kuromoji 2.0.0-dev.1
kuromoji: ^2.0.0-dev.1 copied to clipboard
Tokenize Japanese text using dictionaries defined by the Atilika project.
2.0.0-dev.1 #
TokenizerBuilderhas been removed, useTokenizer.buildSyncdirectly.Tokenizer.tokenizenow returns aList<UnknownToken>instead ofList<Map<String, Object?>>. To have the old behaviour, you can calltoJsonon the token instance.- Fixed an issue with surrogate pairs not being handled correctly.
- This also fixed an issue when parsing the character definitions file, which would cause the tokenizer to not work correctly in some cases.
1.0.5 #
- This adds web support and moves to a base64 embedding for the dictionaries. Thanks to Bruno D'Luka for his PR.
1.0.4 #
- This reverts the change made in 1.0.3 to remove embedded dicts, AoT compilation would fail.
1.0.3 #
- Removed embedded dicts.
1.0.2 #
- Fixed a condition that'd add more tokens than necessary
1.0.1 #
- Lowered the minimum sdk constraint.
1.0.0 #
- Initial version.