elasticsearch使用JAVA API从Elastic Search建议搜索响应中提取源数据

Question

我在Java中使用CompletionSuggester来自动完成应用程序，我能够使用JavaAPI从搜索响应中提取建议文本。在检查原始响应时，我看到建议响应包含源数据（完整文档，而不仅仅是建议字符串）。如何从建议搜索响应中提取源数据

请在下面找到我用来获取建议文本的代码-

SearchRequest searchRequest = new SearchRequest("my_entitiy");
CompletionSuggestionBuilder suggestionBuilder = new CompletionSuggestionBuilder("nameSuggest");
suggestionBuilder.size(10).prefix(input).skipDuplicates(true);

SearchSourceBuilder searchSourceBuilder = new SearchSourceBuilder();
searchSourceBuilder.suggest(
        new SuggestBuilder().addSuggestion(SUGGESTION_NAME, suggestionBuilder));
searchRequest.source(searchSourceBuilder);
SearchResponse searchResponse = elasticClient.search(searchRequest, RequestOptions.DEFAULT);

Suggest suggest = searchResponse.getSuggest();
Suggest.Suggestion<Suggest.Suggestion.Entry<Suggest.Suggestion.Entry.Option>> suggesition =
        suggest.getSuggestion(SUGGESTION_NAME);
List<String> suggestionList =  new ArrayList<>();
for (Suggest.Suggestion.Entry<Suggest.Suggestion.Entry.Option> entry : suggesition.getEntries()) {
  for(Suggest.Suggestion.Entry.Option option:entry.getOptions()){
    suggestionList.add(option.getText().toString());
  }
}

在选项中，有几种方法可用于提取分数、文本和突出显示的内容。是否可以从选项中获取_源数据？我看到一个toXContent函数，可以使用它来获取源数据吗

上面的代码片段正在将建议的字符串保存到一个列表中，我想知道是否有可能获得完整的docjson

Answer 1

我通过将Suggest.Suggestion.Entry.Option转换为它的实现CompletionSuggestion.Entry.Option实现了这个技巧，它公开了SearchHit。下面是将源映射到对象的步骤

CompletionSuggestion.Entry.Option completionSuggestionEntryOption =
              (CompletionSuggestion.Entry.Option) option;
          SearchHit searchHit = completionSuggestionEntryOption.getHit();
          String hitJson = searchHit.getSourceAsString();
          ObjectMapper objectMapper = new ObjectMapper();
          T object = objectMapper.readValue(hitJson, clazz);

clazz是要映射到的类

Answer 2

您可以使用JHLC（java高级rest客户端）的fetch source API，它是Elasticsearch的官方java客户端

根据链接

This API helps to get only the _source field of a document.

您可以从之前的响应中获取id，然后使用获取这些文档的来源

GetSourceRequest getSourceRequest = new GetSourceRequest(
    "posts", 
    "1"); 

在Elastic and的样本上方，其中posts是索引名，1是文档id

Answer 3

您始终可以使用源筛选来筛选搜索结果中要返回的字段。在elastic中，可以在_source上下文中添加includes、excludes或两者。例如，您只想获取field1和field2，您可以将_source与查询一起设置如下：

{
  "query":{
    // your query goes here
  },
  "_source":{
    "includes":["field1", "field2"]
  }
}

使用高级rest客户端，可以实现以下相同功能：

String[] includes = {"field1", "field2"};
searchSourceBuilder.fetchSource(new FetchSourceContext(true, includes, null));