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
代码语言：javascript复制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
代码语言：javascript复制# 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
代码语言：javascript复制 <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。
代码语言：javascript复制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
代码语言：javascript复制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
代码语言：javascript复制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：
代码语言：javascript复制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：
代码语言：javascript复制 @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：
代码语言：javascript复制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：
代码语言：javascript复制 @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：
代码语言：javascript复制# 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：
代码语言：javascript复制 @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：
代码语言：javascript复制# 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：
代码语言：javascript复制 @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 ( ) ：
代码语言：javascript复制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。
代码语言：javascript复制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：
代码语言：javascript复制 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。
你可以直接通过 new MatchQueryBuilder（）的形式创建，但是没必要，因为ES为我们提供了构建者：QueryBuilders。
3.4 QueryBuilders
使用非常方便，不知道如何传参，直接进去看方法说明即可：
MatchQueryBuilder matchQueryBuilder = QueryBuilders.matchQuery("fieldName", "search keyword");
可以看到，QueryBuilders 几乎提供了所有查询的构建方法：
3.5 BoolQueryBuilder
bool查询在我们日常的查询中用得是非常多的，直接通过QueryBuilders即可构建：
BoolQueryBuilder boolQuery = QueryBuilders.boolQuery();
一看就懂系列：
3.6 总结
在不熟悉RestHighLevelClient之前，先根据检索需求，写出DSL语句，按照DSL语句，逐个封装SearchSourceBuilder即可。
在我们开发的过程中，可以通过 SearchRequest 将我们的DSL语句打印出来，方便我们验证DSL语句是否拼写正确。
logger.info("DSL语句为：{}",searchRequest.source().toString());
以上就是今天TeHero为大家分享的内容，通过搭建博客搜索系统，去学习RestHighLevelClient的使用，如果你有更好的使用经验或者疑问，欢迎加入【ES学习社群】，和大家一起交流学习。
下期预告：ES中的聚合查询【关注公众号：方才编程，系统学习ES】
