Kuromojure is a general purpose, practical and concise Clojure wrapper around Kuromoji (version 0.7.7), a Japanese Morphological Analyser designed for search.
Warning: kuromojure is very unstable. The API will probably change some times in the near future, so visit this page often to get the updates.
Add the following entry to the :dependencies vector of your project.clj file:
Tokenizer modes are :normal, :search and :extended. Omitting
mode defaults to :normal.
user=> (use 'kuromojure)
user=> (with-tokenizer :search
(-> "日本" tokenize first))
;;=> {:surface "日本"
; :known true
; :reading "ニッポン"
; :features [:noun :proper :place :country nil nil "日本" "ニッポン" "ニッポン"]
; :base-form "日本"
; :part-of-speech [:noun :proper :place :country]}- Upgrade Kuromoji to version 0.8.0 (will release Kuromojure 0.8.0 then)
The intermediate fns logic for translation applies here too.
Use clj-tokenize to tokenize the text into `clj-token's and from there do the
transformations you would like to the tokens:
(with-tokenizer
(-> "日本" clj-tokenize first)) =>
; {:surfaceForm "日本"
; :known true
; :reading "ニッポン"
; :allFeatures "名詞,固有名詞,地域,国,*,*,日本,ニッポン,ニッポン"
; :baseForm "日本"
; :partOfSpeech "名詞,固有名詞,地域,国"
; :position 0
; :allFeaturesArray ["名詞" "固有名詞" "地域" "国" "*" "*" "日本" "ニッポン" "ニッポン"]
; :class org.atilika.kuromoji.Token
; :user false
; :unknown false}- Christian Moen continuously provided unvaluable information, including the Part of Speech translations to English from Lucene.
- Kuromoji example in Clojure by bouzuya
Copyright (C) 2017 Carlos C. Fontes.
Double licensed under the Eclipse Public License (the same as Clojure) or the Apache Public License 2.0.