResponse( + content={ + "deleted_count": total_deleted, + "by_type": deleted_by_type, + } + ) @router.post("/orphaned-annotations/reattach") diff --git a/backends/advanced/src/advanced_omi_backend/services/plugin_service.py b/backends/advanced/src/advanced_omi_backend/services/plugin_service.py index 9a8cc205..a2061c5d 100644 --- a/backends/advanced/src/advanced_omi_backend/services/plugin_service.py +++ b/backends/advanced/src/advanced_omi_backend/services/plugin_service.py @@ -18,6 +18,7 @@ from advanced_omi_backend.config_loader import get_plugins_yml_path from advanced_omi_backend.plugins import BasePlugin, PluginRouter +from advanced_omi_backend.plugins.events import PluginEvent from advanced_omi_backend.plugins.services import PluginServices logger = logging.getLogger(__name__) @@ -176,7 +177,9 @@ def replacer(match): return value -def load_plugin_config(plugin_id: str, orchestration_config: Dict[str, Any]) -> Dict[str, Any]: +def load_plugin_config( + plugin_id: str, orchestration_config: Dict[str, Any] +) -> Dict[str, Any]: """ Load complete plugin configuration from multiple sources. @@ -215,9 +218,13 @@ def load_plugin_config(plugin_id: str, orchestration_config: Dict[str, Any]) -> with open(plugin_config_path, "r") as f: plugin_config = yaml.safe_load(f) or {} config.update(plugin_config) - logger.debug(f"Loaded {len(plugin_config)} config keys for '{plugin_id}'") + logger.debug( + f"Loaded {len(plugin_config)} config keys for '{plugin_id}'" + ) else: - logger.debug(f"No config.yml found for plugin '{plugin_id}' at {plugin_config_path}") + logger.debug( + f"No config.yml found for plugin '{plugin_id}' at {plugin_config_path}" + ) except Exception as e: logger.warning(f"Failed to load config.yml for plugin '{plugin_id}': {e}") @@ -398,7 +405,9 @@ def load_schema_yml(plugin_id: str) -> Optional[Dict[str, Any]]: return None -def infer_schema_from_config(plugin_id: str, config_dict: Dict[str, Any]) -> Dict[str, Any]: +def infer_schema_from_config( + plugin_id: str, config_dict: Dict[str, Any] +) -> Dict[str, Any]: """Infer configuration schema from plugin config.yml. This function analyzes the config.yml file to generate a JSON schema @@ -480,8 +489,7 @@ def mask_secrets_in_config( if env_var and env_var in secret_env_vars: # Check if env var is set in per-plugin .env or os.environ is_set = bool( - (plugin_env and plugin_env.get(env_var)) - or os.environ.get(env_var) + (plugin_env and plugin_env.get(env_var)) or os.environ.get(env_var) ) masked_config[key] = "ā€¢ā€¢ā€¢ā€¢ā€¢ā€¢ā€¢ā€¢ā€¢ā€¢ā€¢ā€¢" if is_set else "" @@ -527,7 +535,9 @@ def get_plugin_metadata( # Mask secrets in current config current_config = load_plugin_config(plugin_id, orchestration_config) - masked_config = mask_secrets_in_config(current_config, config_schema, plugin_env=plugin_env) + masked_config = mask_secrets_in_config( + current_config, config_schema, plugin_env=plugin_env + ) # Mark which env vars are set (check per-plugin .env first, then os.environ) for env_var_name, env_var_schema in config_schema.get("env_vars", {}).items(): @@ -727,17 +737,24 @@ def _build_plugin_router() -> Optional[PluginRouter]: # Let plugin register its prompts with the prompt registry try: - from advanced_omi_backend.prompt_registry import get_prompt_registry + from advanced_omi_backend.prompt_registry import ( + get_prompt_registry, + ) + plugin.register_prompts(get_prompt_registry()) except Exception as e: - logger.debug(f"Plugin '{plugin_id}' prompt registration skipped: {e}") + logger.debug( + f"Plugin '{plugin_id}' prompt registration skipped: {e}" + ) # Note: async initialization happens in app_factory lifespan or reload_plugins router.register_plugin(plugin_id, plugin) logger.info(f"Plugin '{plugin_id}' registered successfully") except Exception as e: - logger.error(f"Failed to register plugin '{plugin_id}': {e}", exc_info=True) + logger.error( + f"Failed to register plugin '{plugin_id}': {e}", exc_info=True + ) logger.info( f"Plugin registration complete: {len(router.plugins)} plugin(s) registered" @@ -806,6 +823,62 @@ async def ensure_plugin_router() -> Optional[PluginRouter]: return plugin_router +async def dispatch_plugin_event( + event: PluginEvent, + user_id: str, + data: dict, + metadata: dict = None, + description: str = "", + require_router: bool = False, +) -> Optional[list]: + """Dispatch an event to the plugin system with standard logging. + + Handles the common pattern of: ensure router -> dispatch event -> log results. + + Args: + event: Plugin event to dispatch + user_id: User ID for the event + data: Event-specific data dict + metadata: Optional metadata dict + description: Log context (e.g., "conversation=abc123, memories=5") + require_router: If True and no router, raise RuntimeError instead of returning None + + Returns: + List of plugin results, or None if no router available + + Raises: + RuntimeError: If require_router=True and no plugin router is available + """ + plugin_router = await ensure_plugin_router() + + if not plugin_router: + if require_router: + raise RuntimeError( + f"Plugin router could not be initialized in worker process. " + f"{event.value} event will NOT be dispatched!" + ) + return None + + logger.info(f"šŸ”Œ DISPATCH: {event.value} event ({description})") + + plugin_results = await plugin_router.dispatch_event( + event=event, + user_id=user_id, + data=data, + metadata=metadata or {}, + ) + + result_count = len(plugin_results) if plugin_results else 0 + logger.info(f"šŸ”Œ RESULT: {event.value} dispatched to {result_count} plugins") + + if plugin_results: + for result in plugin_results: + if result.message: + logger.info(f" Plugin result: {result.message}") + + return plugin_results + + async def cleanup_plugin_router() -> None: """Clean up the plugin router and all registered plugins.""" global _plugin_router @@ -934,7 +1007,9 @@ def signal_worker_restart() -> None: try: timestamp = time.strftime("%Y-%m-%dT%H:%M:%S") client.set(WORKER_RESTART_KEY, timestamp) - logger.info(f"Worker restart signal sent via Redis key '{WORKER_RESTART_KEY}'") + logger.info( + f"Worker restart signal sent via Redis key '{WORKER_RESTART_KEY}'" + ) finally: client.close() except Exception as e: diff --git a/backends/advanced/src/advanced_omi_backend/utils/job_utils.py b/backends/advanced/src/advanced_omi_backend/utils/job_utils.py index c9028909..695906c4 100644 --- a/backends/advanced/src/advanced_omi_backend/utils/job_utils.py +++ b/backends/advanced/src/advanced_omi_backend/utils/job_utils.py @@ -10,7 +10,31 @@ logger = logging.getLogger(__name__) -async def check_job_alive(redis_client, current_job, session_id: Optional[str] = None) -> bool: +def update_job_meta(**kwargs) -> None: + """Update the current RQ job's metadata with the given key-value pairs. + + Handles the common boilerplate of: get_current_job() -> null check -> + meta init -> update -> save_meta. + + Args: + **kwargs: Key-value pairs to merge into job.meta + + Example: + update_job_meta(conversation_id="abc", processing_time=1.5) + """ + from rq import get_current_job + + current_job = get_current_job() + if current_job: + if not current_job.meta: + current_job.meta = {} + current_job.meta.update(kwargs) + current_job.save_meta() + + +async def check_job_alive( + redis_client, current_job, session_id: Optional[str] = None +) -> bool: """ Check if current RQ job still exists in Redis. @@ -44,12 +68,19 @@ async def check_job_alive(redis_client, current_job, session_id: Optional[str] = if session_id: session_key = f"audio:session:{session_id}" session_status = await redis_client.hget(session_key, "status") - if session_status and session_status.decode() in ["finalizing", "finished"]: + if session_status and session_status.decode() in [ + "finalizing", + "finished", + ]: # Session ended naturally - not a zombie, just natural cleanup - logger.debug(f"šŸ“‹ Job {current_job.id} ending naturally (session closed)") + logger.debug( + f"šŸ“‹ Job {current_job.id} ending naturally (session closed)" + ) return False # True zombie - job deleted while session still active - logger.error(f"šŸ§Ÿ Zombie job detected - job {current_job.id} deleted from Redis while session still active, exiting") + logger.error( + f"šŸ§Ÿ Zombie job detected - job {current_job.id} deleted from Redis while session still active, exiting" + ) return False return True diff --git a/backends/advanced/src/advanced_omi_backend/workers/annotation_jobs.py b/backends/advanced/src/advanced_omi_backend/workers/annotation_jobs.py index 3681ab5f..10d8f65e 100644 --- a/backends/advanced/src/advanced_omi_backend/workers/annotation_jobs.py +++ b/backends/advanced/src/advanced_omi_backend/workers/annotation_jobs.py @@ -4,14 +4,14 @@ These jobs run periodically via the cron scheduler to: 1. Surface potential errors in transcripts and memories for user review 2. Fine-tune error detection models using accepted/rejected annotations - -TODO: Implement actual LLM-based error detection and model training logic. """ +import json import logging from datetime import datetime, timedelta, timezone from typing import List +from advanced_omi_backend.llm_client import async_generate from advanced_omi_backend.models.annotation import ( Annotation, AnnotationSource, @@ -20,100 +20,182 @@ ) from advanced_omi_backend.models.conversation import Conversation from advanced_omi_backend.models.user import User +from advanced_omi_backend.prompt_registry import get_prompt_registry logger = logging.getLogger(__name__) +LOOKBACK_DAYS = 7 +MAX_SEGMENTS_PER_PROMPT = 30 +MAX_SUGGESTIONS_PER_RUN = 50 + +PROMPT_ID = "annotation.transcript_error_detection" + async def surface_error_suggestions(): """ - Generate AI suggestions for potential transcript/memory errors. - Runs daily, creates PENDING annotations for user review. + Generate AI suggestions for potential transcript errors. - This is a PLACEHOLDER implementation. To fully implement: - 1. Query recent transcripts and memories (last N days) - 2. Use LLM to analyze content for potential errors: - - Hallucinations (made-up facts) - - Misheard words (audio transcription errors) - - Grammar/spelling issues - - Inconsistencies with other memories - 3. For each potential error: - - Create PENDING annotation with MODEL_SUGGESTION source - - Store original_text and suggested corrected_text - 4. Users can review suggestions in UI (accept/reject) - 5. Accepted suggestions improve future model accuracy - - TODO: Implement LLM-based error detection logic. + Runs daily via cron. For each user, queries recent conversations + and uses the LLM to identify potential transcription errors. + Creates PENDING annotations with MODEL_SUGGESTION source for + user review in the swipe UI. """ - logger.info("šŸ“ Checking for annotation suggestions (placeholder)...") + logger.info("Checking for annotation suggestions...") + total_created = 0 try: - # Get all users users = await User.find_all().to_list() - logger.info(f" Found {len(users)} users to analyze") + logger.info(f"Found {len(users)} users to analyze") for user in users: - # TODO: Query recent conversations for this user (last 7 days) - # recent_conversations = await Conversation.find( - # Conversation.user_id == str(user.id), - # Conversation.created_at >= datetime.now(timezone.utc) - timedelta(days=7) - # ).to_list() - - # TODO: For each conversation, analyze transcripts - # for conversation in recent_conversations: - # active_transcript = conversation.get_active_transcript() - # if not active_transcript: - # continue - # - # # TODO: Use LLM to identify potential errors - # # suggestions = await llm_provider.analyze_transcript_for_errors( - # # segments=active_transcript.segments, - # # context=conversation.summary - # # ) - # - # # TODO: Create PENDING annotations for each suggestion - # # for suggestion in suggestions: - # # annotation = Annotation( - # # annotation_type=AnnotationType.TRANSCRIPT, - # # user_id=str(user.id), - # # conversation_id=conversation.conversation_id, - # # segment_index=suggestion.segment_index, - # # original_text=suggestion.original_text, - # # corrected_text=suggestion.suggested_text, - # # source=AnnotationSource.MODEL_SUGGESTION, - # # status=AnnotationStatus.PENDING - # # ) - # # await annotation.save() - - # TODO: Query recent memories for this user - # recent_memories = await memory_service.get_recent_memories( - # user_id=str(user.id), - # days=7 - # ) - - # TODO: Use LLM to identify potential errors in memories - # for memory in recent_memories: - # # TODO: Analyze memory content for hallucinations/errors - # # suggestions = await llm_provider.analyze_memory_for_errors( - # # content=memory.content, - # # metadata=memory.metadata - # # ) - # - # # TODO: Create PENDING annotations - # # ... - - # Placeholder logging - logger.debug(f" Analyzed user {user.id} (placeholder)") - - logger.info("āœ… Suggestion check complete (placeholder implementation)") - logger.info( - " ā„¹ļø TODO: Implement LLM-based error detection to create actual suggestions" - ) + user_id = str(user.id) + cutoff = datetime.now(timezone.utc) - timedelta(days=LOOKBACK_DAYS) + + recent_conversations = await Conversation.find( + Conversation.user_id == user_id, + Conversation.created_at >= cutoff, + Conversation.deleted != True, + ).to_list() + + if not recent_conversations: + logger.info( + f"User {user.email or user_id}: no recent conversations, skipping" + ) + continue + + logger.info( + f"User {user.email or user_id}: {len(recent_conversations)} conversations in last {LOOKBACK_DAYS} days" + ) + + # Get conversation IDs that already have pending model suggestions + existing = await Annotation.find( + Annotation.user_id == user_id, + Annotation.source == AnnotationSource.MODEL_SUGGESTION, + Annotation.status == AnnotationStatus.PENDING, + ).to_list() + skip_conversation_ids = { + a.conversation_id for a in existing if a.conversation_id + } + if skip_conversation_ids: + logger.info( + f" Skipping {len(skip_conversation_ids)} conversations with existing pending suggestions" + ) + + created_for_user = 0 + for conversation in recent_conversations: + if total_created >= MAX_SUGGESTIONS_PER_RUN: + logger.info( + f" Reached max suggestions per run ({MAX_SUGGESTIONS_PER_RUN}), stopping" + ) + break + if conversation.conversation_id in skip_conversation_ids: + continue + + active_transcript = conversation.active_transcript + if not active_transcript or not active_transcript.segments: + logger.debug( + f" Conversation '{conversation.title or conversation.conversation_id}': no transcript/segments, skipping" + ) + continue + + seg_count = len(active_transcript.segments) + logger.info( + f" Analyzing '{conversation.title or 'Untitled'}' " + f"({seg_count} segments, id={conversation.conversation_id[:8]}...)" + ) + + suggestions = await _analyze_transcript(conversation, active_transcript) + + if not suggestions: + logger.info(f" No issues found") + else: + logger.info(f" LLM found {len(suggestions)} potential issues") + + for suggestion in suggestions: + if total_created >= MAX_SUGGESTIONS_PER_RUN: + break + + seg_idx = suggestion.get("segment_index") + if seg_idx is None or seg_idx >= len(active_transcript.segments): + logger.debug(f" Skipping invalid segment_index={seg_idx}") + continue + + annotation = Annotation( + annotation_type=AnnotationType.TRANSCRIPT, + user_id=user_id, + conversation_id=conversation.conversation_id, + segment_index=seg_idx, + original_text=suggestion.get("original_text", ""), + corrected_text=suggestion.get("corrected_text", ""), + source=AnnotationSource.MODEL_SUGGESTION, + status=AnnotationStatus.PENDING, + ) + await annotation.save() + total_created += 1 + created_for_user += 1 + logger.info( + f" Created suggestion: segment {seg_idx} - " + f"'{suggestion.get('reason', 'unknown')}'" + ) + + logger.info( + f"User {user.email or user_id}: {created_for_user} suggestions created" + ) + + logger.info(f"Suggestion check complete: {total_created} annotations created") except Exception as e: - logger.error(f"āŒ Error in surface_error_suggestions: {e}", exc_info=True) + logger.error(f"Error in surface_error_suggestions: {e}", exc_info=True) raise +async def _analyze_transcript(conversation, transcript) -> list[dict]: + """Use LLM to analyze a transcript for potential errors.""" + segments = transcript.segments[:MAX_SEGMENTS_PER_PROMPT] + segments_text = "\n".join( + f"{i}: {seg.speaker} - {seg.text}" + for i, seg in enumerate(segments) + if seg.text.strip() + ) + + if not segments_text: + logger.debug(f" No non-empty segments to analyze") + return [] + + registry = get_prompt_registry() + prompt = await registry.get_prompt( + PROMPT_ID, + title=conversation.title or "Untitled", + segments_text=segments_text, + ) + + try: + logger.debug(f" Sending {len(segments)} segments to LLM for analysis...") + response = await async_generate(prompt) + logger.debug(f" LLM response length: {len(response)} chars") + # Parse JSON from response, handling markdown code blocks + text = response.strip() + if text.startswith("```"): + text = text.split("\n", 1)[1] if "\n" in text else text[3:] + text = text.rsplit("```", 1)[0] + suggestions = json.loads(text) + if not isinstance(suggestions, list): + logger.warning(f" LLM returned non-list response, ignoring") + return [] + return suggestions + except json.JSONDecodeError as e: + logger.warning( + f" Failed to parse LLM JSON for '{conversation.title or conversation.conversation_id}': {e}" + ) + logger.debug(f" Raw LLM response: {response[:500]}") + return [] + except Exception as e: + logger.warning( + f" LLM call failed for '{conversation.title or conversation.conversation_id}': {e}" + ) + return [] + + async def finetune_hallucination_model(): """ Fine-tune error detection model using accepted/rejected annotations. @@ -199,15 +281,11 @@ async def finetune_hallucination_model(): # Calculate acceptance rate if accepted_count + rejected_count > 0: - acceptance_rate = ( - accepted_count / (accepted_count + rejected_count) - ) * 100 + acceptance_rate = (accepted_count / (accepted_count + rejected_count)) * 100 logger.info(f" Suggestion acceptance rate: {acceptance_rate:.1f}%") logger.info("āœ… Training check complete (placeholder implementation)") - logger.info( - " ā„¹ļø TODO: Implement model fine-tuning using user feedback data" - ) + logger.info(" ā„¹ļø TODO: Implement model fine-tuning using user feedback data") except Exception as e: logger.error(f"āŒ Error in finetune_hallucination_model: {e}", exc_info=True) @@ -216,6 +294,7 @@ async def finetune_hallucination_model(): # Additional helper functions for future implementation + async def analyze_common_error_patterns() -> List[dict]: """ Analyze accepted annotations to identify common error patterns. diff --git a/backends/advanced/src/advanced_omi_backend/workers/conversation_jobs.py b/backends/advanced/src/advanced_omi_backend/workers/conversation_jobs.py index 2142ce07..5f7487e5 100644 --- a/backends/advanced/src/advanced_omi_backend/workers/conversation_jobs.py +++ b/backends/advanced/src/advanced_omi_backend/workers/conversation_jobs.py @@ -24,7 +24,7 @@ from advanced_omi_backend.observability.otel_setup import set_otel_session from advanced_omi_backend.plugins.events import PluginEvent from advanced_omi_backend.services.plugin_service import ( - ensure_plugin_router, + dispatch_plugin_event, get_plugin_router, ) from advanced_omi_backend.utils.conversation_utils import ( @@ -35,6 +35,7 @@ track_speech_activity, update_job_progress_metadata, ) +from advanced_omi_backend.utils.job_utils import update_job_meta logger = logging.getLogger(__name__) @@ -1161,27 +1162,16 @@ async def generate_title_summary_job( processing_time = time.time() - start_time # Update job metadata - from rq import get_current_job - - current_job = get_current_job() - if current_job: - if not current_job.meta: - current_job.meta = {} - current_job.meta.update( - { - "conversation_id": conversation_id, - "title": conversation.title, - "summary": conversation.summary, - "detailed_summary_length": ( - len(conversation.detailed_summary) - if conversation.detailed_summary - else 0 - ), - "segment_count": len(segments), - "processing_time": processing_time, - } - ) - current_job.save_meta() + update_job_meta( + conversation_id=conversation_id, + title=conversation.title, + summary=conversation.summary, + detailed_summary_length=( + len(conversation.detailed_summary) if conversation.detailed_summary else 0 + ), + segment_count=len(segments), + processing_time=processing_time, + ) logger.info( f"āœ… Title/summary generation completed for {conversation_id} in {processing_time:.2f}s" @@ -1264,64 +1254,25 @@ async def dispatch_conversation_complete_event_job( user_email = user.email if user else "" # Prepare plugin event data (same format as open_conversation_job) + actual_end_reason = end_reason or "file_upload" try: - plugin_router = await ensure_plugin_router() - - # CRITICAL CHECK: Fail loudly if no router - if not plugin_router: - error_msg = ( - f"āŒ Plugin router could not be initialized in worker process. " - f"conversation.complete event for {conversation_id[:12]} will NOT be dispatched!" - ) - logger.error(error_msg) - - return { - "success": False, - "skipped": True, - "reason": "No plugin router", - "conversation_id": conversation_id, - "error": error_msg, - } - - plugin_data = { - "conversation": { - "client_id": client_id, - "user_id": user_id, - }, - "transcript": conversation.transcript if conversation else "", - "duration": 0, # Duration not tracked for file uploads - "conversation_id": conversation_id, - } - - # Use provided end_reason or default to 'file_upload' for backward compatibility - actual_end_reason = end_reason or "file_upload" - - logger.info( - f"šŸ”Œ DISPATCH: conversation.complete event for {conversation_id[:12]} " - f"(end_reason={actual_end_reason}, user={user_id}, client={client_id})" - ) - - plugin_results = await plugin_router.dispatch_event( + plugin_results = await dispatch_plugin_event( event=PluginEvent.CONVERSATION_COMPLETE, user_id=user_id, - data=plugin_data, + data={ + "conversation": { + "client_id": client_id, + "user_id": user_id, + }, + "transcript": conversation.transcript if conversation else "", + "duration": 0, # Duration not tracked for file uploads + "conversation_id": conversation_id, + }, metadata={"end_reason": actual_end_reason}, + description=f"conversation={conversation_id[:12]}, end_reason={actual_end_reason}", + require_router=True, ) - logger.info( - f"šŸ”Œ RESULT: conversation.complete dispatched to {len(plugin_results) if plugin_results else 0} plugins" - ) - if plugin_results: - logger.info( - f"šŸ“Œ Triggered {len(plugin_results)} conversation-level plugins" - ) - for result in plugin_results: - logger.info( - f" Plugin result: success={result.success}, message={result.message}" - ) - if result.message: - logger.info(f" Plugin result: {result.message}") - processing_time = time.time() - start_time logger.info( f"āœ… Conversation complete event dispatched for {conversation_id} in {processing_time:.2f}s" @@ -1334,6 +1285,15 @@ async def dispatch_conversation_complete_event_job( "processing_time_seconds": processing_time, } + except RuntimeError as e: + logger.error(f"āŒ {e}") + return { + "success": False, + "skipped": True, + "reason": "No plugin router", + "conversation_id": conversation_id, + "error": str(e), + } except Exception as e: logger.warning(f"āš ļø Error dispatching conversation complete event: {e}") return { diff --git a/backends/advanced/src/advanced_omi_backend/workers/memory_jobs.py b/backends/advanced/src/advanced_omi_backend/workers/memory_jobs.py index 492dc650..1fcfe510 100644 --- a/backends/advanced/src/advanced_omi_backend/workers/memory_jobs.py +++ b/backends/advanced/src/advanced_omi_backend/workers/memory_jobs.py @@ -27,7 +27,7 @@ set_otel_session, ) from advanced_omi_backend.plugins.events import PluginEvent -from advanced_omi_backend.services.plugin_service import ensure_plugin_router +from advanced_omi_backend.services.plugin_service import dispatch_plugin_event logger = logging.getLogger(__name__) @@ -391,11 +391,12 @@ async def process_memory_job( logger.warning(f"āš ļø Knowledge graph extraction failed (non-fatal): {e}") # Trigger memory-level plugins (ALWAYS dispatch when success, even with 0 new memories) + memory_count = len(created_memory_ids) if created_memory_ids else 0 try: - plugin_router = await ensure_plugin_router() - - if plugin_router: - plugin_data = { + await dispatch_plugin_event( + event=PluginEvent.MEMORY_PROCESSED, + user_id=user_id, + data={ "memories": created_memory_ids or [], "conversation": { "conversation_id": conversation_id, @@ -403,39 +404,15 @@ async def process_memory_job( "user_id": user_id, "user_email": user_email, }, - "memory_count": ( - len(created_memory_ids) if created_memory_ids else 0 - ), + "memory_count": memory_count, "conversation_id": conversation_id, - } - - logger.info( - f"šŸ”Œ DISPATCH: memory.processed event " - f"(conversation={conversation_id[:12]}, memories={len(created_memory_ids) if created_memory_ids else 0})" - ) - - plugin_results = await plugin_router.dispatch_event( - event=PluginEvent.MEMORY_PROCESSED, - user_id=user_id, - data=plugin_data, - metadata={ - "processing_time": processing_time, - "memory_provider": memory_provider, - }, - ) - - logger.info( - f"šŸ”Œ RESULT: memory.processed dispatched to {len(plugin_results) if plugin_results else 0} plugins" - ) - - if plugin_results: - logger.info( - f"šŸ“Œ Triggered {len(plugin_results)} memory-level plugins" - ) - for result in plugin_results: - if result.message: - logger.info(f" Plugin result: {result.message}") - + }, + metadata={ + "processing_time": processing_time, + "memory_provider": memory_provider, + }, + description=f"conversation={conversation_id[:12]}, memories={memory_count}", + ) except Exception as e: logger.warning(f"āš ļø Error triggering memory-level plugins: {e}") diff --git a/backends/advanced/src/advanced_omi_backend/workers/speaker_jobs.py b/backends/advanced/src/advanced_omi_backend/workers/speaker_jobs.py index 0a4192fb..1576d9eb 100644 --- a/backends/advanced/src/advanced_omi_backend/workers/speaker_jobs.py +++ b/backends/advanced/src/advanced_omi_backend/workers/speaker_jobs.py @@ -12,22 +12,17 @@ from advanced_omi_backend.auth import generate_jwt_for_user from advanced_omi_backend.models.conversation import Conversation from advanced_omi_backend.models.job import async_job -from advanced_omi_backend.services.audio_stream import ( - TranscriptionResultsAggregator, -) +from advanced_omi_backend.services.audio_stream import TranscriptionResultsAggregator from advanced_omi_backend.speaker_recognition_client import SpeakerRecognitionClient from advanced_omi_backend.users import get_user_by_id +from advanced_omi_backend.utils.job_utils import update_job_meta logger = logging.getLogger(__name__) @async_job(redis=True, beanie=True) async def check_enrolled_speakers_job( - session_id: str, - user_id: str, - client_id: str, - *, - redis_client=None + session_id: str, user_id: str, client_id: str, *, redis_client=None ) -> Dict[str, Any]: """ Check if any enrolled speakers are present in the current audio stream. @@ -54,19 +49,23 @@ async def check_enrolled_speakers_job( # Check for enrolled speakers speaker_client = SpeakerRecognitionClient() - enrolled_present, speaker_result = await speaker_client.check_if_enrolled_speaker_present( - redis_client=redis_client, - client_id=client_id, - session_id=session_id, - user_id=user_id, - transcription_results=raw_results + enrolled_present, speaker_result = ( + await speaker_client.check_if_enrolled_speaker_present( + redis_client=redis_client, + client_id=client_id, + session_id=session_id, + user_id=user_id, + transcription_results=raw_results, + ) ) # Check for errors from speaker service if speaker_result and speaker_result.get("error"): error_type = speaker_result.get("error") error_message = speaker_result.get("message", "Unknown error") - logger.error(f"šŸŽ¤ [SPEAKER CHECK] Speaker service error: {error_type} - {error_message}") + logger.error( + f"šŸŽ¤ [SPEAKER CHECK] Speaker service error: {error_type} - {error_message}" + ) # For connection failures, assume no enrolled speakers but allow conversation to proceed # Speaker filtering is optional - if service is down, conversation should still be created @@ -82,7 +81,7 @@ async def check_enrolled_speakers_job( "enrolled_present": False, "identified_speakers": [], "skip_reason": f"Speaker service unavailable: {error_type}", - "processing_time_seconds": time.time() - start_time + "processing_time_seconds": time.time() - start_time, } # For other processing errors, also assume no enrolled speakers @@ -93,7 +92,7 @@ async def check_enrolled_speakers_job( "error_details": error_message, "enrolled_present": False, "identified_speakers": [], - "processing_time_seconds": time.time() - start_time + "processing_time_seconds": time.time() - start_time, } # Extract identified speakers @@ -101,31 +100,31 @@ async def check_enrolled_speakers_job( if speaker_result and "segments" in speaker_result: for seg in speaker_result["segments"]: identified_as = seg.get("identified_as") - if identified_as and identified_as != "Unknown" and identified_as not in identified_speakers: + if ( + identified_as + and identified_as != "Unknown" + and identified_as not in identified_speakers + ): identified_speakers.append(identified_as) processing_time = time.time() - start_time if enrolled_present: - logger.info(f"āœ… Enrolled speaker(s) found: {', '.join(identified_speakers)} ({processing_time:.2f}s)") + logger.info( + f"āœ… Enrolled speaker(s) found: {', '.join(identified_speakers)} ({processing_time:.2f}s)" + ) else: logger.info(f"ā­ļø No enrolled speakers found ({processing_time:.2f}s)") # Update job metadata for timeline tracking - from rq import get_current_job - current_job = get_current_job() - if current_job: - if not current_job.meta: - current_job.meta = {} - current_job.meta.update({ - "session_id": session_id, - "client_id": client_id, - "enrolled_present": enrolled_present, - "identified_speakers": identified_speakers, - "speaker_count": len(identified_speakers), - "processing_time": processing_time - }) - current_job.save_meta() + update_job_meta( + session_id=session_id, + client_id=client_id, + enrolled_present=enrolled_present, + identified_speakers=identified_speakers, + speaker_count=len(identified_speakers), + processing_time=processing_time, + ) return { "success": True, @@ -133,7 +132,7 @@ async def check_enrolled_speakers_job( "enrolled_present": enrolled_present, "identified_speakers": identified_speakers, "speaker_result": speaker_result, - "processing_time_seconds": processing_time + "processing_time_seconds": processing_time, } @@ -144,7 +143,7 @@ async def recognise_speakers_job( transcript_text: str = "", words: list = None, *, - redis_client=None + redis_client=None, ) -> Dict[str, Any]: """ RQ job function for identifying speakers in a transcribed conversation. @@ -168,12 +167,16 @@ async def recognise_speakers_job( Dict with processing results """ - logger.info(f"šŸŽ¤ RQ: Starting speaker recognition for conversation {conversation_id}") + logger.info( + f"šŸŽ¤ RQ: Starting speaker recognition for conversation {conversation_id}" + ) start_time = time.time() # Get the conversation - conversation = await Conversation.find_one(Conversation.conversation_id == conversation_id) + conversation = await Conversation.find_one( + Conversation.conversation_id == conversation_id + ) if not conversation: logger.error(f"Conversation {conversation_id} not found") return {"success": False, "error": "Conversation not found"} @@ -201,7 +204,7 @@ async def recognise_speakers_job( "conversation_id": conversation_id, "version_id": version_id, "speaker_recognition_enabled": False, - "processing_time_seconds": 0 + "processing_time_seconds": 0, } # Get provider capabilities from metadata @@ -222,7 +225,9 @@ async def recognise_speakers_job( # If we have existing segments from provider, proceed to identification if transcript_version.segments: - logger.info(f"šŸŽ¤ Using {len(transcript_version.segments)} segments from provider") + logger.info( + f"šŸŽ¤ Using {len(transcript_version.segments)} segments from provider" + ) # Continue to speaker identification below (after this block) else: logger.warning(f"šŸŽ¤ Provider claimed diarization but no segments found") @@ -237,32 +242,35 @@ async def recognise_speakers_job( if not actual_words and transcript_version.words: # Convert Word objects to dicts for speaker service API actual_words = [ - { - "word": w.word, - "start": w.start, - "end": w.end, - "confidence": w.confidence - } + {"word": w.word, "start": w.start, "end": w.end, "confidence": w.confidence} for w in transcript_version.words ] - logger.info(f"šŸ”¤ Loaded {len(actual_words)} words from transcript version.words field") + logger.info( + f"šŸ”¤ Loaded {len(actual_words)} words from transcript version.words field" + ) # Backward compatibility: Fall back to metadata if words field is empty (old data) elif not actual_words and transcript_version.metadata.get("words"): actual_words = transcript_version.metadata.get("words", []) - logger.info(f"šŸ”¤ Loaded {len(actual_words)} words from transcript version metadata (legacy)") + logger.info( + f"šŸ”¤ Loaded {len(actual_words)} words from transcript version metadata (legacy)" + ) # Backward compatibility: Extract from segments if that's all we have (old streaming data) elif not actual_words and transcript_version.segments: for segment in transcript_version.segments: if segment.words: for w in segment.words: - actual_words.append({ - "word": w.word, - "start": w.start, - "end": w.end, - "confidence": w.confidence - }) + actual_words.append( + { + "word": w.word, + "start": w.start, + "end": w.end, + "confidence": w.confidence, + } + ) if actual_words: - logger.info(f"šŸ”¤ Extracted {len(actual_words)} words from segments (legacy)") + logger.info( + f"šŸ”¤ Extracted {len(actual_words)} words from segments (legacy)" + ) if not actual_transcript_text: logger.warning(f"šŸŽ¤ No transcript text found in version {version_id}") @@ -271,7 +279,7 @@ async def recognise_speakers_job( "conversation_id": conversation_id, "version_id": version_id, "error": "No transcript text available", - "processing_time_seconds": 0 + "processing_time_seconds": 0, } # Check if we can run pyannote diarization @@ -290,7 +298,7 @@ async def recognise_speakers_job( "conversation_id": conversation_id, "version_id": version_id, "error": "No word timestamps and no segments available", - "processing_time_seconds": time.time() - start_time + "processing_time_seconds": time.time() - start_time, } # Has existing segments - fall through to run identification on them logger.info( @@ -303,6 +311,7 @@ async def recognise_speakers_job( # 1. Config toggle (per_segment_speaker_id) enables per-segment globally # 2. Manual reprocess trigger also enables per-segment for that run from advanced_omi_backend.config import get_misc_settings + misc_config = get_misc_settings() per_segment_config = misc_config.get("per_segment_speaker_id", False) @@ -323,7 +332,11 @@ async def recognise_speakers_job( # Have existing segments and can't/shouldn't run pyannote - do identification only # Covers: provider already diarized, no word timestamps but segments exist, etc. # Only send speech segments for identification; skip event/note segments - speech_segments = [s for s in transcript_version.segments if getattr(s, 'segment_type', 'speech') == 'speech'] + speech_segments = [ + s + for s in transcript_version.segments + if getattr(s, "segment_type", "speech") == "speech" + ] logger.info( f"šŸŽ¤ Using segment-level speaker identification on {len(speech_segments)} speech segments " f"(skipped {len(transcript_version.segments) - len(speech_segments)} non-speech)" @@ -341,10 +354,7 @@ async def recognise_speakers_job( ) else: # Standard path: full diarization + identification via speaker service - transcript_data = { - "text": actual_transcript_text, - "words": actual_words - } + transcript_data = {"text": actual_transcript_text, "words": actual_words} # Generate backend token for speaker service to fetch audio try: @@ -356,35 +366,41 @@ async def recognise_speakers_job( "conversation_id": conversation_id, "version_id": version_id, "error": "User not found", - "processing_time_seconds": time.time() - start_time + "processing_time_seconds": time.time() - start_time, } backend_token = generate_jwt_for_user(user_id, user.email) logger.info(f"šŸ” Generated backend token for speaker service") except Exception as token_error: - logger.error(f"Failed to generate backend token: {token_error}", exc_info=True) + logger.error( + f"Failed to generate backend token: {token_error}", exc_info=True + ) return { "success": False, "conversation_id": conversation_id, "version_id": version_id, "error": f"Token generation failed: {token_error}", - "processing_time_seconds": time.time() - start_time + "processing_time_seconds": time.time() - start_time, } - logger.info(f"šŸŽ¤ Calling speaker recognition service with conversation_id...") + logger.info( + f"šŸŽ¤ Calling speaker recognition service with conversation_id..." + ) speaker_result = await speaker_client.diarize_identify_match( conversation_id=conversation_id, backend_token=backend_token, transcript_data=transcript_data, - user_id=user_id + user_id=user_id, ) # Check for errors from speaker service if speaker_result.get("error"): error_type = speaker_result.get("error") error_message = speaker_result.get("message", "Unknown error") - logger.error(f"šŸŽ¤ Speaker recognition service error: {error_type} - {error_message}") + logger.error( + f"šŸŽ¤ Speaker recognition service error: {error_type} - {error_message}" + ) # Connection/timeout errors ā†’ skip gracefully (existing behavior) if error_type in ("connection_failed", "timeout", "client_error"): @@ -401,7 +417,7 @@ async def recognise_speakers_job( "identified_speakers": [], "skip_reason": f"Speaker service unavailable: {error_type}", "error_type": error_type, - "processing_time_seconds": time.time() - start_time + "processing_time_seconds": time.time() - start_time, } # Validation errors ā†’ fail job, don't retry @@ -414,7 +430,7 @@ async def recognise_speakers_job( "error": f"Validation error: {error_message}", "error_type": error_type, "retryable": False, # Don't retry validation errors - "processing_time_seconds": time.time() - start_time + "processing_time_seconds": time.time() - start_time, } # Resource errors ā†’ fail job, can retry later @@ -427,7 +443,7 @@ async def recognise_speakers_job( "error": f"Resource error: {error_message}", "error_type": error_type, "retryable": True, # Can retry later when resources available - "processing_time_seconds": time.time() - start_time + "processing_time_seconds": time.time() - start_time, } # Unknown errors ā†’ fail job @@ -439,11 +455,15 @@ async def recognise_speakers_job( "error": f"Speaker recognition failed: {error_type}", "error_details": error_message, "error_type": error_type, - "processing_time_seconds": time.time() - start_time + "processing_time_seconds": time.time() - start_time, } # Service worked but found no segments (legitimate empty result) - if not speaker_result or "segments" not in speaker_result or not speaker_result["segments"]: + if ( + not speaker_result + or "segments" not in speaker_result + or not speaker_result["segments"] + ): logger.warning(f"šŸŽ¤ Speaker recognition returned no segments") return { "success": True, @@ -451,7 +471,7 @@ async def recognise_speakers_job( "version_id": version_id, "speaker_recognition_enabled": True, "identified_speakers": [], - "processing_time_seconds": time.time() - start_time + "processing_time_seconds": time.time() - start_time, } speaker_segments = speaker_result["segments"] @@ -486,12 +506,16 @@ async def recognise_speakers_job( continue # Skip segments with invalid structure - if not isinstance(seg.get("start"), (int, float)) or not isinstance(seg.get("end"), (int, float)): + if not isinstance(seg.get("start"), (int, float)) or not isinstance( + seg.get("end"), (int, float) + ): empty_segment_count += 1 logger.debug(f"Filtered segment with invalid timing: {seg}") continue - speaker_name = seg.get("identified_as") or unknown_label_map.get(seg.get("speaker", "Unknown"), "Unknown Speaker") + speaker_name = seg.get("identified_as") or unknown_label_map.get( + seg.get("speaker", "Unknown"), "Unknown Speaker" + ) # Extract words from speaker service response (already matched to this segment) words_data = seg.get("words", []) @@ -500,13 +524,14 @@ async def recognise_speakers_job( word=w.get("word", ""), start=w.get("start", 0.0), end=w.get("end", 0.0), - confidence=w.get("confidence") + confidence=w.get("confidence"), ) for w in words_data ] # Classify segment type from content from advanced_omi_backend.utils.segment_utils import classify_segment_text + seg_classification = classify_segment_text(text) seg_type = "event" if seg_classification == "event" else "speech" @@ -519,18 +544,21 @@ async def recognise_speakers_job( segment_type=seg_type, identified_as=seg.get("identified_as"), confidence=seg.get("confidence"), - words=segment_words # Use words from speaker service + words=segment_words, # Use words from speaker service ) ) if empty_segment_count > 0: - logger.info(f"šŸ”‡ Filtered out {empty_segment_count} empty segments from speaker recognition") + logger.info( + f"šŸ”‡ Filtered out {empty_segment_count} empty segments from speaker recognition" + ) # Re-insert non-speech segments (event/note) that were skipped during identification # They need to be merged back into position based on timestamps non_speech_segments = [ - s for s in transcript_version.segments - if getattr(s, 'segment_type', 'speech') != 'speech' + s + for s in transcript_version.segments + if getattr(s, "segment_type", "speech") != "speech" ] if non_speech_segments: for ns_seg in non_speech_segments: @@ -541,7 +569,9 @@ async def recognise_speakers_job( insert_pos = i break updated_segments.insert(insert_pos, ns_seg) - logger.info(f"šŸŽ¤ Re-inserted {len(non_speech_segments)} non-speech segments") + logger.info( + f"šŸŽ¤ Re-inserted {len(non_speech_segments)} non-speech segments" + ) # Update the transcript version transcript_version.segments = updated_segments @@ -559,24 +589,31 @@ async def recognise_speakers_job( sr_metadata = { "enabled": True, - "identification_mode": "per_segment" if use_per_segment else "majority_vote", + "identification_mode": ( + "per_segment" if use_per_segment else "majority_vote" + ), "identified_speakers": list(identified_speakers), "speaker_count": len(identified_speakers), "total_segments": len(speaker_segments), - "processing_time_seconds": time.time() - start_time + "processing_time_seconds": time.time() - start_time, } if speaker_result.get("partial_errors"): sr_metadata["partial_errors"] = speaker_result["partial_errors"] transcript_version.metadata["speaker_recognition"] = sr_metadata # Set diarization source if pyannote ran (provider didn't do diarization) - if not provider_has_diarization and transcript_version.diarization_source != "provider": + if ( + not provider_has_diarization + and transcript_version.diarization_source != "provider" + ): transcript_version.diarization_source = "pyannote" await conversation.save() processing_time = time.time() - start_time - logger.info(f"āœ… Speaker recognition completed for {conversation_id} in {processing_time:.2f}s") + logger.info( + f"āœ… Speaker recognition completed for {conversation_id} in {processing_time:.2f}s" + ) return { "success": True, @@ -585,22 +622,18 @@ async def recognise_speakers_job( "speaker_recognition_enabled": True, "identified_speakers": list(identified_speakers), "segment_count": len(updated_segments), - "processing_time_seconds": processing_time + "processing_time_seconds": processing_time, } except asyncio.TimeoutError as e: logger.error(f"āŒ Speaker recognition timeout: {e}") # Add timeout metadata to job - from rq import get_current_job - current_job = get_current_job() - if current_job: - current_job.meta.update({ - "error_type": "timeout", - "audio_duration": conversation.audio_total_duration if conversation else None, - "timeout_occurred_at": time.time() - }) - current_job.save_meta() + update_job_meta( + error_type="timeout", + audio_duration=conversation.audio_total_duration if conversation else None, + timeout_occurred_at=time.time(), + ) return { "success": False, @@ -608,13 +641,16 @@ async def recognise_speakers_job( "version_id": version_id, "error": "Speaker recognition timeout", "error_type": "timeout", - "audio_duration": conversation.audio_total_duration if conversation else None, - "processing_time_seconds": time.time() - start_time + "audio_duration": ( + conversation.audio_total_duration if conversation else None + ), + "processing_time_seconds": time.time() - start_time, } except Exception as speaker_error: logger.error(f"āŒ Speaker recognition failed: {speaker_error}") import traceback + logger.debug(traceback.format_exc()) return { @@ -622,5 +658,5 @@ async def recognise_speakers_job( "conversation_id": conversation_id, "version_id": version_id, "error": str(speaker_error), - "processing_time_seconds": time.time() - start_time + "processing_time_seconds": time.time() - start_time, } diff --git a/backends/advanced/src/advanced_omi_backend/workers/transcription_jobs.py b/backends/advanced/src/advanced_omi_backend/workers/transcription_jobs.py index fb644ec1..ff9ecc45 100644 --- a/backends/advanced/src/advanced_omi_backend/workers/transcription_jobs.py +++ b/backends/advanced/src/advanced_omi_backend/workers/transcription_jobs.py @@ -37,7 +37,7 @@ from advanced_omi_backend.models.job import BaseRQJob, JobPriority, async_job from advanced_omi_backend.plugins.events import PluginEvent from advanced_omi_backend.services.audio_stream import TranscriptionResultsAggregator -from advanced_omi_backend.services.plugin_service import ensure_plugin_router +from advanced_omi_backend.services.plugin_service import dispatch_plugin_event from advanced_omi_backend.services.transcription import ( get_transcription_provider, is_transcription_available, @@ -50,6 +50,7 @@ analyze_speech, mark_conversation_deleted, ) +from advanced_omi_backend.utils.job_utils import update_job_meta logger = logging.getLogger(__name__) @@ -295,55 +296,26 @@ def _on_batch_progress(event: dict) -> None: # Trigger transcript-level plugins BEFORE speech validation # This ensures wake-word commands execute even if conversation gets deleted - logger.info( - f"šŸ” DEBUG: About to trigger plugins - transcript_text exists: {bool(transcript_text)}" - ) if transcript_text: try: - plugin_router = await ensure_plugin_router() - - if plugin_router: - logger.info( - f"šŸ” DEBUG: Preparing to trigger transcript plugins for conversation {conversation_id}" - ) - plugin_data = { + await dispatch_plugin_event( + event=PluginEvent.TRANSCRIPT_BATCH, + user_id=user_id, + data={ "transcript": transcript_text, "segment_id": f"{conversation_id}_batch", "conversation_id": conversation_id, "segments": segments, "word_count": len(words), - } - - logger.info( - f"šŸ”Œ DISPATCH: transcript.batch event " - f"(conversation={conversation_id[:12]}, words={len(words)})" - ) - - plugin_results = await plugin_router.dispatch_event( - event=PluginEvent.TRANSCRIPT_BATCH, - user_id=user_id, - data=plugin_data, - metadata={"client_id": client_id}, - ) - - logger.info( - f"šŸ”Œ RESULT: transcript.batch dispatched to {len(plugin_results) if plugin_results else 0} plugins" - ) - - if plugin_results: - logger.info( - f"āœ… Triggered {len(plugin_results)} transcript plugins in batch mode" - ) - for result in plugin_results: - if result.message: - logger.info(f" Plugin: {result.message}") + }, + metadata={"client_id": client_id}, + description=f"conversation={conversation_id[:12]}, words={len(words)}", + ) except Exception as e: logger.exception( f"āš ļø Error triggering transcript plugins in batch mode: {e}" ) - logger.info(f"šŸ” DEBUG: Plugin processing complete, moving to speech validation") - # Validate meaningful speech BEFORE any further processing transcript_data = {"text": transcript_text, "words": words} speech_analysis = analyze_speech(transcript_data) @@ -573,21 +545,14 @@ def _on_batch_progress(event: dict) -> None: ) # Update job metadata with title and summary for UI display - current_job = get_current_job() - if current_job: - if not current_job.meta: - current_job.meta = {} - current_job.meta.update( - { - "conversation_id": conversation_id, - "title": conversation.title, - "summary": conversation.summary, - "transcript_length": len(transcript_text), - "word_count": len(words), - "processing_time": processing_time, - } - ) - current_job.save_meta() + update_job_meta( + conversation_id=conversation_id, + title=conversation.title, + summary=conversation.summary, + transcript_length=len(transcript_text), + word_count=len(words), + processing_time=processing_time, + ) return { "success": True, @@ -972,18 +937,12 @@ async def stream_speech_detection_job( ) # Update job metadata to show status - if current_job: - if not current_job.meta: - current_job.meta = {} - current_job.meta.update( - { - "status": "listening_for_speech", - "session_id": session_id, - "client_id": client_id, - "session_level": True, # Mark as session-level job - } - ) - current_job.save_meta() + update_job_meta( + status="listening_for_speech", + session_id=session_id, + client_id=client_id, + session_level=True, # Mark as session-level job + ) # Track when session closes for graceful shutdown session_closed_at = None diff --git a/backends/advanced/webui/package-lock.json b/backends/advanced/webui/package-lock.json index 54ca06ae..c3bd503e 100644 --- a/backends/advanced/webui/package-lock.json +++ b/backends/advanced/webui/package-lock.json @@ -18,6 +18,7 @@ "d3-selection": "^3.0.0", "d3-time-format": "^4.1.0", "d3-zoom": "^3.0.0", + "framer-motion": "^11.0.0", "lucide-react": "^0.294.0", "react": "^18.2.0", "react-dom": "^18.2.0", @@ -3355,6 +3356,33 @@ "url": "https://github.com/sponsors/rawify" } }, + "node_modules/framer-motion": { + "version": "11.18.2", + "resolved": "https://registry.npmjs.org/framer-motion/-/framer-motion-11.18.2.tgz", + "integrity": "sha512-5F5Och7wrvtLVElIpclDT0CBzMVg3dL22B64aZwHtsIY8RB4mXICLrkajK4G9R+ieSAGcgrLeae2SeUTg2pr6w==", + "license": "MIT", + "dependencies": { + "motion-dom": "^11.18.1", + "motion-utils": "^11.18.1", + "tslib": "^2.4.0" + }, + "peerDependencies": { + "@emotion/is-prop-valid": "*", + "react": "^18.0.0 || ^19.0.0", + "react-dom": "^18.0.0 || ^19.0.0" + }, + "peerDependenciesMeta": { + "@emotion/is-prop-valid": { + "optional": true + }, + "react": { + "optional": true + }, + "react-dom": { + "optional": true + } + } + }, "node_modules/fs.realpath": { "version": "1.0.0", "resolved": "https://registry.npmjs.org/fs.realpath/-/fs.realpath-1.0.0.tgz", @@ -4028,6 +4056,21 @@ "node": ">=16 || 14 >=14.17" } }, + "node_modules/motion-dom": { + "version": "11.18.1", + "resolved": "https://registry.npmjs.org/motion-dom/-/motion-dom-11.18.1.tgz", + "integrity": "sha512-g76KvA001z+atjfxczdRtw/RXOM3OMSdd1f4DL77qCTF/+avrRJiawSG4yDibEQ215sr9kpinSlX2pCTJ9zbhw==", + "license": "MIT", + "dependencies": { + "motion-utils": "^11.18.1" + } + }, + "node_modules/motion-utils": { + "version": "11.18.1", + "resolved": "https://registry.npmjs.org/motion-utils/-/motion-utils-11.18.1.tgz", + "integrity": "sha512-49Kt+HKjtbJKLtgO/LKj9Ld+6vw9BjH5d9sc40R/kVyH8GLAXgT42M2NnuPcJNuA3s9ZfZBUcwIgpmZWGEE+hA==", + "license": "MIT" + }, "node_modules/ms": { "version": "2.1.3", "resolved": "https://registry.npmjs.org/ms/-/ms-2.1.3.tgz", @@ -5560,7 +5603,6 @@ "version": "2.8.1", "resolved": "https://registry.npmjs.org/tslib/-/tslib-2.8.1.tgz", "integrity": "sha512-oJFu94HQb+KVduSUQL7wnpmqnfmLsOA/nAh6b6EH0wCEoK0/mPeXU6c3wKDV83MkOuHPRHtSXKKU99IBazS/2w==", - "dev": true, "license": "0BSD" }, "node_modules/type-check": { diff --git a/backends/advanced/webui/package.json b/backends/advanced/webui/package.json index ca2c77a5..7c497790 100644 --- a/backends/advanced/webui/package.json +++ b/backends/advanced/webui/package.json @@ -13,6 +13,7 @@ "@tanstack/react-query": "^5.90.20", "axios": "^1.6.2", "clsx": "^2.0.0", + "framer-motion": "^11.0.0", "cronstrue": "^2.50.0", "d3-array": "^3.2.4", "d3-axis": "^3.0.0", diff --git a/backends/advanced/webui/src/components/UserLoopModal.tsx b/backends/advanced/webui/src/components/UserLoopModal.tsx new file mode 100644 index 00000000..6e671432 --- /dev/null +++ b/backends/advanced/webui/src/components/UserLoopModal.tsx @@ -0,0 +1,533 @@ +import { useState, useEffect, useCallback, useMemo, useRef } from 'react' +import { motion, AnimatePresence, PanInfo } from 'framer-motion' +import { X, Check, Heart, HeartCrack, Pencil, Play, Pause } from 'lucide-react' +import { api, BACKEND_URL } from '../services/api' +import { getStorageKey } from '../utils/storage' + +type DiffToken = { text: string; type: 'equal' | 'added' | 'removed' } + +/** Simple word-level diff using LCS to highlight changes. */ +function computeWordDiff(original: string, corrected: string): { originalTokens: DiffToken[]; correctedTokens: DiffToken[] } { + const a = original.split(/(\s+)/) + const b = corrected.split(/(\s+)/) + + // Build LCS table + const m = a.length, n = b.length + const dp: number[][] = Array.from({ length: m + 1 }, () => Array(n + 1).fill(0)) + for (let i = 1; i <= m; i++) { + for (let j = 1; j <= n; j++) { + dp[i][j] = a[i - 1] === b[j - 1] ? dp[i - 1][j - 1] + 1 : Math.max(dp[i - 1][j], dp[i][j - 1]) + } + } + + // Backtrack to get diff + const originalTokens: DiffToken[] = [] + const correctedTokens: DiffToken[] = [] + let i = m, j = n + const origReverse: DiffToken[] = [] + const corrReverse: DiffToken[] = [] + + while (i > 0 || j > 0) { + if (i > 0 && j > 0 && a[i - 1] === b[j - 1]) { + origReverse.push({ text: a[i - 1], type: 'equal' }) + corrReverse.push({ text: b[j - 1], type: 'equal' }) + i--; j-- + } else if (j > 0 && (i === 0 || dp[i][j - 1] >= dp[i - 1][j])) { + corrReverse.push({ text: b[j - 1], type: 'added' }) + j-- + } else { + origReverse.push({ text: a[i - 1], type: 'removed' }) + i-- + } + } + + originalTokens.push(...origReverse.reverse()) + correctedTokens.push(...corrReverse.reverse()) + return { originalTokens, correctedTokens } +} + +const AUTO_SHOW_KEY = 'userloop-auto-show' + +interface Suggestion { + id: string + annotation_type: string + conversation_id: string + segment_index: number | null + original_text: string + corrected_text: string + created_at: string + conversation_title: string | null + transcript_snippet: string | null + segment_start: number | null + segment_end: number | null +} + +/** Read auto-show preference from localStorage (default: false). */ +function getAutoShow(): boolean { + try { + return localStorage.getItem(AUTO_SHOW_KEY) === 'true' + } catch { + return false + } +} + +export default function UserLoopModal() { + const [isOpen, setIsOpen] = useState(false) + const [suggestions, setSuggestions] = useState([]) + const [currentIndex, setCurrentIndex] = useState(0) + const [direction, setDirection] = useState(0) + const [isAnimating, setIsAnimating] = useState(false) + const [particles, setParticles] = useState<{ id: number; x: number; y: number; type: 'heart' | 'heart-break' }[]>([]) + const [isEditing, setIsEditing] = useState(false) + const [editText, setEditText] = useState('') + const [isPlaying, setIsPlaying] = useState(false) + const audioRef = useRef(null) + + const stopAudio = useCallback(() => { + if (audioRef.current) { + audioRef.current.pause() + audioRef.current = null + } + setIsPlaying(false) + }, []) + + const fetchSuggestions = useCallback(async (): Promise => { + try { + const response = await api.get('/api/annotations/suggestions', { params: { limit: 20 } }) + const data = response.data + if (Array.isArray(data) && data.length > 0) { + setSuggestions(data) + setCurrentIndex(0) + return data + } + return [] + } catch { + return [] + } + }, []) + + // Auto-show: only poll & auto-open when the user has opted in via localStorage + useEffect(() => { + if (!getAutoShow()) return + + const check = async () => { + const data = await fetchSuggestions() + if (data.length > 0) setIsOpen(true) + } + check() + const interval = setInterval(check, 60000) + return () => clearInterval(interval) + }, [fetchSuggestions]) + + // Explicit trigger from Fine-tuning page (always works regardless of auto-show) + useEffect(() => { + const handler = () => { + fetchSuggestions().then(data => { + if (data.length > 0) setIsOpen(true) + }) + } + window.addEventListener('open-swipe-ui', handler) + return () => window.removeEventListener('open-swipe-ui', handler) + }, [fetchSuggestions]) + + // Stop audio on unmount + useEffect(() => { + return () => { stopAudio() } + }, [stopAudio]) + + // Stop audio when card changes + useEffect(() => { + stopAudio() + }, [currentIndex, stopAudio]) + + // Clean up particles + useEffect(() => { + const timer = setTimeout(() => setParticles([]), 1000) + return () => clearTimeout(timer) + }, [particles]) + + // Close modal when no suggestions left + useEffect(() => { + if (suggestions.length === 0 && isOpen) { + stopAudio() + setIsOpen(false) + } + }, [suggestions.length, isOpen, stopAudio]) + + // Keyboard shortcuts + useEffect(() => { + if (!isOpen || suggestions.length === 0) return + + const handleKeyDown = (e: KeyboardEvent) => { + // Don't capture keys when editing (textarea handles its own keys) + if (isEditing) return + + switch (e.key) { + case 'ArrowDown': + e.preventDefault() + handleSkip() + break + case 'ArrowUp': + e.preventDefault() + setEditText(suggestions[currentIndex]?.corrected_text || '') + setIsEditing(true) + break + case 'ArrowLeft': + e.preventDefault() + handleAction('reject', -1) + break + case 'ArrowRight': + e.preventDefault() + handleAction('accept', 1) + break + } + } + + window.addEventListener('keydown', handleKeyDown) + return () => window.removeEventListener('keydown', handleKeyDown) + }, [isOpen, suggestions, currentIndex, isEditing, isAnimating]) + + const createParticles = (type: 'heart' | 'heart-break') => { + setParticles( + Array.from({ length: 8 }, (_, i) => ({ + id: Date.now() + i, + x: Math.random() * 400 - 200, + y: Math.random() * 200 - 100, + type, + })) + ) + } + + const handleSkip = () => { + if (isAnimating) return + setIsEditing(false) + stopAudio() + if (currentIndex < suggestions.length - 1) { + setCurrentIndex(prev => prev + 1) + } else { + setIsOpen(false) + setSuggestions([]) + } + } + + const handleEditSave = async () => { + const suggestion = suggestions[currentIndex] + if (!suggestion) return + try { + await api.patch(`/api/annotations/${suggestion.id}`, { corrected_text: editText }) + // Update local state so diff re-renders with new text + setSuggestions(prev => prev.map((s, i) => i === currentIndex ? { ...s, corrected_text: editText } : s)) + } catch (error) { + console.error('Failed to save edit:', error) + } + setIsEditing(false) + } + + const togglePlay = () => { + const s = suggestions[currentIndex] + if (!s || s.segment_start == null || s.segment_end == null) return + + if (isPlaying && audioRef.current) { + stopAudio() + return + } + + const token = localStorage.getItem(getStorageKey('token')) || '' + const url = `${BACKEND_URL}/api/audio/chunks/${s.conversation_id}?start_time=${s.segment_start}&end_time=${s.segment_end}&token=${token}` + const audio = new Audio(url) + audioRef.current = audio + audio.addEventListener('ended', () => setIsPlaying(false)) + audio.play().then(() => setIsPlaying(true)).catch(() => setIsPlaying(false)) + } + + const handleAction = async (action: 'accept' | 'reject', swipeDirection: number) => { + const suggestion = suggestions[currentIndex] + if (!suggestion || isAnimating) return + + setIsAnimating(true) + setDirection(swipeDirection) + createParticles(action === 'accept' ? 'heart' : 'heart-break') + + try { + const status = action === 'accept' ? 'accepted' : 'rejected' + await api.patch(`/api/annotations/${suggestion.id}/status`, null, { + params: { status }, + }) + } catch (error) { + console.error(`Failed to ${action} suggestion:`, error) + } + + setTimeout(() => { + if (currentIndex < suggestions.length - 1) { + setCurrentIndex(prev => prev + 1) + } else { + setIsOpen(false) + setSuggestions([]) + } + setIsAnimating(false) + setDirection(0) + }, 400) + } + + const onPanEnd = (_event: MouseEvent | TouchEvent | PointerEvent, info: PanInfo) => { + if (isAnimating) return + const threshold = 100 + if (info.offset.x > threshold) { + handleAction('accept', 1) + } else if (info.offset.x < -threshold) { + handleAction('reject', -1) + } + } + + const diff = useMemo(() => { + if (!isOpen || suggestions.length === 0) return null + const current = suggestions[currentIndex] + return computeWordDiff(current.original_text, current.corrected_text) + }, [isOpen, suggestions, currentIndex]) + + if (!isOpen || suggestions.length === 0) return null + + const current = suggestions[currentIndex] + + const cardVariants = { + enter: (dir: number) => ({ x: dir > 0 ? 1000 : -1000, opacity: 0, scale: 0.8 }), + center: { zIndex: 1, x: 0, opacity: 1, scale: 1 }, + exit: (dir: number) => ({ zIndex: 0, x: dir > 0 ? 1000 : -1000, opacity: 0, scale: 0.8 }), + } + + return ( + + {isOpen && ( + +
+ {/* Particles */} + + {particles.map(p => ( + + {p.type === 'heart' ? ( + + ) : ( + + )} + + ))} + + + {/* Card */} + + {/* Status Overlays */} + + {direction > 0 && ( + + GOOD + + )} + {direction < 0 && ( + + NOPE + + )} + + + {/* Content */} + +
+ Review Suggestion +
+ + {current.conversation_title && ( +
+ {current.conversation_title} +
+ )} + + {/* Original vs corrected with diff highlighting */} +
+
+
Original
+
+ {diff?.originalTokens.map((t, i) => + t.type === 'removed' ? ( + {t.text} + ) : ( + {t.text} + ) + )} +
+
+
+
+ {isEditing ? ( + <> + + Editing... + + ) : ( + 'Suggested' + )} +
+ {isEditing ? ( +
+