发布了一个基于jieba分词的ElasticSearch插件

github地址: https://github.com/hongfuli/elasticsearch-analysis-jieba

基于 jiebaelasticsearch 中文分词插件。

集成到ElasticSearch

git clone git@github.com:hongfuli/elasticsearch-analysis-jieba.git
cd elasticsearch-analysis-jieba
mvn package

把release/elasticsearch-analysis-jieba-{version}.zip文件解压到 elasticsearch 的 plugins 目录下,重启elasticsearch即可。

直接使用Tokenizer分词

可直接使用 com.github.hongfuli.jieba.Tokenizer 对文本字符进行分词,方法参数完全和 jieba python 一致。

imort com.github.hongfuli.jieba.Tokenizer

Tokenizer t = new Tokenizer();
t.cut("这是一个伸手不见五指的黑夜。我叫孙悟空,我爱北京,我爱Python和C++。", false, true);

集成到Lucene

import com.github.hongfuli.jieba.lucene.JiebaAnalyzer;

Analyzer analyzer = new JiebaAnalyzer();
try(TokenStream ts = analyzer.tokenStream("field", "这是一个伸手不见五指的黑夜。我叫孙悟空,我爱北京,我爱Python和C++。")) {
      StringBuilder b = new StringBuilder();
      CharTermAttribute termAtt = ts.getAttribute(CharTermAttribute.class);
      PositionIncrementAttribute posIncAtt = ts.getAttribute(PositionIncrementAttribute.class);
      PositionLengthAttribute posLengthAtt = ts.getAttribute(PositionLengthAttribute.class);
      OffsetAttribute offsetAtt = ts.getAttribute(OffsetAttribute.class);
      assertNotNull(offsetAtt);
      ts.reset();
      int pos = -1;
      while (ts.incrementToken()) {
        pos += posIncAtt.getPositionIncrement();
        b.append(termAtt);
        b.append(" at pos=");
        b.append(pos);
        if (posLengthAtt != null) {
          b.append(" to pos=");
          b.append(pos + posLengthAtt.getPositionLength());
        }
        b.append(" offsets=");
        b.append(offsetAtt.startOffset());
        b.append('-');
        b.append(offsetAtt.endOffset());
        b.append('\n');
      }
      ts.end();
      return b.toString();
    }
posted @ 2017-06-30 15:26  海鸟  阅读(2946)  评论(0编辑  收藏  举报