关于结巴分词 ElasticSearch 插件:
https://github.com/huaban/elasticsearch-analysis-jieba
该插件由huaban开发。支持Elastic Search 版本<=2.3.5。
结巴分词分析器
结巴分词插件提供3个分析器:jieba_index、jieba_search和jieba_other。
- jieba_index: 用于索引分词,分词粒度较细;
- jieba_search: 用于查询分词,分词粒度较粗;
- jieba_other: 全角转半角、大写转小写、字符分词;
使用jieba_index或jieba_search分析器,可以实现基本的分词效果。
以下是最小配置示例:
{ "mappings": { "test": { "_all": { "enabled": false }, "properties": { "name": { "type": "string", "analyzer": "jieba_index", "search_analyzer": "jieba_index" } } } } }
在生产化境中,因为业务的需要,需要考虑实现以下功能:
- 支持同义词;
- 支持字符过滤器;
结巴插件提供的分析器jieba_index、jieba_search无法实现以上功能。
自定义分析器
当jieba_index、jieba_search分析器不满足生成环境的需求时,我们可以使用自定义分析器来解决以上问题。
分析器是由字符过滤器,分词器,词元过滤器组成的。
一个分词器允许包含多个字符过滤器+一个分词器+多个词元过滤器。
因业务的需求,我们需要使用映射字符过滤器来实现分词前某些字符串的替换操作。如将用户输入的c#替换为csharp,c++替换为cplus。
下面逐一介绍分析器各个组成部分。
1. 映射字符过滤器Mapping Char Filter
这个是Elastic Search内置的映射字符过滤器,位于settings –> analysis -> char_filter下:
PUT /my_index { "settings": { "analysis": { "char_filter": { "mapping_filter": { "type": "mapping", "mappings": [ "c# => csharp", "c++ => cplus" ] } } } } }
也可以通过文件载入字符映射表。
PUT /my_index { "settings": { "analysis": { "char_filter": { "mapping_filter": { "type": "mapping", "mappings_path": "mappings.txt" } } } } }
文件默认存放config目录下,即config/ mappings.txt。
2. 结巴分词词元过滤器JiebaTokenFilter
JiebaTokenFilter接受一个SegMode参数,该参数有两个可选值:Index和Search。
我们预先定义两个词元过滤器:jieba_index_filter和jieba_search_filter。
PUT /my_index { "settings": { "analysis": { "filter": { "jieba_index_filter": { "type": "jieba", "seg_mode": "index" }, "jieba_search_filter": { "type": "jieba", "seg_mode": "search" } } } } }
这两个词元过滤器将分别用于索引分析器和查询分析器。
3. stop 停用词词元过滤器
因分词词元过滤器JiebaTokenFilter并不处理停用词。因此我们在自定义分析器时,需要定义停用词词元过滤器来处理停用词。
Elastic Search提供了停用词词元过滤器,我们可以这样来定义:
PUT /my_index { "settings": { "analysis": { "filter": { "stop_filter": { "type": "stop", "stopwords": ["and", "is", "the"] } } } } }
也可以通过文件载入停用词列表
PUT /my_index { "settings": { "analysis": { "filter": { "stop_filter": { "type": "stop", "stopwords_path": "stopwords.txt" } } } } }
文件默认存放config目录下,即config/ stopwords.txt。
4. synonym 同义词词元过滤器
我们使用ElasticSearch内置同义词词元过滤器来实现同义词的功能。
PUT /my_index { "settings": { "analysis": { "filter": { "synonym_filter": { "type": "synonym", "stopwords": [ "中文,汉语,汉字" ] } } } } }
如果同义词量比较大时,推荐使用文件的方式载入同义词库。
PUT /my_index { "settings": { "analysis": { "filter": { "synonym_filter ": { "type": "synonym", "stopwords_path": "synonyms.txt" } } } } }
5. 重新定义分析器jieba_index和jieba_search
Elastic Search支持多级分词,我们使用whitespace分词作为分词器;并在词元过滤器加入定义好的Jiebie分词词元过滤器:jieba_index_filter和jieba_search_filter。
PUT /my_index { "settings": { "analysis": { "analyzer": { "jieba_index": { "char_filter": [ "mapping_filter" ], "tokenizer": "whitespace", "filter": [ "jieba_index_filter", "stop_filter", "synonym_filter" ] }, "jieba_search": { "char_filter": [ "mapping_filter" ], "tokenizer": "whitespace", "filter": [ "jieba_search_filter", "stop_filter", "synonym_filter" ] } } } } }
注意,上面分析器的命名依然使用jieba_index和jieba_search,以便覆盖结巴分词插件提供的分析器。
当存在多个同名的分析器时,Elastic Search会优先使用索引配置中定义的分析器。
这样在代码调用层面便无需再更改。
下面是完整的配置:
PUT /my_index { "settings": { "analysis": { "char_filter": { "mapping_filter": { "type": "mapping", "mappings_path": "mappings.txt" } } "filter": { "synonym_filter ": { "type": "synonym", "stopwords_path": "synonyms.txt" }, "stop_filter": { "type": "stop", "stopwords_path": "stopwords.txt" }, "jieba_index_filter": { "type": "jieba", "seg_mode": "index" }, "jieba_search_filter": { "type": "jieba", "seg_mode": "search" } } "analyzer": { "jieba_index": { "char_filter": [ "mapping_filter" ], "tokenizer": "whitespace", "filter": [ "jieba_index_filter", "stop_filter", "synonym_filter" ] }, "jieba_search": { "char_filter": [ "mapping_filter" ], "tokenizer": "whitespace", "filter": [ "jieba_search_filter", "stop_filter", "synonym_filter" ] } } } } }
参考资料:
https://www.elastic.co/guide/en/elasticsearch/reference/2.3/index.html