如何使用 Elastic Search Low Level API 构造请求进行搜索

场景：

需要在客户端构造请求调解，调用 Elastic Search 的 API 取到结果，并且能够使用 ES 的授权机制。

方案：

一、在客户端构造 Low Level API 。

二、Low Level API 直接请求 ES 的 HTTP地址，传入 ES 的用户名，密码。

三、Low Level API 取到 JSON 格式的返回结果后，结果反序列化成 Java 对象。这个步骤里面的 Java 类定义，使用前面博客里提到的方法，从 json 文件中来生成代码。

另外也需要在客户端，根据不同的条件构造出来 Json 格式的请求调解。

我们分4个步骤来看。

1). 构造 Low Level API Client

public class ElasticSearchRestClient {private static final Logger logger = LoggerFactory.getLogger(ElasticSearchRestClient.class);private RestClient restClient;public void init(String elasticsearchAddress, String userName, String password){try {// Elasticsearch集群需要basic auth验证。final CredentialsProvider credentialsProvider = new BasicCredentialsProvider();//访问用户名和密码为您创建Elasticsearch实例时设置的用户名和密码，也是Kibana控制台的登录用户名和密码。credentialsProvider.setCredentials(AuthScope.ANY, new UsernamePasswordCredentials(userName, password));// 通过builder创建rest client，配置http client的HttpClientConfigCallback。RestClientBuilder builder = RestClient.builder(new HttpHost(elasticsearchAddress, 9200)).setHttpClientConfigCallback(new RestClientBuilder.HttpClientConfigCallback() {@Overridepublic HttpAsyncClientBuilder customizeHttpClient(HttpAsyncClientBuilder httpClientBuilder) {return httpClientBuilder.setDefaultCredentialsProvider(credentialsProvider);}});// lower level clientrestClient = builder.build();} catch (Exception e) {logger.error("Failed to initialize Elastic Search client, server:{}, userName:{}", elasticsearchAddress, userName, e);}}
}

2).指定要搜索的索引发出搜索请求

            // low level api:Request request = new Request("POST", "/" + indexName +  "/_search");request.setJsonEntity(queryJson );Response response = restClient.performRequest(request);

3).取到 JSON 格式的返回结果，结果反序列化成 Java 对象

前面的文章有介绍，如何从 Json 格式的数据定义中生产 Java 类的代码。

我们以 ES HTTP 请求返回的 Json 为例，ES 返回的 Json 格式数据跟 Kibana 控制台 Dev Tools -> Console 里面发送请求，看到的返回结果格式是一样的。比如发送这个请求：

GET  media/_search
{"query": {"match_all": {}}}

能看到返回的这个结果

右边返回的 Json 内容保存为文件 elastic.response.json，内容如下：

{"took" : 113,"timed_out" : false,"_shards" : {"total" : 10,"successful" : 9,"skipped" : 0,"failed" : 0},"hits" : {"total" : 1,"max_score" : 16.761106,"hits" : [{"_index" : "media","_type" : "doc","_id" : "176","_score" : 16.761106,"_source" : {"publish_state" : "Published","media_id" : "226bb012c46e1ef4e52b788cf3312345","id" : 176,"url" : "http://www.abcd.com/20190317152813791092","owner_id" : 1438747659963495,"media_info" : "","description" : null,"env" : 1,"cover_url" : null,"gmt_create" : "2016-03-17T07:29:18.000Z","gmt_modified" : "2016-06-16T06:50:48.000Z","state" : "Normal","title" : "20160317152813791092","tags" : null,"@version" : "1","@timestamp" : "2021-04-06T14:36:17.393Z"}}]}
}

然后执行命令：

jsonschema2pojo -T json --source elastic.response.json --target java

就能够自动生成这几个类：

ElasticResponse.java, Hit.java, Hits.java, Shards.java, Source.java

生成的 ElasticResponse 类是这样的，可以稍做调整，或者重构适应实际的需求：

 @Generated("jsonschema2pojo")public class ElasticResponse {@JsonProperty("took")private Integer took;@JsonProperty("timed_out")private Boolean timedOut;@JsonProperty("_shards")private Shards shards;@JsonProperty("hits")private Hits hits;@JsonProperty("took")public Integer getTook() {return took;}@JsonProperty("took")public void setTook(Integer took) {this.took = took;}@JsonProperty("timed_out")public Boolean getTimedOut() {return timedOut;}@JsonProperty("timed_out")public void setTimedOut(Boolean timedOut) {this.timedOut = timedOut;}@JsonProperty("_shards")public Shards getShards() {return shards;}//....
}

4). 构造 JSON 格式请求条件

以上请求发起调用的完整代码:

    public ElasticResponse search(String indexName, String queryJson){try {if(StringUtils.isEmpty(queryJson)){ElasticResponse response = new ElasticResponse();JSONObject errorInfo = new JSONObject();errorInfo.put("message", "empty json");response.setError(errorInfo);response.setError(errorInfo);return response;}// low level api:Request request = new Request("POST", "/" + indexName +  "/_search");request.setJsonEntity(queryJson );Response response = restClient.performRequest(request);String jsonResponse = EntityUtils.toString(response.getEntity());ElasticResponse finalResponse = JSON.parseObject(jsonResponse, ElasticResponse.class);return finalResponse;} catch (IOException e) {logger.error("Failed to search for index:{}", indexName, e);}return null;}

那么这里 json 格式的请求条件 queryJson 如何构造，比如要查询指定 owner_id, 且 publish_state不是删除状态，id 小于某个限定值的数据，可以这样构造：

        MatchQueryBuilder ownerQuery = QueryBuilders.matchQuery("owner_id", ownerId);BoolQueryBuilder finalQuery = QueryBuilders.boolQuery();finalQuery.must(ownerQuery);// id <= #id#if(upperId>1){RangeQueryBuilder lessThanUpperIdQuery = QueryBuilders.rangeQuery("id").to(upperId, true);finalQuery.must(lessThanUpperIdQuery);}// publish_state !='Deleted'MatchQueryBuilder inDeletedQuery = QueryBuilders.matchQuery("publish_state", "Deleted");finalQuery.mustNot(inDeletedQuery);SearchSourceBuilder searchSourceBuilder = new SearchSourceBuilder();searchSourceBuilder.query(finalQuery);//ORDER BY gmt_create DESC,id DESCFieldSortBuilder createSort = SortBuilders.fieldSort("gmt_create").order(SortOrder.DESC);searchSourceBuilder.sort(createSort);FieldSortBuilder idSort = SortBuilders.fieldSort("id").order(SortOrder.DESC);searchSourceBuilder.sort(idSort);String queryJson = searchSourceBuilder.toString();

然后就能使用这个构造出来的 queryJson 来发送 ElasticSearch 的请求了。