raise ValueError(f"Error parsing file retrieve response: {str(e)}") This makes it impossible to know when an uploaded file becomes ready (ACTIVE), causing https://litellm/v1beta/models/gemini-2.5-flash:generateContent to fail with the error: The File <file_id> is not in an ACTIVE state and usage is not allowed

Fix Plan

The fix involves modifying the transform_retrieve_file_response function in transformation.py to handle the response from the Gemini API correctly.

• Modify the transform_retrieve_file_response function to catch and handle any exceptions that occur during response parsing.
• Add error logging to track any issues that arise during file retrieval.

Example code:

# litellm/llms/gemini/files/transformation.py

import logging

def transform_retrieve_file_response(raw_response, logging_obj, litellm_params):
    try:
        # existing code to parse the response
        # ...
    except Exception as e:
        logging_obj.error(f"Error parsing file retrieve response: {str(e)}")
        # return a default or error response
        return {"error": "Failed to parse response"}

Verification

To verify the fix, follow these steps:

• Upload a file to Gemini using the LiteLLM Proxy.
• Call the GET https://litellm/v1/files/<file_id> endpoint to retrieve the file status.
• Check the response to ensure it no longer returns a 500 error.
• Verify that the file status is correctly updated to "ACTIVE" after upload.

Extra Tips

• Ensure that the Gemini API is correctly configured and responding as expected.
• Review the LiteLLM Proxy logs to identify any other potential issues that may be contributing to the problem.
• Consider adding additional error handling and logging to the transform_retrieve_file_response function to improve debuggability.