Error processing file: 500: Internal Server Error

Gemini API
,
1

Hi, I’m having problems processing audio files. It doesn’t always, but randomly returns error 500. I’m attaching log lines and part of the app code.
What could be happening?

LOG:
2024-08-13 18:06:54,869 - uvicorn.access - INFO - xxx.xxx.xxx.xxx:35136 - “POST /ai_api_audio_file/ HTTP/1.1” 200
2024-08-13 18:09:48,882 - google.auth.compute_engine._metadata - WARNING - Compute Engine Metadata server unavailable on attempt 1 of 3. Reason: timed out
2024-08-13 18:09:52,864 - google.auth.compute_engine._metadata - WARNING - Compute Engine Metadata server unavailable on attempt 2 of 3. Reason: timed out
2024-08-13 18:09:57,942 - google.auth.compute_engine._metadata - WARNING - Compute Engine Metadata server unavailable on attempt 3 of 3. Reason: timed out
2024-08-13 18:09:57,942 - google.auth._default - WARNING - Authentication failed using Compute Engine authentication due to unavailable metadata server.
2024-08-13 18:09:57,942 - gemini_api - ERROR - Error processing file:
No API_KEY or ADC found. Please either:
- Set the GOOGLE_API_KEY environment variable.
- Manually pass the key with genai.configure(api_key=my_api_key).
- Or set up Application Default Credentials, see Authentication with OAuth quickstart  |  Gemini API  |  Google AI for Developers for more information.
2024-08-13 18:09:57,943 - gemini_api - ERROR - Error processing file: 500: Internal Server Error
2024-08-13 18:09:57,943 - uvicorn.access - INFO - xxx.xxx.xxx.xxx:33504 - “POST /ai_api_audio_file/ HTTP/1.1” 500

CODE:
@app.post(“/ai_api_audio_file/”)
async def ai_api_audio_file(file: UploadFile = File(…), seconds: int = Form(…), username: str = Depends(get_current_username)):

file_extension = file.filename.split(".")[-1]
file_path = f'temp_file_{uuid.uuid4()}.{file_extension}'

try:
    
    async with aiofiles.open(file_path, 'wb') as f:
        content = await file.read()
        await f.write(content)

    prompt= "Make a transcription of the attached audio."
    
    result = await process_file(file_path, prompt)

except Exception as e:
    logger.error(f"Error processing file: {e}")
    raise HTTPException(status_code=500, detail="Internal Server Error")

finally:
    # Eliminar el archivo temporal
    if os.path.exists(file_path):
        os.remove(file_path)

return result

async def process_file(file_path: str, prompt: str):
try:

    uploaded_file = genai.upload_file(path=file_path)

    is_valid = await validate_limits()
    if not is_valid:
        return JSONResponse(
            status_code=200,
            content={"text": "Limite RPD"}
        )
    
    genai.configure(api_key=GOOGLE_API_KEY)
    model = genai.GenerativeModel('models/gemini-1.5-flash')
    response = model.generate_content(
        [prompt, uploaded_file],
        generation_config=generation_config,
        safety_settings=safety_settings,
        request_options={"timeout": 120},
        stream=False
    )
    
    
    total_token_count = response.usage_metadata.total_token_count
    api_call_registry.add_call(total_token_count)

    return handle_response(response)

except Exception as e:
    logger.error(f"Error processing file: {e}")
    raise HTTPException(status_code=500, detail="Internal Server Error")
2

The error message says how to proceed. The actual error code the servers generate for missing API key is 403 forbidden.

Hope that helps!

3

Thanks Nebulosa Orangia!
Yes, I saw the message and proceeded both ways and the result, randomly, is always the same.
I tried with the lines
load_dotenv()
GOOGLE_API_KEY = os.environ[‘GOOGLE_API_KEY’]
genai.configure(api_key=GOOGLE_API_KEY) and with

genai.configure(api_key=‘xxxxxxx’).

This is why I wrote here if someone had happened to them or knows where the problem could come from, beyond what the log indicates.