raise validation error if source file is invalid

maestro_worker_python/convert_files.py

@@ -4,6 +4,7 @@
 from dataclasses import dataclass
 from typing import List
 from subprocess import check_call
+from .response import ValidationError
 
 logger = logging.getLogger(__name__)
 
@@ -45,6 +46,24 @@ def _run_subprocess(command):
     try:
         process = subprocess.run(command, shell=True, capture_output=True, check=True)
     except subprocess.CalledProcessError as exc:
+        if "Invalid data found when processing input" in exc.stderr.decode():
+            logger.warning(
+                f"Could not convert file because of invalid data",
+                extra={'props': {'stderr': exc.stderr.decode(), 'stdout': exc.stdout.decode()}}
+            )
+            raise ValidationError(
+                f"Could not convert file because of invalid data: {exc.stderr.decode()}"
+            ) from exc
+
+        if "Output file #0 does not contain any stream" in exc.stderr.decode():
+            logger.warning(
+                f"Could not convert file because it has no audio data",
+                extra={'props': {'stderr': exc.stderr.decode(), 'stdout': exc.stdout.decode()}}
+            )
+            raise ValidationError(
+                f"Could not convert file because it has no audio data: {exc.stderr.decode()}"
+            ) from exc
+
         logger.error(
             f"Fatal error during conversion",
             extra={'props': {'stderr': exc.stderr.decode(), 'stdout': exc.stdout.decode()}}

maestro_worker_python/serve.py

@@ -50,7 +50,7 @@ async def internal_exception_handler(request: Request, exc: Exception):
 
 @app.exception_handler(ValidationError)
 async def validation_error_handler(request: Request, exc: ValidationError):
-    return JSONResponse(status_code=200, content=jsonable_encoder({"error":  exc.reason}))
+    return JSONResponse(status_code=400, content=jsonable_encoder({"error":  exc.reason}))
 
 
 @app.post("/inference", response_model=WorkerResponse)

tests/test_convert_files.py

@@ -3,15 +3,16 @@
 import subprocess
 from pathlib import Path
 from maestro_worker_python.convert_files import convert_files, FileToConvert
+from maestro_worker_python.response import ValidationError
 
 
 TEST_PATH = Path(__file__).resolve().parent
 
 
 @pytest.fixture(scope="session")
 def invalid_audio_file(tmp_path_factory):
-    fn = tmp_path_factory.mktemp("data") / "invalid_audio.mp3"
-    fn.write_bytes(b"invalid")
+    fn = tmp_path_factory.mktemp("data") / "invalid_audio.mov"
+    fn.write_bytes(b"")
     return fn
 
 
@@ -20,7 +21,7 @@ def test_should_re_raise_exceptions_in_thread(invalid_audio_file, file_format, c
     with pytest.raises(subprocess.CalledProcessError) as exc:
         convert_files(
             [FileToConvert(
-                input_file_path=invalid_audio_file,
+                input_file_path=TEST_PATH / "foobar.mp3",
                 output_file_path=f"{invalid_audio_file}.wav",
                 file_format=file_format,
             )]
@@ -30,6 +31,37 @@ def test_should_re_raise_exceptions_in_thread(invalid_audio_file, file_format, c
     assert caplog.records[0].message == "Fatal error during conversion"
 
 
+@pytest.mark.parametrize("file_format", ["m4a", "wav"])
+def test_should_raise_validation_error_if_audio_file_is_invalid(invalid_audio_file, file_format, caplog):
+    with pytest.raises(ValidationError) as exc:
+        convert_files(
+            [FileToConvert(
+                input_file_path=invalid_audio_file,
+                output_file_path=f"{invalid_audio_file}.wav",
+                file_format=file_format,
+            )]
+        )
+
+    assert "Could not convert file because of invalid data" in str(exc.value)
+    #assert caplog.records[0].levelname == "WARNING"
+    #assert caplog.records[0].message == "foobar"
+
+
+@pytest.mark.parametrize("file_format", ["m4a", "wav"])
+def test_should_raise_validation_error_if_source_has_no_audio(file_format, caplog):
+    input_file_path, output_file_path = TEST_PATH / "video-no-audio.mp4", TEST_PATH / "output.wav"
+    with pytest.raises(ValidationError) as exc:
+        convert_files(
+            [FileToConvert(
+                input_file_path=input_file_path,
+                output_file_path=output_file_path,
+                file_format=file_format,
+            )]
+        )
+
+    assert "Could not convert file because it has no audio data" in str(exc.value)
+
+
 def test_should_convert_valid_audio_file():
     input_file_path, output_file_path = TEST_PATH / "silent.ogg", TEST_PATH / "silent.wav"
     convert_files(