Namespace Lucene.Net.Analysis.Cjk

Analyzer for Chinese, Japanese, and Korean, which indexes bigrams. This analyzer generates bigram terms, which are overlapping groups of two adjacent Han, Hiragana, Katakana, or Hangul characters.

Three analyzers are provided for Chinese, each of which treats Chinese text in a different way.

  • ChineseAnalyzer (in the Lucene.Net.Analysis.Cn namespace): Index unigrams (individual Chinese characters) as a token.
  • CJKAnalyzer (in the Lucene.Net.Analysis.Cjk namespace): Index bigrams (overlapping groups of two adjacent Chinese characters) as tokens.
  • SmartChineseAnalyzer (in the Lucene.Net.Analysis.SmartCn package): Index words (attempt to segment Chinese text into words) as tokens.

Example phrase： "我是中国人"

  1. ChineseAnalyzer: 我－是－中－国－人
  2. CJKAnalyzer: 我是－是中－中国－国人
  3. SmartChineseAnalyzer: 我－是－中国－人

Classes

CJKAnalyzer

An Lucene.Net.Analysis.Analyzer that tokenizes text with StandardTokenizer, normalizes content with CJKWidthFilter, folds case with LowerCaseFilter, forms bigrams of CJK with CJKBigramFilter, and filters stopwords with StopFilter

CJKBigramFilter

Forms bigrams of CJK terms that are generated from StandardTokenizer or ICUTokenizer.

CJK types are set by these tokenizers, but you can also use CJKBigramFilter(TokenStream, CJKScript) to explicitly control which of the CJK scripts are turned into bigrams.

By default, when a CJK character has no adjacent characters to form a bigram, it is output in unigram form. If you want to always output both unigrams and bigrams, set the 

outputUnigrams
flag in CJKBigramFilter(TokenStream, CJKScript, Boolean). This can be used for a combined unigram+bigram approach.

In all cases, all non-CJK input is passed thru unmodified.

CJKBigramFilterFactory

Factory for CJKBigramFilter.

<fieldType name="text_cjk" class="solr.TextField">
  <analyzer>
    <tokenizer class="solr.StandardTokenizerFactory"/>
    <filter class="solr.CJKWidthFilterFactory"/>
    <filter class="solr.LowerCaseFilterFactory"/>
    <filter class="solr.CJKBigramFilterFactory" 
      han="true" hiragana="true" 
      katakana="true" hangul="true" outputUnigrams="false" />
  </analyzer>
</fieldType>

CJKTokenizer

CJKTokenizer is designed for Chinese, Japanese, and Korean languages.


The tokens returned are every two adjacent characters with overlap match.

Example: "java C1C2C3C4" will be segmented to: "java" "C1C2" "C2C3" "C3C4".

Additionally, the following is applied to Latin text (such as English):
  • Text is converted to lowercase.
  • Numeric digits, '+', '#', and '_' are tokenized as letters.
  • Full-width forms are converted to half-width forms.
For more info on Asian language (Chinese, Japanese, and Korean) text segmentation: please search google

CJKTokenizerFactory

Factory for CJKTokenizer. 

<fieldType name="text_cjk" class="solr.TextField" positionIncrementGap="100">
  <analyzer>
    <tokenizer class="solr.CJKTokenizerFactory"/>
  </analyzer>
</fieldType>

CJKWidthFilter

A Lucene.Net.Analysis.TokenFilter that normalizes CJK width differences:

  • Folds fullwidth ASCII variants into the equivalent basic latin
  • Folds halfwidth Katakana variants into the equivalent kana

NOTE: this filter can be viewed as a (practical) subset of NFKC/NFKD Unicode normalization. See the normalization support in the ICU package for full normalization.

CJKWidthFilterFactory

Factory for CJKWidthFilter.

<fieldType name="text_cjk" class="solr.TextField">
  <analyzer>
    <tokenizer class="solr.StandardTokenizerFactory"/>
    <filter class="solr.CJKWidthFilterFactory"/>
    <filter class="solr.LowerCaseFilterFactory"/>
    <filter class="solr.CJKBigramFilterFactory"/>
  </analyzer>
</fieldType>

Enums

CJKScript