Elasticsearch7.15.2 ik中文分词器 定制化分词器之扩展词库

Elasticsearch7.15.2 ik中文分词器 定制化分词器之扩展词库
Elasticsearch7.15.2 ik中文分词器 定制化分词器之扩展词库

背景: IK分词提供的两个分词器,并不支持一些新的词汇,有时候也不能满足实际业务需要,这时候,我们可以定义自定义词库来完成目标。
目标: 定制化中文分词器,使得我们的中文分词器支持扩展的词汇

文章目录

一、搜索现状
1. 搜索关键词
# 搜索凯悦相关的酒店
GET /shop/_search
{
  "query":{
    "match": {"name":"凯悦"}
  }
}
2. 数据结果
{
  "took" : 7,
  "timed_out" : false,
  "_shards" : {
    "total" : 1,
    "successful" : 1,
    "skipped" : 0,
    "failed" : 0
  },
  "hits" : {
    "total" : {
      "value" : 5,
      "relation" : "eq"
    },
    "max_score" : 3.3362136,
    "hits" : [
      {
        "_index" : "shop",
        "_type" : "_doc",
        "_id" : "9",
        "_score" : 3.3362136,
        "_source" : {
          "price_per_man" : 176,
          "remark_score" : 2.2,
          "category_name" : "酒店",
          "@version" : "1",
          "seller_disabled_flag" : 0,
          "@timestamp" : "2021-11-21T04:10:03.916Z",
          "tags" : "落地大窗",
          "location" : "31.306172,121.525843",
          "seller_remark_score" : 3.0,
          "id" : 9,
          "name" : "凯悦酒店",
          "seller_id" : 17,
          "category_id" : 2
        }
      },
      {
        "_index" : "shop",
        "_type" : "_doc",
        "_id" : "10",
        "_score" : 2.836244,
        "_source" : {
          "price_per_man" : 182,
          "remark_score" : 0.5,
          "category_name" : "酒店",
          "@version" : "1",
          "seller_disabled_flag" : 0,
          "@timestamp" : "2021-11-21T04:10:03.918Z",
          "tags" : "自助餐",
          "location" : "31.196742,121.322846",
          "seller_remark_score" : 3.0,
          "id" : 10,
          "name" : "凯悦嘉轩酒店",
          "seller_id" : 17,
          "category_id" : 2
        }
      },
      {
        "_index" : "shop",
        "_type" : "_doc",
        "_id" : "11",
        "_score" : 2.836244,
        "_source" : {
          "price_per_man" : 74,
          "remark_score" : 1.0,
          "category_name" : "酒店",
          "@version" : "1",
          "seller_disabled_flag" : 0,
          "@timestamp" : "2021-11-21T04:10:03.920Z",
          "tags" : "自助餐",
          "location" : "31.156899,121.238362",
          "seller_remark_score" : 3.0,
          "id" : 11,
          "name" : "新虹桥凯悦酒店",
          "seller_id" : 17,
          "category_id" : 2
        }
      },
      {
        "_index" : "shop",
        "_type" : "_doc",
        "_id" : "12",
        "_score" : 2.638537,
        "_source" : {
          "price_per_man" : 71,
          "remark_score" : 2.0,
          "category_name" : "美食2",
          "@version" : "1",
          "seller_disabled_flag" : 0,
          "@timestamp" : "2021-11-21T04:10:03.923Z",
          "tags" : "有包厢",
          "location" : "30.679819,121.651921",
          "seller_remark_score" : 3.0,
          "id" : 12,
          "name" : "凯悦咖啡(新建西路店)",
          "seller_id" : 17,
          "category_id" : 1
        }
      },
      {
        "_index" : "shop",
        "_type" : "_doc",
        "_id" : "4",
        "_score" : 1.3119392,
        "_source" : {
          "price_per_man" : 152,
          "remark_score" : 2.0,
          "category_name" : "美食2",
          "@version" : "1",
          "seller_disabled_flag" : 0,
          "@timestamp" : "2021-11-21T04:10:03.907Z",
          "tags" : "落地大窗 有WIFI",
          "location" : "31.306419,121.524878",
          "seller_remark_score" : 2.0,
          "id" : 4,
          "name" : "花悦庭果木烤鸭",
          "seller_id" : 2,
          "category_id" : 1
        }
      }
    ]
  }
}
3. 数据分析

上面数据中有一条不符的结果数据,此数据中无**“凯悦”**关键词,但是,搜索后还是显示在页面上,不符合预期搜索结果。

 {
        "_index" : "shop",
        "_type" : "_doc",
        "_id" : "4",
        "_score" : 1.3119392,
        "_source" : {
          "price_per_man" : 152,
          "remark_score" : 2.0,
          "category_name" : "美食2",
          "@version" : "1",
          "seller_disabled_flag" : 0,
          "@timestamp" : "2021-11-21T04:10:03.907Z",
          "tags" : "落地大窗 有WIFI",
          "location" : "31.306419,121.524878",
          "seller_remark_score" : 2.0,
          "id" : 4,
          "name" : "花悦庭果木烤鸭",
          "seller_id" : 2,
          "category_id" : 1
        }
      }
4. ES IK分词
# 查阅凯悦分词
GET /shop/_analyze
{
  "analyzer": "ik_smart",
  "text": "凯悦"
}
5. IK分词结果+分析
{
  "tokens" : [
    {
      "token" : "凯",
      "start_offset" : 0,
      "end_offset" : 1,
      "type" : "CN_CHAR",
      "position" : 0
    },
    {
      "token" : "悦",
      "start_offset" : 1,
      "end_offset" : 2,
      "type" : "CN_CHAR",
      "position" : 1
    }
  ]
}

从上面数据可以看出,使用ik_smart分词api,分词“凯”,“悦”,并没有将“凯悦”关键词当做一个分词元素,主要原因就是,es安装的ik中文分词库中没有将“凯悦”放入分词库。

二、定制化分词器
2.1. 新增分词词典库
cd /app/elasticsearch-7.15.2/config/analysis-ik/
vim new_word.dic

添加自定义分词

凯悦
2.2. 词典配置

使用ik加载我们自定义的分词词典库

vim IKAnalyzer.cfg.xml

内容:

<?xml version="1.0" encoding="UTF-8"?>
<!DOCTYPE properties SYSTEM "http://java.sun.com/dtd/properties.dtd">
<properties>
        <comment>IK Analyzer 扩展配置</comment>
        <!--用户可以在这里配置自己的扩展字典 -->
        <entry key="ext_dict">new_word.dic</entry>
         <!--用户可以在这里配置自己的扩展停止词字典-->
        <entry key="ext_stopwords"></entry>
        <!--用户可以在这里配置远程扩展字典 -->
        <!-- <entry key="remote_ext_dict">words_location</entry> -->
        <!--用户可以在这里配置远程扩展停止词字典-->
        <!-- <entry key="remote_ext_stopwords">words_location</entry> -->
</properties>

Elasticsearch7.15.2 ik中文分词器 定制化分词器之扩展词库

2.3. 重启es7
ps -ef|grep elasticsearch
kill -9 es进程号
cd /app/elasticsearch-7.15.2/
bin/elasticsearch -d

Elasticsearch7.15.2 ik中文分词器 定制化分词器之扩展词库

2.4. 重新查看分词结果
# 查阅凯悦分词
GET /shop/_analyze
{
  "analyzer": "ik_smart",
  "text": "凯悦"
}


GET /shop/_analyze
{
  "analyzer": "ik_max_word",
  "text": "凯悦"
}

Elasticsearch7.15.2 ik中文分词器 定制化分词器之扩展词库

2.5. 重新搜索
GET /shop/_search
{
  "query":{
    "match": {"name":"凯悦"}
  }
}

Elasticsearch7.15.2 ik中文分词器 定制化分词器之扩展词库

GET /shop/_search

Elasticsearch7.15.2 ik中文分词器 定制化分词器之扩展词库

发现一条数据都没查询出来,但是,数据都还在。

2.6. 重建分词索引

索引创建的时候,是在ik分词器上当时没有“凯悦”这个词的时候。目前,我们凯悦酒店这条记录对应的记录”凯和悦”已经在索引成型,单字的”凯”和单字“悦”。因为在擦黄建索引的时候,并没有做分词的扩展分词库加载。

目前的问题,现在索引中存储的是”凯和悦”分开的,但是我搜索的时候,执行的凯悦,却是按照搜索当前search的分词器,也就是分出来的是”凯和悦”连字存在。

我搜索是2个字,但是倒排索引的时候是按照单字做搜引得,因此导致搜素数据为空。

解决方案:

第一种(第一次):把索引全部删除,然后全量同步分词索引
第二种(推荐):针对搜索的索引中包含“凯“或者“悦“的索引执行重建索引,其他的索引不重建索引。

# 重建凯悦分析索引
POST /shop/_update_by_query
{
  "query": {
    "bool": {
      "must": [
        {"term":{"name":"凯"}},
          {"term":{"name":"悦"}}
      ]
    }
  }
}
2.7. 再次查询
GET /shop/_search
{
  "query":{
    "match": {"name":"凯悦"}
  }
}
2.8. 数据分析
{
  "took" : 3,
  "timed_out" : false,
  "_shards" : {
    "total" : 1,
    "successful" : 1,
    "skipped" : 0,
    "failed" : 0
  },
  "hits" : {
    "total" : {
      "value" : 4,
      "relation" : "eq"
    },
    "max_score" : 2.0709352,
    "hits" : [
      {
        "_index" : "shop",
        "_type" : "_doc",
        "_id" : "9",
        "_score" : 2.0709352,
        "_source" : {
          "price_per_man" : 176,
          "remark_score" : 2.2,
          "category_name" : "酒店",
          "@version" : "1",
          "seller_disabled_flag" : 0,
          "@timestamp" : "2021-11-21T04:10:03.916Z",
          "tags" : "落地大窗",
          "location" : "31.306172,121.525843",
          "seller_remark_score" : 3.0,
          "id" : 9,
          "name" : "凯悦酒店",
          "seller_id" : 17,
          "category_id" : 2
        }
      },
      {
        "_index" : "shop",
        "_type" : "_doc",
        "_id" : "10",
        "_score" : 1.7177677,
        "_source" : {
          "price_per_man" : 182,
          "remark_score" : 0.5,
          "category_name" : "酒店",
          "@version" : "1",
          "seller_disabled_flag" : 0,
          "@timestamp" : "2021-11-21T04:10:03.918Z",
          "tags" : "自助餐",
          "location" : "31.196742,121.322846",
          "seller_remark_score" : 3.0,
          "id" : 10,
          "name" : "凯悦嘉轩酒店",
          "seller_id" : 17,
          "category_id" : 2
        }
      },
      {
        "_index" : "shop",
        "_type" : "_doc",
        "_id" : "11",
        "_score" : 1.7177677,
        "_source" : {
          "price_per_man" : 74,
          "remark_score" : 1.0,
          "category_name" : "酒店",
          "@version" : "1",
          "seller_disabled_flag" : 0,
          "@timestamp" : "2021-11-21T04:10:03.920Z",
          "tags" : "自助餐",
          "location" : "31.156899,121.238362",
          "seller_remark_score" : 3.0,
          "id" : 11,
          "name" : "新虹桥凯悦酒店",
          "seller_id" : 17,
          "category_id" : 2
        }
      },
      {
        "_index" : "shop",
        "_type" : "_doc",
        "_id" : "12",
        "_score" : 1.5828056,
        "_source" : {
          "price_per_man" : 71,
          "remark_score" : 2.0,
          "category_name" : "美食2",
          "@version" : "1",
          "seller_disabled_flag" : 0,
          "@timestamp" : "2021-11-21T04:10:03.923Z",
          "tags" : "有包厢",
          "location" : "30.679819,121.651921",
          "seller_remark_score" : 3.0,
          "id" : 12,
          "name" : "凯悦咖啡(新建西路店)",
          "seller_id" : 17,
          "category_id" : 1
        }
      }
    ]
  }
}

从上面数据可以看出,符合预期结果!

上一篇:Pr 入门教程「37」了解在 Pr 中合成图像的重要概念和功能


下一篇:mysql索引使用