Troubleshooting

This section covers common issues and their solutions when using the OpenGithubModelsApi.jl package.

Common Errors

Authentication Errors

Symptom: 401 Unauthorized errors

Solution:

• Verify your GitHub token has the necessary permissions
• Check that the token hasn't expired
• Ensure you're using the correct token format (ghp_...)

Invalid Model ID Format

Symptom: Error about model ID format

Solution:

• Model IDs must be in the format {publisher}/{model_name}
• Verify the model exists using list_models()
• Check for typos in the model ID

Parameter Validation Errors

Symptom: Errors about invalid parameters

Common issues:

• Temperature outside 0-1 range
• Invalid message roles
• Unsupported modalities

Solution:

• Check parameter constraints in the documentation
• Validate inputs before making API calls
• Use the type-safe constructors to catch errors early

Streaming Not Supported

Symptom: "Error: Streaming Not Supported" message

Solution:

• This package does not support streaming responses
• Set stream=false or omit the parameter (default is false)

Debugging Tips

Enable Verbose Mode

For more detailed response information:

response = create_chat_completion(client, request, verbose=true)

Inspect Model Data

To understand available models:

models = list_models(client)
for model in models
    println("Model ID: $(model.id)")
    println("Supported input modalities: $(model.supported_input_modalities)")
    println("Supported output modalities: $(model.supported_output_modalities)")
    println("---")
end

Validate Request Before Sending

try
    # This will validate the request structure
    request = InferenceRequest(
        model="openai/gpt-4.1",
        messages=[Message(role="user", content="Hello")],
        temperature=0.5
    )
    # If we get here, the request is valid
    response = create_chat_completion(client, request)
catch e
    println("Request validation failed: $e")
end

Frequently Asked Questions

Q: Why am I getting "Invalid message role" errors?

A: Message roles must be one of: "assistant", "developer", "system", or "user". Check your message definitions for typos.

Q: Can I use streaming responses with this package?

A: No, streaming responses are explicitly not supported by this package. Attempting to set stream=true will result in an error.

Q: How do I find available models?

A: Use the list_models(client) function to retrieve a list of all available models with their metadata.

Q: What's the difference between create_chat_completion and org_create_chat_completion?

A: org_create_chat_completion is for organization-scoped requests and requires an additional organization name parameter, while create_chat_completion is for standard requests.

Reporting Issues

If you encounter problems that aren't covered here:

1. Check the GitHub issues page
2. Include:
   • Package version
   • Julia version
   • Code snippet that reproduces the issue
   • Full error message
3. Open a new issue with the details