ElasticSearch使用completion实现自动补全功能

文章目录

文章目录

最终效果

completion使用

Java实现搜索补全代码

热门搜索推荐

最终效果

实现的最终效果如下图百度搜索相似，输入词的时候返回提示。同时输入拼音也会有相同的提示效果。

completion使用

安装ik中文分词器：https://github.com/medcl/elasticsearch-analysis-ik

安装拼音分词器：https://github.com/medcl/elasticsearch-analysis-pinyin

定义关键词索引并自定义ik+pinyin分词器，要完成补全搜索，必须要用到特殊的数据类型completion，要汉字拼音都能补全，必须要使用自定义的ik+pinyin分词器。

1.  
   PUT suggest
2.  
   {
3.  
   "settings": {
4.  
   "number_of_replicas": 0,
5.  
   "number_of_shards": 1,
6.  
   "analysis": {
7.  
   "analyzer": {
8.  
   "ik_pinyin_analyzer": {
9.  
   "type": "custom",
10.  
    "tokenizer": "ik_max_word",
11.  
    "filter": ["my_pinyin", "word_delimiter"]
12.  
    }
13.  
    },
14.  
    "filter": {
15.  
    "my_pinyin": {
16.  
    "type": "pinyin",
17.  
    "first letter": "prefix",
18.  
    "padding_char": " "
19.  
    }
20.  
    }
21.  
    }
22.  
    },
23.  
    "mappings": {
24.  
    "suggest": {
25.  
    "properties": {
26.  
    "keyword": {
27.  
    "type": "completion",
28.  
    "analyzer": "ik_pinyin_analyzer",
29.  
    "fields": {
30.  
    "key": {
31.  
    "type": "keyword"
32.  
    }
33.  
    }
34.  
    },
35.  
    "id": {
36.  
    "type": "keyword"
37.  
    },
38.  
    "createDate": {
39.  
    "type": "date",
40.  
    "format": "yyyy-MM-dd HH:mm:ss"
41.  
    }
42.  
    }
43.  
    }
44.  
    }
45.  
    }

初始化部分数据

1.  
   POST _bulk/?refresh=true
2.  
   { "index": { "_index": "suggest", "_type": "suggest" }}
3.  
   { "keyword": "项目"}
4.  
   { "index": { "_index": "suggest", "_type": "suggest" }}
5.  
   { "keyword": "项目进度"}
6.  
   { "index": { "_index": "suggest", "_type": "suggest" }}
7.  
   { "keyword": "项目管理"}
8.  
   { "index": { "_index": "suggest", "_type": "suggest" }}
9.  
   { "keyword": "项目进度及调整 汇总.doc_文档"}
10.  
    { "index": { "_index": "suggest", "_type": "suggest" }}
11.  
    { "keyword": "项目"}

使用suggest获取搜索补全建议，并对同一词语去重。 

1.  
   GET /suggest/_search
2.  
   {
3.  
   "suggest": {
4.  
   "my-suggest": {
5.  
   "prefix": "项目",
6.  
   "completion": {
7.  
   "field": "keyword",
8.  
   "size": 20,
9.  
   "skip_duplicates": true
10.  
    }
11.  
    }
12.  
    }
13.  
    }

Java实现搜索补全代码

1.  
   /**
2.  
   * 获取相关搜索，最多返回9条
3.  
   * @param key
4.  
   * @return
5.  
   */
6.  
   public JSONObject getSearchSuggest(String key) {
7.  
   CompletionSuggestionBuilder suggestion = SuggestBuilders
8.  
   .completionSuggestion("keyword").prefix(key).size(20).skipDuplicates(true);
9.  
   SuggestBuilder suggestBuilder = new SuggestBuilder();
10.  
    suggestBuilder.addSuggestion("suggest", suggestion);
11.  
    SearchResponse response = template.suggest(suggestBuilder, EsConstants.SUGGEST);
12.  
    Suggest suggest = response.getSuggest();
13.  
     
14.  
    Set<String> keywords = null;
15.  
    if (suggest != null) {
16.  
    keywords = new HashSet<>();
17.  
    List<? extends Suggest.Suggestion.Entry<? extends Suggest.Suggestion.Entry.Option>> entries = suggest.getSuggestion("suggest").getEntries();
18.  
     
19.  
    for (Suggest.Suggestion.Entry<? extends Suggest.Suggestion.Entry.Option> entry: entries) {
20.  
    for (Suggest.Suggestion.Entry.Option option: entry.getOptions()) {
21.  
    /** 最多返回9个推荐，每个长度最大为20 */
22.  
    String keyword = option.getText().string();
23.  
    if (!StringUtils.isEmpty(keyword) && keyword.length() <= 20) {
24.  
    /** 去除输入字段 */
25.  
    if (keyword.equals(key)) continue;
26.  
    keywords.add(keyword);
27.  
    if (keywords.size() >= 9) {
28.  
    break;
29.  
    }
30.  
    }
31.  
    }
32.  
    }
33.  
    }
34.  
    return ApiResult.OK(keywords, "获取推荐词组成功");
35.  
    }
36.  
     

热门搜索推荐

上面自定义的索引中，通过fields属性专门存储了类型为keyword的字段keyword.key。可以通过统计keyword.key来获取搜索次数最多的句子。下面是java实现方式

1.  
   public JSONObject searchHot(Map<String, Object> map) {
2.  
   Integer size = 10;
3.  
   if (!StringUtils.isEmpty(map.get("size"))) {
4.  
   size = (Integer)map.get("size");
5.  
   }
6.  
    
7.  
   /** 获取最近一个月时间 */
8.  
   String preMonth = LocalDateTime.now().minusMonths(1).format(EsConstants.fomatter);
9.  
   String now = LocalDateTime.now().format(EsConstants.fomatter);
10.  
     
11.  
    /** 统计最近一个月的热门搜索，长度最大10，方便显示 */
12.  
    SearchRequestBuilder requestBuilder = template.getClient().prepareSearch(EsConstants.SUGGEST)
13.  
    .setQuery(QueryBuilders.rangeQuery("createDate").get(preMonth).lte(now));
14.  
     
15.  
    SearchResponse searchResponse = requestBuilder.addAggregation(AggregationBuilders
16.  
    .terms("hotSearch").field("keyword.key").size(size)).execute().actionGet();
17.  
     
18.  
    Aggregations aggregations = searchResponse.getAggregations();
19.  
    Set<String> keywords = null;
20.  
    if (aggregations != null) {
21.  
    keywords = new HashSet<>();
22.  
    Terms hotSearch = aggregations.get("hotSearch");
23.  
    List<? extends Terms.Bucket> buckets = hotSearch.getBuckets();
24.  
    for (Terms.Bucket bucket: buckets) {
25.  
    if (bucket.getKey().toString().length() <= 10) {
26.  
    keywords.add((String)bucket.getKey());
27.  
    }
28.  
    }
29.  
    }
30.  
    return ApiResult.OK(keywords, "热门搜索获取成功");
31.  
    }