release: 2.5.1

Release v2.2.0

Bump version to 2.1.2

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

更新Python版本与类型检查

Release v2.5.0 - Multi-Modal PDF Analysis

Features:
- Multi-modal PDF analysis with PyMuPDF (10x faster)
- Image and table extraction
- Intelligent LLM auto-selection
- LLM capability detection

Bug Fixes:
- Gmail sender filter env var support
- Enhanced deduplicate log formatting
- Fixed collection name quotes

Testing:
- All workflows tested successfully
- 100+ new tests

chore: bump version to 2.4.0

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>

Release v2.4.0.r2

- Fix critical deduplication bug (collection key extraction)
- Improve deduplicate log output with Chinese labels
- Restore 10 items lost due to previous bug
- Items now correctly moved to 06_TRASHES collection

Release v2.3.0

Codebase cleanup and optimization:
- Remove dead code (metrics.py, format_response, auto_analyze.py)
- Consolidate documentation (AGENTS.md, GITHUB-ACTIONS-SETUP.md)
- Update tooling references (Black/isort → Ruff, pip → uv)
- Fix type annotation issues
- Simplify CLAUDE.md from 479 to 83 lines
- Remove unused docs/ and scripts/ directories

Total: 376 lines of dead code removed, documentation consolidated