Lindera SudachiDict
Lindera SudachiDict is a Japanese dictionary crate based on SudachiDict, the actively maintained dictionary used by the Sudachi morphological analyzer. It is updated several times a year upstream, so recent vocabulary (令和, スマホ, 推し活, ...) is in-vocabulary. Each entry has 19 fields, including SudachiDict-specific columns such as the normalized form, split references, and synonym group IDs.
The dictionary source files are mirrored at lindera/sudachidict. The bundled version is 20260723 (small + core + notcore lexicons).
[!NOTE] This dictionary reproduces SudachiDict's vocabulary and lattice behavior, but not the Sudachi engine plugins (katakana/numeric joining, input normalization, A/B/C split modes). See Behavioral differences from Sudachi for details.
Contents
- Dictionary Format -- Field definitions for system and user dictionaries
- Build -- How to build the dictionary from source
- Examples -- Tokenization examples