Expand description
Japanese text → hiragana, katakana or romaji, optionally annotating kanji with their readings.
Readings come from lindera with the IPADIC dictionary embedded in the binary: nothing to download, no setup.
use sabiyomi::{Mode, Options, RomajiSystem, Sabiyomi, Target};
let s = Sabiyomi::new()?; // create once, reuse (it is `Send + Sync`)
assert_eq!(s.convert("日本語を勉強する", &Options::default())?, "にほんごをべんきょうする");
let romaji = Options { to: Target::Romaji, ..Options::default() };
assert_eq!(s.convert("東京へ行こう", &romaji)?, "tōkyōeikō");
let kunrei = Options { system: RomajiSystem::Kunrei, ..romaji };
assert_eq!(s.convert("富士山に", &kunrei)?, "huzisanni");
let annotated = Options { mode: Mode::Annotated, ..Options::default() };
assert_eq!(s.convert("漢字を読む", &annotated)?, "漢字(かんじ)を読(よ)む");Dictionary-free helpers (character classes, kana ↔ kana, kana → romaji)
live in kana.
Modules§
- kana
- Character classification and kana conversion. No dictionary involved, so everything here is cheap and never fails.
Structs§
- Error
- Morphological analysis failed (dictionary load or segmentation).
- Options
- Conversion settings. Start from
Options::defaultand override fields. - Parse
Option Error - A string didn’t name a known
Target,Mode,RomajiSystemorLongVowels. - Sabiyomi
- The converter. Holds the dictionary; create one and share it.
- Token
- One morpheme. Build these yourself to plug in any analyzer via
convert_tokens.
Enums§
- Long
Vowels - How long vowels are written in romaji.
- Mode
- Output layout.
- Romaji
System - Romanization system.
- Target
- Output script.
Functions§
- convert_
tokens - Converts pre-tokenized input. Use this to bring your own analyzer; tokens only need a surface, and a reading wherever there is kanji.