Skip to content

v1.3.0

Choose a tag to compare

@lucianotonet lucianotonet released this 03 Jun 20:35
· 41 commits to main since this release

🚀 Release v1.3.0

This release brings significant improvements to the Files/Batches module, including expanded support for batch processing of chat completions and audio operations.

🔄 Breaking Changes in Files/Batches Module

New JSONL Format Requirements (Fixes #14)

  • Each line must now follow Groq's official batch format:
    {
      "custom_id": "your-tracking-id",
      "method": "POST",
      "url": "/v1/chat/completions",
      "body": {
        "model": "llama-3.1-8b-instant",
        "messages": [...]
      }
    }

File Restrictions

  • Only .jsonl extension is now supported
  • Removed support for .json, .txt, .ndjson

✨ New Features

  • Batch processing support for:
    • Chat completions
    • Audio transcription
    • Audio translation
  • Enhanced validation and error handling

🛠️ Improvements

  • Additional MIME types support
  • Better JSONL file type detection
  • Expanded test coverage
  • Updated documentation

📝 Migration Guide

  1. Update your JSONL files:

    • Add custom_id field
    • Set method to "POST"
    • Use proper endpoint URL
    • Move request parameters inside body object
  2. Supported Endpoints:

    • /v1/chat/completions
    • /v1/audio/transcriptions
    • /v1/audio/translations

📚 Example JSONL File

{"custom_id": "chat-1", "method": "POST", "url": "/v1/chat/completions", "body": {"model": "llama-3.1-8b-instant", "messages": [{"role": "user", "content": "Hello"}]}}
{"custom_id": "audio-1", "method": "POST", "url": "/v1/audio/transcriptions", "body": {"model": "whisper-large-v3", "language": "en", "url": "https://example.com/audio.wav"}}

📋 Notes

  • Changes only affect the Files/Batches module
  • Core library features remain unchanged
  • All tests passing: 39 tests, 218 assertions

🔗 Links