Fix audio upload validation cleanup

docs/v3/documentation/features/advanced/file-uploads.mdx

@@ -103,7 +103,7 @@ Our approach involves...
 
 **JSON Files**: Structured data is converted to string format.
 
-**Audio Files**: `.mp3` and `.wav` uploads are transcribed by OpenAI during upload using `whisper-1` by default. Large audio files may be split into multiple audio segments before transcription and then merged back into ordered transcript text before message chunking.
+**Audio Files**: `.mp3` and `.wav` uploads are transcribed directly by OpenAI during upload using `whisper-1` by default. The resulting transcript is then split into normal message-sized text chunks before messages are created.
 
 ### Chunking Strategy
 

src/utils/files.py

@@ -246,16 +246,14 @@ async def is_validated_audio_upload(file: UploadFile) -> bool:
 
     content_type = processor.normalize_content_type(filename, file.content_type or "")
     suffix = processor.get_output_suffix(filename, content_type)
-
-    with tempfile.NamedTemporaryFile(suffix=suffix, delete=False) as temp_file:
-        temp_path = Path(temp_file.name)
-        try:
-            while chunk := await file.read(UPLOAD_VALIDATION_CHUNK_BYTES):
-                temp_file.write(chunk)
-        finally:
-            await file.seek(0)
+    temp_path: Path | None = None
 
     try:
+        with tempfile.NamedTemporaryFile(suffix=suffix, delete=False) as temp_file:
+            temp_path = Path(temp_file.name)
+            while chunk := await file.read(UPLOAD_VALIDATION_CHUNK_BYTES):
+                temp_file.write(chunk)
+
         await asyncio.to_thread(
             processor.probe_audio_duration_seconds_from_path,
             temp_path,
@@ -266,7 +264,9 @@ async def is_validated_audio_upload(file: UploadFile) -> bool:
             return False
         raise
     finally:
-        temp_path.unlink(missing_ok=True)
+        await file.seek(0)
+        if temp_path is not None:
+            temp_path.unlink(missing_ok=True)
 
 
 class FileProcessingService:

tests/routes/test_files.py

@@ -16,6 +16,7 @@ from starlette.datastructures import Headers
 
 from src import models, schemas
 from src.config import settings
+from src.exceptions import ValidationException
 from src.models import Peer, Workspace
 from src.routers.messages import create_messages_with_file
 from src.utils.files import ExtractedFileText
@@ -946,7 +947,13 @@ async def test_malformed_audio_upload_returns_validation_error(
     form_data = {"peer_id": test_peer.name}
 
     url = _get_upload_url(test_workspace.name, session_name)
-    with patch.dict("src.utils.files.CLIENTS", {"openai": object()}, clear=False):
+    with (
+        patch.dict("src.utils.files.CLIENTS", {"openai": object()}, clear=False),
+        patch(
+            "src.utils.files.AudioProcessor._probe_audio_duration_seconds",
+            side_effect=ValidationException("Uploaded audio is invalid or unreadable"),
+        ),
+    ):
         response = client.post(url, files=files, data=form_data)
 
     assert response.status_code == 422

tests/utils/test_audio_processing.py

@@ -1,5 +1,6 @@
 import asyncio
 import io
+from types import TracebackType
 from unittest.mock import AsyncMock, patch
 
 import pytest
@@ -7,6 +8,7 @@ from fastapi import UploadFile
 from openai import AsyncOpenAI
 from starlette.datastructures import Headers
 
+import src.utils.files as file_utils
 from src.config import settings
 from src.exceptions import ValidationException
 from src.utils.clients import CLIENTS, transcribe_audio
@@ -299,3 +301,39 @@ async def test_audio_processor_normalizes_small_mime_only_audio_filename():
 
     assert extracted.text == "normalized"
     assert extracted.metadata["audio_segment_count"] == 1
+
+
+@pytest.mark.asyncio
+async def test_validated_audio_upload_cleans_up_temp_file_on_write_failure():
+    file = UploadFile(
+        file=io.BytesIO(b"audio-bytes"),
+        filename="voice.mp3",
+        headers=Headers({"content-type": "audio/mpeg"}),
+    )
+
+    class FailingTempFile:
+        name: str = "/tmp/test-audio-validation.mp3"
+
+        def __enter__(self) -> "FailingTempFile":
+            return self
+
+        def __exit__(
+            self,
+            exc_type: type[BaseException] | None,
+            exc: BaseException | None,
+            tb: TracebackType | None,
+        ) -> bool:
+            return False
+
+        def write(self, _chunk: bytes) -> int:
+            raise OSError("disk full")
+
+    with (
+        patch("src.utils.files.tempfile.NamedTemporaryFile", return_value=FailingTempFile()),
+        patch("src.utils.files.Path.unlink") as mock_unlink,
+        pytest.raises(OSError, match="disk full"),
+    ):
+        await file_utils.is_validated_audio_upload(file)
+
+    mock_unlink.assert_called_once_with(missing_ok=True)
+    assert file.file.tell() == 0