Translation Fails or Hangs - protospatial/NodeToCode GitHub Wiki

Troubleshooting → Translation Fails or Hangs

Common Error Responses

When using cloud-based LLM providers (Anthropic, OpenAI, DeepSeek), you may encounter these standard HTTP error responses:

Tip: Check the Output Log with MinSeverity set to Debug to see detailed error responses.

Code Type Description Resolution
400 invalid_request_error Issue with request format/content Check plugin settings and reduce complexity if needed
401 authentication_error Invalid or missing API key Verify API key in Project Settings
403 permission_error Insufficient permissions Check API key permissions and usage tier
404 not_found_error Resource not found Verify endpoint URLs and model availability
413 request_too_large Request exceeds size limits Reduce reference files or translation depth
429 rate_limit_error Rate limit exceeded Wait before retrying or check usage quotas
500 api_error Provider internal error Wait and retry; check provider status
529 overloaded_error Service temporarily overloaded Wait and retry later

Translation Timeout Issues

If your translation appears to hang indefinitely:

[!NOTE] Node to Code's default timeout is 1000 seconds, but translations typically complete within 30-60 seconds depending on graph complexity and LLM provider & model.

Check for:

  • Network connectivity issues
  • LLM provider service status
  • Token limit exceeded
  • Request size too large

Resolution Steps:

  1. Check the Output Log for timeout messages
  2. Verify internet connection
  3. Confirm LLM provider status
  4. Reduce translation complexity if needed:
    • Lower translation depth
    • Remove unnecessary reference files
    • Split large graphs into smaller components
    • Use a model with larger context window

Ollama-Specific Issues

When using local LLM processing through Ollama, unique issues may arise:

  1. Service Connection

    # Verify Ollama is running with:
curl http://localhost:11434/api/version
    • Check endpoint configuration in Plugin Settings
    • Ensure Ollama service is running
    • Verify network access if running on remote machine

  2. Model Management

    # List installed models
ollama list

# Pull required model if missing
ollama pull codellama:code
    • Confirm model is installed
    • Check model name matches Plugin Settings
    • Verify model compatibility

  3. Resource Constraints

    • Monitor system resources:
      • System RAM usage
      • CPU utilization
      • GPU availability and VRAM usage (if using GPU acceleration)
      • Disk space for model storage

  4. Common Ollama Errors

    • You may see errors pertaining to the following:
      • Failed to connect to Ollama service
      • Model not found
      • Out of memory
      • GPU not available

    Resolution Steps:

    1. Restart Ollama service
    2. Free up system resources
    3. Check Ollama logs for detailed error messages
    4. Refer to Ollama's documentation and/or community for further troubleshooting

General Troubleshooting Steps

  1. Check Logs
    • Review Output Log for specific error messages
    • Check HTTP response codes
    • Monitor token usage warnings
    • Look for timeout indicators

[!NOTE] Set MinSeverity to Debug in Plugin Settings for maximum logging detail during troubleshooting.

  1. Validate Configuration

    • Confirm plugin settings
    • Test API connectivity
    • Verify file paths for reference files
    • Check model selection and availability

  2. Reduce Complexity

    • Lower translation depth
    • Remove reference files
    • Simplify Blueprint graph
    • Split into smaller translations

[!NOTE] Tip: Start with a simple graph translation to establish baseline functionality before attempting more complex translations.

Prevention Tips

  1. Regular Maintenance

    • Keep API keys updated
    • Monitor token usage and costs
    • Update plugin regularly
    • Clean up reference files

  2. Best Practices

    • Start with simple translations
    • Gradually increase complexity
    • Document successful configurations
    • Maintain backup API keys

[!NOTE] Keep reference files under 50KB - 100KB total (depending on LLM context window) for optimal performance and reduced token usage.

When to Seek Support

If issues persist after trying the above solutions, prepare the following information before submitting an issue in the Node to Code Discord:

  1. Complete error messages from Output Log
  2. Plugin version
  3. Unreal Engine version
  4. LLM provider and model
  5. Sample Blueprint causing issues (if possible)
  6. Steps to reproduce the problem
  7. Any relevant plugin settings

Remember that most translation issues can be resolved by:

  1. Checking error responses in the Output Log
  2. Validating configuration and credentials
  3. Ensuring proper resource availability
  4. Starting with simpler translations to isolate issues. See Blueprint Selection for guidance.