Enhance speaker recognition service integration and error handling

.github/workflows/robot-tests.yml

@@ -28,6 +28,7 @@ jobs:
       env:
         DEEPGRAM_API_KEY: ${{ secrets.DEEPGRAM_API_KEY }}
         OPENAI_API_KEY: ${{ secrets.OPENAI_API_KEY }}
+        HF_TOKEN: ${{ secrets.HF_TOKEN }}
       run: |
         echo "Verifying required secrets..."
         if [ -z "$DEEPGRAM_API_KEY" ]; then
@@ -38,8 +39,13 @@ jobs:
           echo "❌ ERROR: OPENAI_API_KEY secret is not set"
           exit 1
         fi
+        if [ -z "$HF_TOKEN" ]; then
+          echo "❌ ERROR: HF_TOKEN secret is not set"
+          exit 1
+        fi
         echo "✓ DEEPGRAM_API_KEY is set (length: ${#DEEPGRAM_API_KEY})"
         echo "✓ OPENAI_API_KEY is set (length: ${#OPENAI_API_KEY})"
+        echo "✓ HF_TOKEN is set (length: ${#HF_TOKEN})"
         echo "✓ All required secrets verified"
 
     - name: Set up Docker Buildx
@@ -71,28 +77,6 @@ jobs:
       run: |
         uv pip install --system robotframework robotframework-requests python-dotenv websockets
 
-    - name: Create test environment file
-      working-directory: tests/setup
-      run: |
-        cat > .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
-
-        # API Keys (from GitHub secrets)
-        OPENAI_API_KEY=${{ secrets.OPENAI_API_KEY }}
-        DEEPGRAM_API_KEY=${{ secrets.DEEPGRAM_API_KEY }}
-
-        # Test Configuration
-        TEST_TIMEOUT=120
-        TEST_DEVICE_NAME=robot-test
-        EOF
-
     - name: Create test config.yml
       run: |
         echo "Copying test configuration file..."
@@ -101,122 +85,17 @@ jobs:
         echo "✓ Test config.yml created from tests/configs/deepgram-openai.yml"
         ls -lh config/config.yml
 
-    - name: Start test environment
-      working-directory: backends/advanced
-      env:
-        DEEPGRAM_API_KEY: ${{ secrets.DEEPGRAM_API_KEY }}
-        OPENAI_API_KEY: ${{ secrets.OPENAI_API_KEY }}
-        LLM_PROVIDER: openai
-        TRANSCRIPTION_PROVIDER: deepgram
-        MEMORY_PROVIDER: friend_lite
-      run: |
-        # Debug: Check if secrets are available
-        echo "Checking environment variables..."
-        echo "DEEPGRAM_API_KEY is set: $([ -n "$DEEPGRAM_API_KEY" ] && echo 'YES' || echo 'NO')"
-        echo "OPENAI_API_KEY is set: $([ -n "$OPENAI_API_KEY" ] && echo 'YES' || echo 'NO')"
-        echo "LLM_PROVIDER: $LLM_PROVIDER"
-        echo "TRANSCRIPTION_PROVIDER: $TRANSCRIPTION_PROVIDER"
-
-        # Clean any existing test containers for fresh start
-        echo "Cleaning up any existing test containers..."
-        docker compose -f docker-compose-test.yml down -v || true
-
-        # Start ALL services in parallel - Docker Compose handles dependencies via healthchecks
-        echo "Starting all services in parallel (docker-compose-test.yml)..."
-        echo "Note: Using test compose file with source mounts for faster startup"
-
-        # Export API keys so docker-compose can use them
-        export DEEPGRAM_API_KEY
-        export OPENAI_API_KEY
-        export LLM_PROVIDER
-        export TRANSCRIPTION_PROVIDER
-        export MEMORY_PROVIDER
-
-        DOCKER_BUILDKIT=0 docker compose -f docker-compose-test.yml up -d
-
-        # Show container status
-        echo "Container status:"
-        docker compose -f docker-compose-test.yml ps
-
-        # Single wait for backend readiness (backend depends_on ensures infra is ready)
-        echo "Waiting for backend readiness (up to 120s)..."
-        for i in {1..40}; do
-          if curl -s http://localhost:8001/readiness > /dev/null 2>&1; then
-            echo "✓ Backend is ready (all dependencies satisfied)"
-            break
-          fi
-          # Show logs every 10 attempts to help debug
-          if [ $((i % 10)) -eq 0 ]; then
-            echo "Still waiting... showing recent logs:"
-            docker compose -f docker-compose-test.yml logs --tail=20 chronicle-backend-test
-          fi
-          if [ $i -eq 40 ]; then
-            echo "✗ Backend failed to start - showing full logs:"
-            docker compose -f docker-compose-test.yml logs
-            exit 1
-          fi
-          echo "Attempt $i/40..."
-          sleep 3
-        done
-
-        echo "✓ Backend is ready!"
-
-        # Verify workers are registered with Redis (Robot tests need stable workers)
-        echo "Waiting for workers to register with Redis (up to 60s)..."
-        for i in {1..30}; do
-          WORKER_COUNT=$(docker compose -f docker-compose-test.yml exec -T workers-test uv run python -c 'from rq import Worker; from redis import Redis; import os; r = Redis.from_url(os.getenv("REDIS_URL", "redis://redis-test:6379/0")); print(len(Worker.all(connection=r)))' 2>/dev/null || echo "0")
-
-          if [ "$WORKER_COUNT" -ge 6 ]; then
-            echo "✓ Found $WORKER_COUNT workers registered"
-            # Show worker details
-            docker compose -f docker-compose-test.yml exec -T workers-test uv run python -c 'from rq import Worker; from redis import Redis; import os; r = Redis.from_url(os.getenv("REDIS_URL", "redis://redis-test:6379/0")); workers = Worker.all(connection=r); print(f"Total registered workers: {len(workers)}"); [print(f"  - {w.name}: queues={w.queue_names()}, state={w.get_state()}") for w in workers]'
-            break
-          fi
-
-          if [ $i -eq 30 ]; then
-            echo "✗ Workers failed to register after 60s"
-            echo "Showing worker logs:"
-            docker compose -f docker-compose-test.yml logs --tail=50 workers-test
-            exit 1
-          fi
-
-          echo "Attempt $i/30: $WORKER_COUNT workers registered (waiting for 6+)..."
-          sleep 2
-        done
-
-        echo "✓ All services ready!"
-
-    - name: Verify checked out code
-      working-directory: tests
-      run: |
-        echo "Current git commit:"
-        git log -1 --oneline
-        echo ""
-        echo "Test files in current checkout:"
-        find . -name "*.robot" -type f | head -10
-        echo ""
-        echo "Sample of tags in test files:"
-        grep -h "\[Tags\]" endpoints/*.robot infrastructure/*.robot integration/*.robot 2>/dev/null | head -20 || echo "No tag files found"
-
-    - name: Clean previous test results
-      working-directory: tests
-      run: |
-        echo "Cleaning any previous test results..."
-        rm -rf results
-        mkdir -p results
-        echo "✓ Fresh results directory created"
-
     - name: Run Robot Framework tests
       working-directory: tests
       env:
-        # Required for backend imports in test libraries
-        OPENAI_API_KEY: ${{ secrets.OPENAI_API_KEY }}
-        OPENAI_BASE_URL: https://api.openai.com/v1
-        OPENAI_MODEL: gpt-4o-mini
+        # Required for test runner script
         DEEPGRAM_API_KEY: ${{ secrets.DEEPGRAM_API_KEY }}
+        OPENAI_API_KEY: ${{ secrets.OPENAI_API_KEY }}
+        HF_TOKEN: ${{ secrets.HF_TOKEN }}
+        CLEANUP_CONTAINERS: "false"  # Don't cleanup in CI - handled by workflow
       run: |
-        # Run all tests (don't fail workflow to allow artifact upload)
-        make all OUTPUTDIR=results
+        # Use the unified test script that mirrors local development
+        ./run-robot-tests.sh
         TEST_EXIT_CODE=$?
         echo "test_exit_code=$TEST_EXIT_CODE" >> $GITHUB_ENV
         exit 0  # Don't fail here, we'll fail at the end after uploading artifacts

backends/advanced/docker-compose-test.yml

@@ -39,9 +39,9 @@ services:
       - OPENMEMORY_USER_ID=${OPENMEMORY_USER_ID:-openmemory}
       - MYCELIA_URL=http://mycelia-backend-test:5173
       - MYCELIA_DB=mycelia_test
-      # Disable speaker recognition in test environment to prevent segment duplication
+      # Enable speaker recognition in test environment
       - DISABLE_SPEAKER_RECOGNITION=false
-      - SPEAKER_SERVICE_URL=https://localhost:8085
+      - SPEAKER_SERVICE_URL=http://speaker-service-test:8085
       - CORS_ORIGINS=http://localhost:3001,http://localhost:8001,https://localhost:3001,https://localhost:8001
       # Set low inactivity timeout for tests (2 seconds instead of 60)
       - SPEECH_INACTIVITY_THRESHOLD_SECONDS=2
@@ -54,6 +54,8 @@ services:
         condition: service_healthy
       redis-test:
         condition: service_started
+      speaker-service-test:
+        condition: service_healthy
     healthcheck:
       test: ["CMD", "curl", "-f", "http://localhost:8000/readiness"]
       interval: 10s
@@ -119,6 +121,36 @@ services:
       timeout: 3s
       retries: 5
 
+  speaker-service-test:
+    build:
+      context: ../../extras/speaker-recognition
+      dockerfile: Dockerfile
+      args:
+        PYTORCH_CUDA_VERSION: cpu
+    image: speaker-recognition-test:latest
+    ports:
+      - "8086:8085"  # Avoid conflict with dev speaker service on 8085
+    volumes:
+      - ../../extras/speaker-recognition/src:/app/src
+      - ../../extras/speaker-recognition/model_cache:/models
+      - ../../extras/speaker-recognition/audio_chunks:/app/audio_chunks
+      - ../../extras/speaker-recognition/debug:/app/debug
+      - ../../extras/speaker-recognition/speaker_data:/app/data
+    environment:
+      - HF_HOME=/models
+      - HF_TOKEN=${HF_TOKEN}
+      - SIMILARITY_THRESHOLD=0.15
+      - SPEAKER_SERVICE_HOST=0.0.0.0
+      - SPEAKER_SERVICE_PORT=8085
+      - DEEPGRAM_API_KEY=${DEEPGRAM_API_KEY}
+    healthcheck:
+      test: ["CMD", "curl", "-f", "http://localhost:8085/health"]
+      interval: 30s
+      timeout: 10s
+      retries: 5
+      start_period: 60s
+    restart: unless-stopped
+
   workers-test:
     build:
       context: .
@@ -151,7 +183,7 @@ services:
       - MYCELIA_URL=http://mycelia-backend-test:5173
       - MYCELIA_DB=mycelia_test
       - DISABLE_SPEAKER_RECOGNITION=false
-      - SPEAKER_SERVICE_URL=https://localhost:8085
+      - SPEAKER_SERVICE_URL=http://speaker-service-test:8085
       # Set low inactivity timeout for tests (2 seconds instead of 60)
       - SPEECH_INACTIVITY_THRESHOLD_SECONDS=2
       # Wait for audio queue to drain before timing out (test mode)
@@ -165,6 +197,8 @@ services:
         condition: service_started
       qdrant-test:
         condition: service_started
+      speaker-service-test:
+        condition: service_healthy
     restart: unless-stopped
 
   # Mycelia - AI memory and timeline service (test environment)
@@ -237,6 +271,11 @@ services:
   #       condition: service_healthy
   #   restart: unless-stopped
 
+# Use default bridge network for test isolation (no external network dependency)
+networks:
+  default:
+    driver: bridge
+
 # CI Considerations (for future implementation):
 # - GitHub Actions can run these services in isolated containers
 # - Port conflicts won't exist in CI since each job runs in isolation

backends/advanced/run-test.sh

@@ -166,9 +166,8 @@ if [ ! -f "diarization_config.json" ] && [ -f "diarization_config.json.template"
     print_success "diarization_config.json created"
 fi
 
-# Install dependencies with uv
-print_info "Installing dependencies with uv..."
-uv sync --dev --group test
+# Note: Robot Framework dependencies are managed via tests/test-requirements.txt
+# The integration tests use Docker containers for service dependencies
 
 # Set up environment variables for testing
 print_info "Setting up test environment variables..."
@@ -211,8 +210,9 @@ export DOCKER_BUILDKIT=0
 export TEST_MODE=dev
 
 # Run the Robot Framework integration tests with extended timeout (mem0 needs time for comprehensive extraction)
+# IMPORTANT: Robot tests must be run from the repository root where backends/ and tests/ are siblings
 print_info "Starting Robot Framework integration tests (timeout: 15 minutes)..."
-if timeout 900 uv run robot --outputdir ../../test-results --loglevel INFO ../../tests/integration/integration_test.robot; then
+if (cd ../.. && timeout 900 robot --outputdir test-results --loglevel INFO tests/integration/integration_test.robot); then
     print_success "Integration tests completed successfully!"
 else
     TEST_EXIT_CODE=$?

backends/advanced/src/advanced_omi_backend/services/memory/providers/chronicle.py

@@ -84,9 +84,16 @@ async def initialize(self) -> None:
             vector_ok = await self.vector_store.test_connection()
 
             if not llm_ok:
-                raise RuntimeError("LLM provider connection failed")
+                raise RuntimeError(
+                    f"LLM provider connection failed for {self.config.llm_provider.value}. "
+                    f"Check API keys, network connectivity, and service availability. "
+                    f"Memory processing cannot proceed without a working LLM connection."
+                )
             if not vector_ok:
-                raise RuntimeError("Vector store connection failed")
+                raise RuntimeError(
+                    f"Vector store connection failed for {self.config.vector_store_provider.value}. "
+                    f"Check that Qdrant service is running and accessible."
+                )
 
             self._initialized = True
             memory_logger.info(