Ichiran is a collection of tools for working with text in Japanese language. It contains experimental segmenting and romanization algorithms and uses open source JMdictDB dictionary database to display meanings of words.
The web interface is under development right now. You can try it at ichi.moe.
!!!NEW!!! There's now a blog post which contains detailed instructions how to get Ichiran running on Linux and Windows. It also describes how to use the new ichiran-cli command line interface!
- Download JMDict data from here. If you want to initialize database from scratch download JMDict, and optionally kanjidic2.xml to use ichiran/kanji functionality.
- Create a settings.lisp file based on the provided settings.lisp.template file with the correct paths to the abovementioned files and the database connection parameters.
- The code can be loaded as a regular ASDF system. Use quicklisp to easily install all the dependencies.
-
- Easy mode: Use database dump from the release page to create a suitable database. Make sure
settings.lispcontains the correct connection parameters. Use(ichiran/maintenance:add-errata)to make database up to date. - Hard mode: Use
(ichiran/maintenance:full-init)to completely initialize the database. Use(ichiran/maintenance:load-jmdict)followed by(ichiran/maintenance:load-best-readings)to initialize onlyichiran/dictand notichiran/kanji. Either way, this will take a few hours or so.
- Easy mode: Use database dump from the release page to create a suitable database. Make sure
- Use
(ichiran/test:run-all-tests)to check that the installation satisfies the tests. - Before using any word segmenting functionality, run
(ichiran/dict:init-suffixes t)to create a suffix cache, which will improve the quality of segmentation.
There is no documentation yet. Any API is considered unstable at this point.
The basic functionality is (ichiran:romanize "一覧は最高だぞ" :with-info t), but feel free to explore further.