Use this page to find the right level of detail without searching through the main README.
- Project overview and quick start
- Design philosophy
- Data sources, attribution, and provenance
- Layered lexicon packs
- Unified KDIC v2 language packs
- RAC-only model rebuild
- Migration from 0.1.1 to 0.2
- PyPI release guide
- 0.2 release checklist
- 0.2 release notes
- 0.3 release notes
- 0.3.1 release notes
- Community corpus roadmap
- Evaluation guide: curated gate, legacy diagnostics, metrics, and commands
- Runtime word composition: composition split, keep-list, and completion cap
- Spelling authority: official vs reviewed community spellings
- Safe word breaks: ZWSP layout rules, composition breaks, and verification
- Typo detection: whole-span diagnostics, suggestions, and edit scripts
- Benchmark results: accuracy and runtime measurements
- Development workflows: tests, corpus preparation, frequencies, dictionaries, and unknown-word review
- Prepare dictionaries for Python, C, and Rust: local source preparation, KDIC conversion, testing, and deployment
For application developers: start with the main README, then select the Python, C, or Rust implementation guide.
For NLP evaluation: read the data policy first, followed by the evaluation and benchmark documents.
For contributors: read the development workflow and the porting reference before changing shared dictionary or normalization behavior.