Skip to content

CHANGELOG.md

Latest commit

 

History

History
201 lines (155 loc) · 6.38 KB

CHANGELOG.md

File metadata and controls

201 lines (155 loc) · 6.38 KB

Changelog

All notable changes to smugVision will be documented in this file.

The format is based on Keep a Changelog, and this project adheres to Semantic Versioning.

[0.3.0] - 2025-11-24

MVP Release - Production Ready! 🎉

smugVision is now a fully functional, production-ready tool for automated photo metadata generation!

Added

Core Processing:

  • ImageProcessor class for orchestrating the complete processing pipeline
  • MetadataFormatter for combining AI-generated, EXIF, and face recognition metadata
  • End-to-end processing with detailed statistics tracking
  • Batch processing with progress indicators
  • Processing result tracking with success/skip/error counts

CLI Interface:

  • Production-ready __main__.py CLI entry point
  • Rich formatted output with banners and summaries
  • Support for --gallery (album key) and --url (SmugMug URL) inputs
  • --dry-run flag for previewing changes without updating SmugMug
  • --force-reprocess flag to reprocess already-tagged images
  • --include-videos flag to process video files (skipped by default)
  • --verbose and --quiet modes for logging control
  • --config option for custom configuration files
  • Exit codes for proper shell integration
  • Comprehensive error messages with troubleshooting hints

Face Recognition Enhancements:

  • Person name formatting (converts underscores to spaces)
  • Configurable confidence thresholds via config.yaml
  • Integration of identified people into captions and tags

Testing & Utilities:

  • test_processor.py for testing the full processing pipeline
  • Support for URL-based album resolution in test scripts
  • Dry-run mode with detailed preview output

Documentation:

  • Comprehensive README.md with installation, usage, and troubleshooting
  • QUICKSTART.md for fast setup and first use
  • Updated DESIGN.md with current status and architecture
  • CHANGELOG.md for tracking releases

Changed

  • Album processing now returns BatchProcessingStats with detailed metrics
  • Improved logging throughout the processing pipeline
  • Enhanced error handling with specific exception types
  • Metadata formatting now preserves existing captions/tags (configurable)

Fixed

  • Person names now display with spaces instead of underscores
  • Reference faces directory path now properly expands tilde (~)
  • Video downloads now use LargestVideo endpoint for actual video files
  • Image downloads use correct size-specific URLs from ImageSizes expansion

[0.2.0] - 2025-11-23

Enhanced Media Handling & EXIF Integration

Added

SmugMug Integration:

  • Video file detection via is_video property
  • LargestVideo endpoint support for proper video downloads
  • Configurable video inclusion/exclusion (skip by default)
  • Album resolution from URLs, node IDs, and names
  • Recursive album search within folder structures
  • URL path resolution for folder navigation
  • Pagination support for large datasets
  • Multiple image size options (Thumb through X3Large, Original)

EXIF & Location:

  • EXIF data extraction with GPS coordinates
  • Reverse geocoding for human-readable location names
  • Location context integration into captions and tags
  • HEIC/HEIF image format support
  • Automatic orientation correction

Testing:

  • test_smugmug.py with caching, URL parsing, and album listing
  • --cache flag for local image downloads
  • --size option for configurable download sizes
  • --force flag for re-downloading existing files
  • --include-videos flag in test script

Changed

  • Improved SmugMug API error messages
  • Enhanced album key resolution logic
  • Better handling of SmugMug folder hierarchies

Fixed

  • Video files now download correctly (not as thumbnails)
  • Image downloads use proper size-specific URLs
  • Content-type validation prevents HTML/JSON downloads
  • Pagination now correctly fetches all results

[0.1.0] - 2025-11-22

Initial MVP - Core Infrastructure

Added

Configuration System:

  • YAML-based configuration (~/.smugvision/config.yaml)
  • Interactive setup wizard
  • Default values for all settings
  • Configuration validation
  • Support for required and optional fields

SmugMug API Client:

  • OAuth 1.0a authentication
  • Album and image retrieval
  • Metadata updates (PATCH endpoint)
  • Marker tag system for tracking processed images
  • Error handling with custom exceptions
  • Rate limiting awareness

Vision Model Integration:

  • Factory pattern for vision models
  • Abstract base class for extensibility
  • Llama 3.2 Vision integration via Ollama
  • Caption generation with customizable prompts
  • Tag generation with keyword extraction
  • Temperature and max_tokens configuration

Face Recognition:

  • Face detection and recognition using face_recognition library
  • Reference faces management (folder-based organization)
  • Multiple reference images per person
  • Confidence-based matching
  • Relationship context integration
  • Face encoding optimization script

Cache Management:

  • Local image caching with folder structure mirroring
  • Skip existing files to avoid re-downloads
  • Configurable cache directory
  • Automatic directory creation

Testing & Utilities:

  • test_vision.py for vision model testing
  • debug_face_recognition.py for face detection debugging
  • get_smugmug_tokens.py for OAuth token acquisition
  • find_album_key.py for album discovery
  • optimize_reference_faces.py for face encoding optimization

Documentation:

  • DESIGN.md with architecture and roadmap
  • README_FACE_RECOGNITION.md with face recognition guide
  • config.yaml.example with all configuration options
  • Code documentation and inline comments

Core Components

  • smugvision.config - Configuration management
  • smugvision.smugmug - SmugMug API client and models
  • smugvision.vision - Vision model abstraction and implementations
  • smugvision.face - Face recognition system
  • smugvision.cache - Cache management
  • smugvision.utils - EXIF extraction, geocoding, and utilities

Future Plans

See DESIGN.md for detailed roadmap. Key planned features:

Version 1.0.0 (Future)

  • Folder batch processing
  • Cache cleanup utilities
  • Unit and integration tests
  • Performance optimizations
  • Web UI for monitoring
  • Multiple vision model support
  • Docker deployment option

Legend:

  • 🎉 Major milestone
  • ✨ New feature
  • 🐛 Bug fix
  • 📝 Documentation
  • ♻️ Refactoring
  • ⚡ Performance improvement