es自动完成功能升级

 

自动完成，功能不能使用Suggester的phrase和term，英文或者拼音纠错可以，但是中文表现不行

 

 

es中的 completion Suggester 提供了自动完成功能

 

是直接通过索引内容和FST一起存放在  .tip文件中；去加载到目标内存中，所以响应的速度更快；FST非常适用于前缀查找。

如果是phrase或者term 则需要分解为token后，去字典里进行查找相对效率会稍差些。

 

但我们需要指定类型为： type: completion   在查询过程中使用suggest查询，来搞定“自动完成”功能，但只能是前缀搜索。 如果需要限制类别，可以使用添加分类字段来指定类别，但是input集合中分类词难维护，不建议使用。

 

如果没有结果召回，我们可以使用分词获取token后，去匹配类型，然后根据类型搜索，按照热度，倒排序，然后执行返回top20。

 

POST _analyze
{
  "text": [
    "Lucene is cool",
    "Elasticsearch builds on top of lucene",
    "Elasticsearch rocks",
    "Elastic is the company behind ELK stack",
    "elk rocks",
    "elasticsearch is rock solid"
  ]
}


PUT /blogs_completion/
{
  "mappings": {
    "tech": {
      "properties": {
        "body": {
          "type": "completion"
        },
        "body1": {
          "type": "text",
          "analyzer":"ik_smart"
        }
      }
    }
  }
}

            
DELETE /blogs_completion/


POST _bulk/?refresh=true
{ "index" : { "_index" : "blogs_completion", "_type" : "tech" } }
{ "body": "Lucene is cool","body1": "Lucene is cool"}
{ "index" : { "_index" : "blogs_completion", "_type" : "tech" } }
{ "body": "Elasticsearch builds on top of lucene","body1":"Elasticsearch builds on top of lucene"}
{ "index" : { "_index" : "blogs_completion", "_type" : "tech" } }
{ "body": "Elasticsearch rocks","body1":"Elasticsearch rocks"}
{ "index" : { "_index" : "blogs_completion", "_type" : "tech" } }
{ "body": "Elastic is the company behind ELK stack","body1":"Elastic is the company behind ELK stack"}
{ "index" : { "_index" : "blogs_completion", "_type" : "tech" } }
{ "body": "the elk stack rocks","body1":"the elk stack rocks"}
{ "index" : { "_index" : "blogs_completion", "_type" : "tech" } }
{ "body": "elasticsearch is rock solid","body1":"elasticsearch is rock solid"}




POST blogs_completion/_search?pretty
{ 
  "size": 0,
  "suggest": {
    "blog-suggest": {
      "prefix": "Elasticsearc b",
      "completion": {
        "field": "body"
      }
    }
  }
}

POST /blogs_completion/_search?pretty
{ 
  "suggest": {
    "blog-suggest": {
      "text": "biuds",
      "term": {
        "suggest_mode": "missing",
        "field": "body1"
      }
    }
  }
}

POST blogs_completion/_search
{
  "size": 0,
  "suggest": {
    "blog-suggest": {
      "text": "Elastcserch rock",
      "phrase": {
        "field": "body1"
      }
    }
  }
}

 

 

 

使用es搞定自动完成功能，使用es提供的suggested方式，suggested支持三种匹配模式：
index设置mapping时：检索精准度 completion>phrase>term

completion模式需要设置对应字段type为：completion
phrase模式和term模式需要设置对应字段type为：text

completion直接返回的option数组结果中是根据左前缀匹配出来的；
phrase在涉及的文档中会做词组的匹配；
term会针对单个词的纠错匹配；（实现方式为Levenstein edit distance，在一定范围内移动字符能匹配就可以作为结果返回）
结果召回率上：completion<phrase<term

 

所以在自动完成功能中要有完整的方案，如果没有匹配项，应该使用term分词后的纠错匹配来增加数据召回率。