[0.1.6] - 2025-01-06

Added

• Support for streaming responses via a new wait attribute. When wait is enabled, the API streams partial output incrementally until completion, improving responsiveness for long-running generations.
• Dataset upload capabilities:
  • upload_dataframe() for pandas DataFrame uploads
  • upload_csv() for CSV file uploads
  • upload_excel() for Excel file uploads (in development)
• Machine learning operations:
  • train() method for model training with time series support
  • tune() method for hyperparameter optimization
  • inference() method for batch predictions
• Job management system with status tracking and metrics retrieval:
  • TrainingJob for training operations
  • TuningJob for hyperparameter tuning
  • InferenceJob for batch inference
• Time granularity support (second, minute, hour, daily, weekly, monthly, etc.)
• Comprehensive error handling with custom exceptions:
  • AuthenticationError for auth failures
  • RateLimitError for rate limiting
  • UploadError for upload failures
  • TrainingError for training issues
  • InferenceError for inference problems
  • TuningError for tuning errors

Changed

• Improved API design with Dataset objects
• Enhanced documentation with comprehensive docstrings
• Better separation of concerns with dedicated job classes

Fixed

• File existence validation before upload attempts
• Empty dataframe validation to prevent upload errors
• Proper error propagation from API responses