ES实战系列01：基于SpringBoot和RestHighLevelClient 快速搭建博客搜索系统

方才编程_公众号同名

发布于 2020-11-13 10:47:33

1.4K0

发布于 2020-11-13 10:47:33

文章被收录于专栏：方才编程方才编程

本文目标

通过4个博客检索场景，巩固之前所学的全文搜索 Full Text Queries 和基于词项的 Term lever Queries，同时通过组合查询的Bool query 完成复杂检索，并应用相关度知识对相关性评分进行控制。

通过搭建博客搜索系统，快速掌握RestHighLevelClient的使用，可以快速应用于工作中。

本文知识导航

01 项目简介

本项目基于SpringBoot 2.3，ElasticSearch 7.7.1，同时使用es官网提供的 elasticsearch-rest-high-level-client 客户端，快速搭建一个简单的博客搜索系统。【ps：本文完整代码获取方式，见文末】

1.1 检索场景

1）case1：根据 title 、content 、tag 进行简单检索，使用rescore利用match_phrase进行相关度控制；

2）case2：利用boost参数行相关度控制，提升 tag 的权重为3，title的权重为2；

3）case3：在case2的基础上增加过滤条件：author、tag、createAt、influence

4）case4：在case3的基础上用户指定排序条件：createAt、vote、view

1.2 场景理解

类似于微信的搜一搜功能，case1和case2就相当于下图，使用相关度进行默认排序，当然微信对相关度的控制肯定更复杂的。

case3就好比可以选择文件的类型【文章、视频等】，只是我这里把过滤条件换成了 author、tag、createAt、influence。

case4就是用户自定义排序功能。

1.3 在docker中安装ES

1、在CentOS7安装Docker1）确定你是CentOS7及以上版本	cat /etc/redhat-release	2）yum安装gcc相关	yum -y install gcc	yum -y install gcc-c++
3）卸载旧版本	yum -y remove docker docker-common docker-selinux docker-engine			4）安装需要的软件包	yum install -y yum-utils device-mapper-persistent-data lvm2
5）设置stable镜像仓库	yum-config-manager --add-repo http://mirrors.aliyun.com/docker-ce/linux/centos/docker-ce.repo
6）更新yum软件包索引	yum makecache fast
7）安装DOCKER CE	yum -y install docker-ce
8）启动docker	systemctl start docker
9）测试	docker version
10）配置阿里云镜像加速
sudo mkdir -p /etc/dockersudo tee /etc/docker/daemon.json <<-'EOF'{  "registry-mirrors": ["https://dfr09p8e.mirror.aliyuncs.com"]}EOFsudo systemctl daemon-reloadsudo systemctl restart docker

2、在docker中安装ES7.7.11)拉取镜像	docker pull docker.elastic.co/elasticsearch/elasticsearch7.7.1	2）查看镜像	docker images
3）启动ES	docker run -p 9200:9200 -p 9300:9300 -e "discovery.type=single-node" --name yourEsName -d  fce8d855350b[你的镜像id]说明：-d 后台启动-p 9200:9200 将虚拟机9200端口映射到elasticsearch的9200端口（web通信默认使用9200端口）-p 9300:9300 将虚拟机9300端口映射到elasticsearch的9300端口（分布式情况下，各个节点之间通信默认使用9300端口）--name MyEs 指定一个名字（MyEs 随意指定）
4）进入ES容器 安装各种插件：docker exec -it yourEsName /bin/bash
5）直接复制下面的命令Ik插件：./bin/elasticsearch-plugin install https://github.com/medcl/elasticsearch-analysis-ik/releases/download/v7.7.1/elasticsearch-analysis-ik-7.7.1.zip
拼音插件：./bin/elasticsearch-plugin install https://github.com/medcl/elasticsearch-analysis-pinyin/releases/download/v7.7.1/elasticsearch-analysis-pinyin-7.7.1.zip
6）退出容器，重启ESctrl + P + Q 退出容器重启docker的ES镜像：docker restart a198a70e6fba【es镜像的容器id，docker ps，即可查看】
3、在docker中安装kibana1）拉取镜像	docker pull docker.elastic.co/kibana/kibana:7.7.1	或者docker pull kibana:7.7.1
2）运行kibanadocker run -d -p 5601:5601 --name kibana --link yourEsName:elasticsearch 6de54f813b39（kibana镜像id）

1.4 数据准备

# 1）创建索引
PUT /demo1_blog
{
  "settings": {
    "index": {
      "number_of_shards": 1,
      "number_of_replicas": 1
    }
  },
  "mappings": {
    "dynamic": false,
    "properties": {
      "id": {
        "type": "integer"
      },
      "author": {
        "type": "keyword"
      },
      "influence": {
        "type": "integer_range"
      },
      "title": {
        "type": "text",
        "analyzer": "ik_max_word"
      },
      "content": {
        "type": "text",
        "analyzer": "ik_smart"
      },
      "tag": {
        "type": "text",
        "analyzer": "ik_max_word",
        "fields": {
          "keyword":{
            "type":"keyword"
          }
        }
      },
      "vote": {
        "type": "integer"
      },
      "view": {
        "type": "integer"
      },
      "createAt": {
        "type": "date",
        "format": "yyyy-MM-dd HH:mm"
      }
    }
  }
}

# 2）导入数据
POST _bulk
{"index":{"_index":"demo1_blog","_id":"1"}}
{"id":1,"author":"方才兄","influence":{"gte":10,"lte":12},"title":"ElasticSearch系列01：如何系统学习ES","content":"最后附上小编的学习记录图，后续小编会持续输出ElasticSearch技术系列文章，欢迎关注，共同探讨学习。","tag":["ElasticSearch","入门学习"],"vote":10,"view":100,"createAt":"2020-04-24 10:56"}
{"index":{"_index":"demo1_blog","_id":"2"}}
{"id":2,"author":"方才兄","influence":{"gte":10,"lte":12},"title":"ElasticSearch系列05：倒排序索引与分词Analysis","content":"系统学习ES】一、 倒排索引是什么？倒排索引是 Elasticsearch 中非常重要的索引结构，是从文档单词到文档 ID 的映射过程","tag":["倒排序索引","分词Analysis"],"vote":9,"view":90,"createAt":"2020-05-17 10:56"}
{"index":{"_index":"demo1_blog","_id":"3"}}
{"id":3,"author":"学堂","influence":{"gte":5,"lte":8},"title":"ElasticSearch安装以及和SpringBoot的整合","content":"自己正好学习一下，ElasticSearch也是nosql中的一种","tag":["ElasticSearch安装","springBoot整合"],"vote":0,"view":61,"createAt":"2020-06-01 10:56"}
{"index":{"_index":"demo1_blog","_id":"4"}}
{"id":4,"author":"阿里云","influence":{"gte":20,"lte":35},"title":"使用ElasticSearch快速搭建检索系统","content":"一个好的搜索系统可以直接促进页面的访问量提升","tag":["ElasticSearch","检索系统"],"vote":30,"view":200,"createAt":"2020-02-24 10:56"}
{"index":{"_index":"demo1_blog","_id":"5"}}
{"id":5,"author":" 铭毅天下","influence":{"gte":15,"lte":20},"title":"Elasticsearch学习，请先看这一篇！","content":"Elasticsearch研究有一段时间了，现特将Elasticsearch相关核心知识、原理从初学者认知、学习的角度，从以下9个方面进行详细梳理。","tag":["ElasticSearch","核心知识"],"vote":30,"view":4200,"createAt":"2020-06-04 10:56"}
{"index":{"_index":"demo1_blog","_id":"6"}}
{"id":6,"author":" 方才兄","influence":{"gte":15,"lte":20},"title":"Elasticsearch系列13：彻底掌握相关度","content":"最后，如果你有更好的相关度控制方式，或者在es的学习过程中有疑问，欢迎加入es交流群，和大家一起系统学习ElasticSearch。","tag":["ES","相关度"],"vote":10,"view":170,"createAt":"2020-06-08 10:56"}

1.5 索引简单分析

根据我们一般的检索经验，对于博客的标题 title、内容 content 均使用 ik分词进行分词，对title 进行 ik_max_word 细颗粒度分词，保证查全率；考虑到 content 的内容一般较多，使用 ik_smart 粗颗粒分词即可。

对于博客的标签 tag，在某些博客系统中是可以直接使用标签过滤的，所以 tag 需要 type 为 keyword 的索引，用于精确过滤；同时标签也能被用于检索，使用 ik_max_word 进行分词。所以tag使用 fields 配置两种分词效果。

02 博客检索系统开发

2.1 pom依赖

      <properties>
        <revision>20200607.0900</revision>
        <type>SNAPSHOT</type>
        <java.version>1.8</java.version>
        <es.version>7.7.1</es.version>
        <swagger.version>2.8.0</swagger.version>
        <fastjson.version>1.2.70</fastjson.version>
        <commons-lang3.version>3.10</commons-lang3.version>
    </properties>

  <dependencies>

        <dependency>
            <groupId>org.elasticsearch</groupId>
            <artifactId>elasticsearch</artifactId>
            <version>${es.version}</version>
        </dependency>
        <dependency>
            <groupId>org.elasticsearch.client</groupId>
            <artifactId>elasticsearch-rest-high-level-client</artifactId>
            <version>${es.version}</version>
        </dependency>

        <!-- https://mvnrepository.com/artifact/org.apache.commons/commons-lang3 -->
        <dependency>
            <groupId>org.apache.commons</groupId>
            <artifactId>commons-lang3</artifactId>
            <version>${commons-lang3.version}</version>
        </dependency>

        <dependency>
            <groupId>com.alibaba</groupId>
            <artifactId>fastjson</artifactId>
            <version>${fastjson.version}</version>
        </dependency>
        <dependency>
            <groupId>io.springfox</groupId>
            <artifactId>springfox-swagger2</artifactId>
            <version>${swagger.version}</version>
        </dependency>
        <dependency>
            <groupId>io.springfox</groupId>
            <artifactId>springfox-swagger-ui</artifactId>
            <version>${swagger.version}</version>
        </dependency>

        <dependency>
            <groupId>org.springframework.boot</groupId>
            <artifactId>spring-boot-starter-web</artifactId>
        </dependency>

        <dependency>
            <groupId>org.projectlombok</groupId>
            <artifactId>lombok</artifactId>
            <optional>true</optional>
        </dependency>
        <dependency>
            <groupId>org.springframework.boot</groupId>
            <artifactId>spring-boot-starter-test</artifactId>
            <scope>test</scope>
            <exclusions>
                <exclusion>
                    <groupId>org.junit.vintage</groupId>
                    <artifactId>junit-vintage-engine</artifactId>
                </exclusion>
            </exclusions>
        </dependency>
    </dependencies>

2.2 yml 配置文件

本文只提供一个简单的示例，es的其他配置详解后续专门分享。

server:
  port: 6700

# 关闭es健康检查
management:
  health:
    elasticsearch:
      enabled: false

spring:
  data:
    elasticsearch:
      nodes: 192.168.1.181:9200 # es地址
      repositories:
        enabled: true
    # 开启es健康检查
#    rest:
#      uris: ["http://192.168.1.181:9200"]

2.3 封装RestHighLevelClient

package com.fangcai.es.common.config;

import org.apache.http.HttpHost;
import org.elasticsearch.client.RestClient;
import org.elasticsearch.client.RestHighLevelClient;
import org.slf4j.Logger;
import org.slf4j.LoggerFactory;
import org.springframework.beans.factory.DisposableBean;
import org.springframework.beans.factory.FactoryBean;
import org.springframework.beans.factory.InitializingBean;
import org.springframework.beans.factory.annotation.Value;
import org.springframework.context.annotation.Configuration;
import java.util.ArrayList;
import java.util.List;


/**
 * @author MouFangCai
 * @date 2019/12/6 10:44
 * @description
 */
@Configuration
public class EsConfig implements FactoryBean<RestHighLevelClient>, InitializingBean, DisposableBean {

    private final Logger logger = LoggerFactory.getLogger(this.getClass());

    private final static String SCHEME = "http";

    private RestHighLevelClient restHighLevelClient;

    @Value ("${spring.data.elasticsearch.nodes}")
    private String nodes;


    /**
     * 控制Bean的实例化过程
     *
     * @return
     */
    @Override
    public RestHighLevelClient getObject() {
        return restHighLevelClient;
    }

    /**
     * 获取接口返回的实例的class
     *
     * @return
     */
    @Override
    public Class<?> getObjectType() {
        return RestHighLevelClient.class;
    }

    @Override
    public void destroy() {
        try {
            if (null != restHighLevelClient) {
                restHighLevelClient.close();
            }
        } catch (final Exception e) {
            logger.error("Error closing ElasticSearch client: ", e);
        }
    }

    @Override
    public boolean isSingleton() {
        return false;
    }

    @Override
    public void afterPropertiesSet() {
        restHighLevelClient = buildClient();
    }

    private RestHighLevelClient buildClient() {
        try {
            String[] hosts = nodes.split(",");
            List<HttpHost> httpHosts = new ArrayList<>(hosts.length);
            for (String node : hosts) {
                HttpHost host = new HttpHost(
                        node.split(":")[0],
                        Integer.parseInt(node.split(":")[1]),
                        SCHEME);
                httpHosts.add(host);
            }
            restHighLevelClient = new RestHighLevelClient(
                    RestClient.builder(httpHosts.toArray(new HttpHost[0]))
            );
        } catch (Exception e) {
            logger.error(e.getMessage());
        }
        return restHighLevelClient;
    }
}

2.4 封装EsUtil

提供了查询、聚合、文档的CURD等公用接口

package com.fangcai.es.common.util;

import com.alibaba.fastjson.JSON;
import com.alibaba.fastjson.JSONArray;
import com.alibaba.fastjson.JSONObject;
import com.fangcai.es.common.exception.EsDemoException;
import com.fangcai.es.common.response.PageResponse;
import org.elasticsearch.action.bulk.BulkRequest;
import org.elasticsearch.action.bulk.BulkResponse;
import org.elasticsearch.action.delete.DeleteRequest;
import org.elasticsearch.action.delete.DeleteResponse;
import org.elasticsearch.action.index.IndexRequest;
import org.elasticsearch.action.index.IndexResponse;
import org.elasticsearch.action.search.SearchRequest;
import org.elasticsearch.action.search.SearchResponse;
import org.elasticsearch.client.RequestOptions;
import org.elasticsearch.client.RestHighLevelClient;
import org.elasticsearch.common.xcontent.XContentType;
import org.elasticsearch.rest.RestStatus;
import org.elasticsearch.search.SearchHit;
import org.elasticsearch.search.SearchHits;
import org.elasticsearch.search.aggregations.Aggregations;
import org.elasticsearch.search.aggregations.bucket.terms.Terms;
import org.elasticsearch.search.builder.SearchSourceBuilder;
import org.slf4j.Logger;
import org.slf4j.LoggerFactory;
import org.springframework.beans.factory.annotation.Autowired;
import org.springframework.http.HttpStatus;
import org.springframework.stereotype.Component;
import java.util.ArrayList;
import java.util.HashMap;
import java.util.List;
import java.util.Map;

/**
 * @author MouFangCai
 * @date 2020/6/9 10:52
 * @description es 数据的 CURD API
 *  API 可参考官网：https://www.elastic.co/guide/en/elasticsearch/client/java-rest/7.7/java-rest-high.html
 */
@Component
public class EsUtil {

    private Logger logger = LoggerFactory.getLogger(this.getClass());

    @Autowired
    private RestHighLevelClient esClient;
    private static int retryLimit = 3;


    /**
     * 搜索
     *
     * @param index
     * @param searchSourceBuilder
     * @param clazz 需要封装的obj
     * @param pageNum
     * @param pageSize
     * @return PageResponse<T>
     */
    public <T> PageResponse<T> search(String index, SearchSourceBuilder searchSourceBuilder, Class<T> clazz,
                                      Integer pageNum, Integer pageSize){

        SearchRequest searchRequest = new SearchRequest(index);
        searchRequest.source(searchSourceBuilder);
        logger.info("DSL语句为：{}",searchRequest.source().toString());
        try {
            SearchResponse response = esClient.search(searchRequest, RequestOptions.DEFAULT);
            PageResponse<T> pageResponse = new PageResponse<>();
            pageResponse.setPageNum(pageNum);
            pageResponse.setPageSize(pageSize);
            pageResponse.setTotal(response.getHits().getTotalHits().value);
            List<T> dataList = new ArrayList<>();
            SearchHits hits = response.getHits();
            for(SearchHit hit : hits){
                dataList.add(JSONObject.parseObject(hit.getSourceAsString(), clazz));
            }
            pageResponse.setData(dataList);
            return pageResponse;
        } catch (Exception e) {
            logger.error(e.getMessage());
            throw new EsDemoException(String.valueOf(HttpStatus.BAD_REQUEST),
                    "error to execute searching,because of " + e.getMessage());
        }
    }


    /**
     * 聚合
     *
     * @param index
     * @param searchSourceBuilder
     * @param aggName 聚合名
     * @return Map<Integer, Long>  key:aggName   value: doc_count
     */
    public Map<Integer, Long> aggSearch(String index, SearchSourceBuilder searchSourceBuilder, String aggName){
        SearchRequest searchRequest = new SearchRequest(index);
        searchRequest.source(searchSourceBuilder);
        logger.info("DSL语句为：{}",searchRequest.source().toString());
        try {
            SearchResponse response = esClient.search(searchRequest, RequestOptions.DEFAULT);
            Aggregations aggregations = response.getAggregations();
            Terms terms = aggregations.get(aggName);
            List<? extends Terms.Bucket> buckets = terms.getBuckets();
            Map<Integer, Long> responseMap = new HashMap<>(buckets.size());
            buckets.forEach(bucket-> {
                responseMap.put(bucket.getKeyAsNumber().intValue(), bucket.getDocCount());
            });
            return responseMap;
        } catch (Exception e) {
            logger.error(e.getMessage());
            throw new EsDemoException(String.valueOf(HttpStatus.BAD_REQUEST),
                    "error to execute aggregation searching,because of " + e.getMessage());
        }

    }



    /**
     *  新增或者更新文档
     *
     *  对于更新文档，建议可以直接使用新增文档的API，替代 UpdateRequest
     *  避免因对应id的doc不存在而抛异常：document_missing_exception
     * @param obj
     * @param index
     * @return
     */
    public Boolean addOrUptDocToEs(Object obj, String index){

        try {
            IndexRequest indexRequest = new IndexRequest(index).id(getESId(obj))
                    .source(JSON.toJSONString(obj), XContentType.JSON);
            int times = 0;
            while (times < retryLimit) {
                IndexResponse indexResponse = esClient.index(indexRequest, RequestOptions.DEFAULT);

                if (indexResponse.status().equals(RestStatus.CREATED) || indexResponse.status().equals(RestStatus.OK)) {
                    return true;
                } else {
                    logger.info(JSON.toJSONString(indexResponse));
                    times++;
                }
            }
            return false;
        } catch (Exception e) {
            logger.error("Object = {}, index = {}, id = {} , exception = {}", obj, index, getESId(obj) , e.getMessage());
            throw new EsDemoException(String.valueOf(HttpStatus.BAD_REQUEST),
                    "error to execute add doc,because of " + e.getMessage());
        }

    }


    /**
     *  删除文档
     *
     * @param index
     * @param id
     * @return
     */
    public Boolean deleteDocToEs(Integer id, String index) {
        try {
            DeleteRequest request = new DeleteRequest(index, id.toString());

            int times = 0;
            while (times < retryLimit) {
                DeleteResponse delete = esClient.delete(request, RequestOptions.DEFAULT);

                if (delete.status().equals(RestStatus.OK)) {
                    return true;
                } else {
                    logger.info(JSON.toJSONString(delete));
                    times++;
                }
            }
            return false;
        } catch (Exception e) {
            logger.error("index = {}, id = {} , exception = {}", index, id , e.getMessage());
            throw new EsDemoException(String.valueOf(HttpStatus.BAD_REQUEST),
                    "error to execute update doc,because of " + e.getMessage());
        }
    }


    /**
     * 批量插入 或者 更新
     *
     * @param array 数据集合
     * @param index
     * @return
     */
    public Boolean batchAddOrUptToEs(JSONArray array, String index) {

        try {
            BulkRequest request = new BulkRequest();
            for (Object obj : array) {
                IndexRequest indexRequest = new IndexRequest(index).id(getESId(obj))
                        .source(JSON.toJSONString(obj), XContentType.JSON);
                request.add(indexRequest);
            }
            BulkResponse bulk = esClient.bulk(request, RequestOptions.DEFAULT);

            return bulk.status().equals(RestStatus.OK);
        } catch (Exception e) {
            logger.error("index = {}, exception = {}", index, e.getMessage());
            throw new EsDemoException(String.valueOf(HttpStatus.BAD_REQUEST),
                    "error to execute batch add doc,because of " + e.getMessage());
        }
    }


    /**
     * 批量删除
     * @param deleteIds 待删除的 _id list
     * @param index
     * @return
     */
    public Boolean batchDeleteToEs(List<Integer> deleteIds, String index){
        try {
            BulkRequest request = new BulkRequest();
            for (Integer deleteId : deleteIds) {
                DeleteRequest deleteRequest = new DeleteRequest(index, deleteId.toString());
                request.add(deleteRequest);
            }
            BulkResponse bulk = esClient.bulk(request, RequestOptions.DEFAULT);

            return bulk.status().equals(RestStatus.OK);
        } catch (Exception e) {
            logger.error("index = {}, exception = {}", index, e.getMessage());
            throw new EsDemoException(String.valueOf(HttpStatus.BAD_REQUEST),
                    "error to execute batch update doc,because of " + e.getMessage());
        }
    }


    /**
     * 将obj的id 作为 doc的_id
     * @param obj
     * @return
     */
    private String getESId(Object obj) {
        JSONObject jsonObject = JSON.parseObject(JSON.toJSONString(obj));
        Object id = jsonObject.get("id");
        return JSON.toJSONString(id);
    }
}

2.5 业务代码

ps：以下java代码之所以使用魔法值，是为了方便对照DSL，在实践中，建议使用枚举等常量代替。完整版项目源码获取方式，见文末。

1）场景1

根据 title 、content 、tag 进行简单检索，使用rescore利用match_phrase重新算分排序。

场景分析：为了保证查全率，直接使用对 title 、content 、tag 这3个字段进行 match query 即可；同时为了保证排序的效果更好，使用rescore利用match_phrase重新算分排序。

DSL语句为：

GET /demo1_blog/_search
{
  "query": {
    "multi_match": {
      "query": "系统学习ElasticSearch",
      "fields": [
        "title",
        "content",
        "tag"
      ]
    }
  },
  "rescore": {
    "query": {
      "rescore_query": {
        "multi_match": {
          "query": "系统学习ElasticSearch",
          "fields": [
            "title",
            "content",
            "tag"
          ],
          "type": "phrase"
        }
      }
    },
    "window_size": 10
  }
}

对应java API 为：

    @GetMapping("case1")
    public PageResponse<Blog> case1 (@RequestParam(defaultValue = "1") Integer pageNum,
                                     @RequestParam(defaultValue = "10") Integer pageSize) {

        SearchSourceBuilder searchSourceBuilder = new SearchSourceBuilder();
        
        // 根据 title 、content 、tag 进行 match query
        MultiMatchQueryBuilder multiMatchQuery = QueryBuilders.multiMatchQuery("系统学习ElasticSearch",
                "title","content","tag");
        searchSourceBuilder.query(multiMatchQuery);

        // 使用 reScore 利用 match_phrase 重新算分排
        MultiMatchQueryBuilder reScoreQuery = QueryBuilders.multiMatchQuery("系统学习ElasticSearch",
                "title","content","tag")
                .type(MultiMatchQueryBuilder.Type.PHRASE);
        QueryRescorerBuilder queryRescorerBuilder = new QueryRescorerBuilder(reScoreQuery);
        searchSourceBuilder.addRescorer(queryRescorerBuilder);

        // 分页
        int from = pageSize * (pageNum - 1);
        searchSourceBuilder.size(pageSize).from(from);
        return esUtil.search(EsIndexEnum.BLOG.getIndexName(), searchSourceBuilder,
                Blog.class, pageNum, pageSize);
    }

检索结果为：文档【1，6，4，2，5，3】

2）场景2

通过boost参数控制相关度，提升 tag 的权重为3，title的权重为2，使用默认排序

场景分析：tag 是一篇博客的标识，所以对权重的影响应该是最大的，title 次之。

DSL语句为：

GET /demo1_blog/_search
{
  "query": {
    "multi_match": {
      "query": "系统学习ElasticSearch",
      "fields": [
         "title^2",
            "content",
            "tag^3"
      ]
    }
  }
}
# 等价于
GET /demo1_blog/_search
{
  "query": {
    "bool": {
      "should": [
        {
          "match": {
            "tag": {
              "query": "系统学习ElasticSearch",
              "boost": 3
            }
          }
        },
                {
          "match": {
            "title": {
              "query": "系统学习ElasticSearch",
              "boost": 2
            }
          }
        },
                {
          "match": {
            "content": {
              "query": "系统学习ElasticSearch",
              "boost": 1
            }
          }
        }
      ]
    }
  }
}

对应java API 为：

    @GetMapping("case2")
    public PageResponse<Blog> case2 (@RequestParam(defaultValue = "1") Integer pageNum,
                                     @RequestParam(defaultValue = "10") Integer pageSize) {

        SearchSourceBuilder searchSourceBuilder = new SearchSourceBuilder();

        // 提升 tag 的权重为3，title的权重为2，使用默认排序
        BoolQueryBuilder boolQuery = QueryBuilders.boolQuery();
        boolQuery.should(QueryBuilders.matchQuery("tag", "系统学习ElasticSearch").boost(3))
                .should(QueryBuilders.matchQuery("title", "系统学习ElasticSearch").boost(2))
                .should(QueryBuilders.matchQuery("content", "系统学习ElasticSearch"));
        searchSourceBuilder.query(boolQuery);

        // 分页
        int from = pageSize * (pageNum - 1);
        searchSourceBuilder.size(pageSize).from(from);

        return esUtil.search(EsIndexEnum.BLOG.getIndexName(), searchSourceBuilder,
                Blog.class, pageNum, pageSize);
    }

检索结果为：文档【1，4，5，3，6，2】

ps：上述两个场景，只是为了给大家演示对相关度的控制。我们在实际项目中，可以通过多种方式去控制相关度，以达到我们最想要检索效果。

3）场景3

在case2的基础上增加过滤条件：author、tag、createAt、influence

场景分析：这个检索场景应该是很好理解的，比如说我只想看某个作者的博客，或者像知乎的搜索一样，我只想看最近一个月发布的博客。直接使用 filter 对特定字段过滤即可。

DSL语句为：

# 场景3
GET /demo1_blog/_search
{
  "query": {
    "bool": {
      "must": [
        {    "multi_match": {
      "query": "系统学习ElasticSearch",
      "fields": [
         "title^2",
            "content",
            "tag^3"
      ]
    }}
      ],
      "filter": [
        {
          "term": {
            "author": "方才兄"
          }
        },
       {
         "terms":{
           "tag.keyword":["ElasticSearch","倒排序索引"]
         }
        },
        {
          "range": {
            "createAt": {
              "gte": "now-2M/d",
              "lte": "now"
            }
          }
        }
        ,
        {
          "range": {
            "influence": {
              "gte": 5,
              "lte": 15
            }
          }
        }
      ]
    }

  }
}

对应java API 为：

    @GetMapping("case3")
    public PageResponse<Blog> case3 (@RequestParam(defaultValue = "1") Integer pageNum,
                                     @RequestParam(defaultValue = "10") Integer pageSize) {

        SearchSourceBuilder searchSourceBuilder = new SearchSourceBuilder();

        // 提升 tag 的权重为3，title的权重为2，使用默认排序
        BoolQueryBuilder boolQuery = QueryBuilders.boolQuery();
        boolQuery.should(QueryBuilders.matchQuery("tag", "系统学习ElasticSearch").boost(3))
                .should(QueryBuilders.matchQuery("title", "系统学习ElasticSearch").boost(2))
                .should(QueryBuilders.matchQuery("content", "系统学习ElasticSearch"));

        // 过滤
        boolQuery.filter(QueryBuilders.termQuery("author", "方才兄"));
        boolQuery.filter(QueryBuilders.termsQuery("tag.keyword", "ElasticSearch", "倒排序索引"));
        boolQuery.filter(QueryBuilders.rangeQuery("createAt").gte("now-3M/d").lte("now/d"));
        boolQuery.filter(QueryBuilders.rangeQuery("influence").gte(5).lte(15));

        searchSourceBuilder.query(boolQuery);

        // 分页
        int from = pageSize * (pageNum - 1);
        searchSourceBuilder.size(pageSize).from(from);

        return esUtil.search(EsIndexEnum.BLOG.getIndexName(), searchSourceBuilder,
                Blog.class, pageNum, pageSize);
    }

检索结果为：文档【1，2】

4）场景4

在case3的基础上用户指定排序条件：createAt、vote、view

场景分析：就像微信的搜一搜一样，用户可以选择排序的方式，根据发布时间，或者根据阅读量。在这种情况下，就没必要进行相关性算分了，所以整个检索都应该在 filter context中。

DSL语句为：

# 场景4
GET /demo1_blog/_search
{
  "query": {
    "bool": {
      "filter": [
        {
          "multi_match": {
            "query": "系统学习ElasticSearch",
            "fields": [
              "title^2",
              "content",
              "tag^3"
            ]
          }
        },
        {
          "term": {
            "author": "方才兄"
          }
        },
       {
         "terms":{
           "tag.keyword":["ElasticSearch","倒排序索引"]
         }
        },
        {
          "range": {
            "createAt": {
              "gte": "now-3M/d",
              "lte": "now"
            }
          }
        },
        {
          "range": {
            "influence": {
              "gte": 10,
              "lte": 15
            }
          }
        }
      ]
    }
  },
  "sort": [
    {
      "createAt": {
        "order": "desc"
      }
    }
  ]
}

对应java API 为：

    @GetMapping("case4")
    public PageResponse<Blog> case4 (@RequestParam(defaultValue = "1") Integer pageNum,
                                     @RequestParam(defaultValue = "10") Integer pageSize) {

        SearchSourceBuilder searchSourceBuilder = new SearchSourceBuilder();

        // 通过 filterContext 查询，忽略评分，增加缓存的可能性，提高查询性能
        BoolQueryBuilder boolQuery = QueryBuilders.boolQuery();
        MultiMatchQueryBuilder multiMatchQuery = QueryBuilders.multiMatchQuery("系统学习ElasticSearch",
                "title","content","tag");
        boolQuery.filter(multiMatchQuery);

        // 过滤
        boolQuery.filter(QueryBuilders.termQuery("author", "方才兄"));
        boolQuery.filter(QueryBuilders.termsQuery("tag.keyword", "ElasticSearch", "倒排序索引"));
        boolQuery.filter(QueryBuilders.rangeQuery("createAt").gte("now-3M/d").lte("now/d"));
        boolQuery.filter(QueryBuilders.rangeQuery("influence").gte(5).lte(15));

        searchSourceBuilder.query(boolQuery);
        searchSourceBuilder.sort("view", SortOrder.DESC);
        // 分页
        int from = pageSize * (pageNum - 1);
        searchSourceBuilder.size(pageSize).from(from);

        return esUtil.search(EsIndexEnum.BLOG.getIndexName(), searchSourceBuilder,
                Blog.class, pageNum, pageSize);
    }

检索结果为：文档【2，1】

03 关于elasticsearch-rest-high-level-client

通过上节的内容，不知道小伙伴们发现了没有，elasticsearch-rest-high-level-client 其实已经把各种方法都封装得很简单了，对于各种检索场景，难点在于DSL的编写，然后直接根据DSL开发API即可。

在此，和各位小伙伴分享分享TeHero对 elasticsearch-rest-high-level-client 的使用经验。就以我们常见的查询为例：

3.1 RestHighLevelClient

RestHighLevelClient，简单来说，它包装了一个LowLevelClient【RestClient】，我们使用它来构建我们的Request请求，以及获取响应Response。 RestHighLevelClient 的大多数方法都有两种形式，一个是阻塞【同步】的，一个是异步的。

在idea中，我们可以进入到RestHighLevelClient类，ctrl+F12，即可查看该类所有的方法，同时支持搜索，比如我们常用的 search( ) 方法：

一看该方法的说明，就知道是干嘛的了：

如果不知道自己该用哪个方法怎么办？很简单，直接看官网：Java REST Client-ES官网。通过目录，我们就可以快速定位到我们想要的api是哪个，就比如说我们的 search ( ) ：

SearchResponse response = esClient.search(searchRequest, RequestOptions.DEFAULT);

直接点击查看，都有介绍该方法该如何使用：

3.2 SearchSourceBuilder

通过上图我们可以看到SearchRequest需要一个SearchSourceBuilder。

SearchRequest searchRequest = new SearchRequest(index);
searchRequest.source(searchSourceBuilder);

和DSL对比理解，SearchSourceBuilder就是最外面的一层：

通过 idea 查看该类提供的方法：

通过 query(QueryBuilder query) 方法去构建我们的查询语句

结合实例看下：

        SearchSourceBuilder searchSourceBuilder = new SearchSourceBuilder();
        
        // 根据 title 、content 、tag 进行 match query
        MultiMatchQueryBuilder multiMatchQuery = QueryBuilders.multiMatchQuery("系统学习ElasticSearch",
                "title","content","tag");
        searchSourceBuilder.query(multiMatchQuery);

        // 使用 reScore 利用 match_phrase 重新算分排
        MultiMatchQueryBuilder reScoreQuery = QueryBuilders.multiMatchQuery("系统学习ElasticSearch",
                "title","content","tag")
                .type(MultiMatchQueryBuilder.Type.PHRASE);
        QueryRescorerBuilder queryRescorerBuilder = new QueryRescorerBuilder(reScoreQuery);
        searchSourceBuilder.addRescorer(queryRescorerBuilder);

        // 分页
        int from = pageSize * (pageNum - 1);
        searchSourceBuilder.size(pageSize).from(from);

3.3 QueryBuilder

我们知道 query(QueryBuilder query) 方法需要一个 QueryBuilder ，而 QueryBuilder是一个接口，那么我们只能将它的实现作为参数输入，依然可以直接通过搜索，获取到我们想要的。

比如说 match query，可以很方便的找到：MatchQueryBuilder。