Skip to content

armgit / meet

Public Imported from GitHub
0

Open source video conferencing app powered by LiveKit. Built with Django and React.

🩹(summary) remove temporary audio files saved

lebaudantoinead0c3eea66a1
diff --git a/src/summary/summary/core/celery_worker.py b/src/summary/summary/core/celery_worker.py
index b45bdd41..3fcda952 100644
--- a/src/summary/summary/core/celery_worker.py
+++ b/src/summary/summary/core/celery_worker.py
@@ -1,6 +1,7 @@
"""Celery workers."""
import json
+import os
import tempfile
from pathlib import Path
@@ -28,6 +29,7 @@ celery = Celery(
)
if settings.sentry_dsn and settings.sentry_is_enabled:
+
@signals.celeryd_init.connect
def init_sentry(**_kwargs):
"""Initialize sentry."""
@@ -100,15 +102,20 @@ def process_audio_transcribe_summarize(filename: str, email: str, sub: str):
api_key=settings.openai_api_key, base_url=settings.openai_base_url
)
- logger.debug("Querying transcription …")
- with open(temp_file_path, "rb") as audio_file:
- transcription = openai_client.audio.transcriptions.create(
- model=settings.openai_asr_model, file=audio_file
- )
+ try:
+ logger.debug("Querying transcription …")
+ with open(temp_file_path, "rb") as audio_file:
+ transcription = openai_client.audio.transcriptions.create(
+ model=settings.openai_asr_model, file=audio_file
+ )
- transcription = transcription.text
+ transcription = transcription.text
- logger.debug("Transcription: \n %s", transcription)
+ logger.debug("Transcription: \n %s", transcription)
+ finally:
+ if os.path.exists(temp_file_path):
+ os.remove(temp_file_path)
+ logger.debug("Temporary file removed: %s", temp_file_path)
instructions = get_instructions(transcription)
summary_response = openai_client.chat.completions.create(