`
m635674608
  • 浏览: 4929721 次
  • 性别: Icon_minigender_1
  • 来自: 南京
社区版块
存档分类
最新评论

ElasticSearch安装ik分词插件

 
阅读更多

对于索引可能最关系的就是分词了 一般对于es 来说默认的smartcn  但效果不是很好 

一个是ik的,一个是mmseg的,下面分别介绍下两者的用法,其实都差不多的,先安装插件,命令行:

安装ik插件

plugin -install medcl/elasticsearch-analysis-ik/1.1.0

下载ik相关配置词典文件到config目录

cd config

wget http://github.com/downloads/medcl/elasticsearch-analysis-ik/ik.zip --no-check-certificate

unzip ik.zip

rm ik.zip

分词配置

ik分词配置,在elasticsearch.yml文件中加上

1
2
3
4
5
6
index:
  analysis:                   
    analyzer:      
      ik:
          alias: [ik_analyzer]
          type: org.elasticsearch.index.analysis.IkAnalyzerProvider

1
index.analysis.analyzer.ik.type : “ik”

安装mmseg插件:

bin/plugin -install medcl/elasticsearch-analysis-mmseg/1.1.0

下载相关配置词典文件到config目录

cd config

wget http://github.com/downloads/medcl/elasticsearch-analysis-mmseg/mmseg.zip --no-check-certificate

unzip mmseg.zip

rm mmseg.zip

mmseg分词配置,也是在在elasticsearch.yml文件中

1
2
3
4
5
6
index:
  analysis:
    analyzer:
      mmseg:
          alias: [news_analyzer, mmseg_analyzer]
          type: org.elasticsearch.index.analysis.MMsegAnalyzerProvider

1
index.analysis.analyzer.default.type : "mmseg"

mmseg分词还有些更加个性化的参数设置如下

1
2
3
4
5
6
7
8
9
10
11
12
index:
  analysis:
    tokenizer:
      mmseg_maxword:
          type: mmseg
          seg_type: "max_word"
      mmseg_complex:
          type: mmseg
          seg_type: "complex"
      mmseg_simple:
          type: mmseg
          seg_type: "simple"

这样配置完后插件安装完成,启动es就会加载插件。

定义mapping

在添加索引的mapping时就可以这样定义分词器

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
{
   "page":{
      "properties":{
         "title":{
            "type":"string",
            "indexAnalyzer":"ik",
            "searchAnalyzer":"ik"
         },
         "content":{
            "type":"string",
            "indexAnalyzer":"ik",
            "searchAnalyzer":"ik"
         }
      }
   }
}

indexAnalyzer为索引时使用的分词器,searchAnalyzer为搜索时使用的分词器。

java mapping代码如下:

 

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
XContentBuilder content = XContentFactory.jsonBuilder().startObject()
        .startObject("page")
          .startObject("properties")       
            .startObject("title")
              .field("type""string")           
              .field("indexAnalyzer""ik")
              .field("searchAnalyzer""ik")
            .endObject() 
            .startObject("code")
              .field("type""string")         
              .field("indexAnalyzer""ik")
              .field("searchAnalyzer""ik")
            .endObject()     
          .endObject()
         .endObject()
       .endObject()

 

测试分词可用调用下面api,注意indexname为索引名,随便指定一个索引就行了

http://localhost:9200/indexname/_analyze?analyzer=ik&text=测试elasticsearch分词器

附:

ik分词插件项目地址:https://github.com/medcl/elasticsearch-analysis-ik

mmseg分词插件项目地址:https://github.com/medcl/elasticsearch-analysis-mmseg

 

配置好的es版本,地址如下:https://github.com/medcl/elasticsearch-rtf

 

一、IK简介

    IK Analyzer是一个开源的,基于java语言开发的轻量级的中文分词工具包。从2006年12月推出1.0版开始, IKAnalyzer已经推出了4个大版本。最初,它是以开源项目Luence为应用主体的,结合词典分词和文法分析算法的中文分词组件。从3.0版本开 始,IK发展为面向Java的公用分词组件,独立于Lucene项目,同时提供了对Lucene的默认优化实现。在2012版本中,IK实现了简单的分词 歧义排除算法,标志着IK分词器从单纯的词典分词向模拟语义分词衍化。  

    IK Analyzer 2012特性:

    1.采用了特有的“正向迭代最细粒度切分算法“,支持细粒度和智能分词两种切分模式;

    2.在系统环境:Core2 i7 3.4G双核,4G内存,window 7 64位, Sun JDK 1.6_29 64位 普通pc环境测试,IK2012具有160万字/秒(3000KB/S)的高速处理能力。

    3.2012版本的智能分词模式支持简单的分词排歧义处理和数量词合并输出。

    4.采用了多子处理器分析模式,支持:英文字母、数字、中文词汇等分词处理,兼容韩文、日文字符

    5.优化的词典存储,更小的内存占用。支持用户词典扩展定义。特别的,在2012版本,词典支持中文,英文,数字混合词语。

 

二、安装IK分词插件

    假设读者已经安装好ES,如果没有的话,请参考ElasticSearch入门 —— 集群搭建。安装IK分词需要的资源可以从这里下载,整个安装过程需要三个步骤:

    1、获取分词的依赖包

    通过git clone https://github.com/medcl/elasticsearch-analysis-ik,下载分词器源码,然后进入下载目录,执行命令:mvn clean package,打包生成elasticsearch-analysis-ik-1.2.5.jar。将这个jar拷贝到ES_HOME/plugins/analysis-ik目录下面,如果没有该目录,则先创建该目录。

    2、ik目录拷贝

    将下载目录中的ik目录拷贝到ES_HOME/config目录下面。

    3、分词器配置

    打开ES_HOME/config/elasticsearch.yml文件,在文件最后加入如下内容:

index:
  analysis:                   
    analyzer:      
      ik:
          alias: [ik_analyzer]
          type: org.elasticsearch.index.analysis.IkAnalyzerProvider
      ik_max_word:
          type: ik
          use_smart: false
      ik_smart:
          type: ik
          use_smart: true
index.analysis.analyzer.default.type: ik

 

    ok!插件安装已经完成,请重新启动ES,接下来测试ik分词效果啦!

三、ik分词测试

    1、创建一个索引,名为index。   

curl -XPUT http://localhost:9200/index

 

    2、为索引index创建mapping。

curl -XPOST http://localhost:9200/index/fulltext/_mapping -d'
{
    "fulltext": {
             "_all": {
            "analyzer": "ik"
        },
        "properties": {
            "content": {
                "type" : "string",
                "boost" : 8.0,
                "term_vector" : "with_positions_offsets",
                "analyzer" : "ik",
                "include_in_all" : true
            }
        }
    }
}'

 

    3、测试

curl 'http://localhost:9200/index/_analyze?analyzer=ik&pretty=true' -d '
{
"text":"世界如此之大"
}'

 

    显示结果如下:

{
  "tokens" : [ {
    "token" : "text",
    "start_offset" : 4,
    "end_offset" : 8,
    "type" : "ENGLISH",
    "position" : 1
  }, {
    "token" : "世界",
    "start_offset" : 11,
    "end_offset" : 13,
    "type" : "CN_WORD",
    "position" : 2
  }, {
    "token" : "如此",
    "start_offset" : 13,
    "end_offset" : 15,
    "type" : "CN_WORD",
    "position" : 3
  }, {
    "token" : "之大",
    "start_offset" : 15,
    "end_offset" : 17,
    "type" : "CN_WORD",
    "position" : 4
  } ]
}

 

 

   总结:1.7.1 es 配置插件

    1、复制ik文件夹到conf目录,

    2、修改elasticsearch.yml。增加 index.analysis.analyzer.ik.type : "ik"

    3.  plugins 目录新建analysis-ik文件夹,把ik jar包放到这个文件夹里面

参考资料

    https://github.com/medcl/elasticsearch-analysis-ik

 

欢迎转载,请注明出处:http://my.oschina.net/xiaohui249/blog/232784

 

http://blog.csdn.net/liuzhenfeng/article/details/39404435

 

http://my.oschina.net/sunzy/blog/195341

分享到:
评论

相关推荐

Global site tag (gtag.js) - Google Analytics