ElasticSearch 实现分词全文检索 - 搜素关键字自动补全(Completion Suggest)
目录
ElasticSearch 实现分词全文检索 - 概述
ElasticSearch 实现分词全文检索 - ES、Kibana、IK安装
ElasticSearch 实现分词全文检索 - Restful基本操作
ElasticSearch 实现分词全文检索 - Java SpringBoot ES 索引操作
ElasticSearch 实现分词全文检索 - Java SpringBoot ES 文档操作
ElasticSearch 实现分词全文检索 - 测试数据准备
ElasticSearch 实现分词全文检索 - term、terms查询
ElasticSearch 实现分词全文检索 - match、match_all、multimatch查询
ElasticSearch 实现分词全文检索 - id、ids、prefix、fuzzy、wildcard、range、regexp 查询
ElasticSearch 实现分词全文检索 - Scroll 深分页
ElasticSearch 实现分词全文检索 - delete-by-query
ElasticSearch 实现分词全文检索 - 复合查询
ElasticSearch 实现分词全文检索 - filter查询
ElasticSearch 实现分词全文检索 - 高亮查询
ElasticSearch 实现分词全文检索 - 聚合查询 cardinality
ElasticSearch 实现分词全文检索 - 经纬度查询
ElasticSearch 实现分词全文检索 - 搜素关键字自动补全(suggest)
ElasticSearch 实现分词全文检索 - SpringBoot 完整实现 Demo 附源码
需求
搜素关键字自动补全(suggest)
输入“人工” 自动带出人工开头的关键字
Kibana 界面操作 实现 搜素关键字自动补全(suggest)
ES使用Completion Suggest 做关键字自动补全时,实际应用中搜索性能更加高效,建议多开一个子字段,如下示例,假设要根据title字段做关键字自动补全,不要改原字段的类型,多开一个子字段title.suggest,类型设置为completion,然后之后的suggest针对title.suggest字段做操作
- Term Suggester:词条建议器。对给输入的文本进进行分词,为每个分词提供词项建议, 基于编辑距离,对analyze过的单个term去提供建议,并不会考虑多个term/词组之间的关系。quert -> query
- Phrase Suggester:短语建议器,在term的基础上,会考量多个term之间的关系在Term Suggester的基础上,通过ngram以词组为单位返回建议。noble prize -> nobel prize
- Completion Suggester:它主要针对的应用场景就是"Auto Completion",FST数据结构,类似Trie树,不用打开倒排,快速返回,前缀匹配
- Context Suggester:上下文建议器,在Completion Suggester的基础上,用于filter和boost
创建索引
创建索引并指定结构
PUT /article-index
{
"settings": {
"number_of_shards": 3,
"number_of_replicas": 0
},
"mappings": {
"properties":{
"id":{
"type":"keyword"
},
"title":{
"type":"text",
"analyzer":"ik_max_word",
"fields": { 扩展一个字段,用于关键字自动补全查询
"suggest" : {
"type" : "completion",
"analyzer": "ik_max_word"
}
}
},
"summary":{
"type":"text",
"analyzer":"ik_max_word"
},
"createDate":{
"type":"date",
"format":"yyyy-MM-dd HH:mm:ss||yyyy-MM-dd"
}
}
}
}
添加数据
JSON { 括号里面的内容,不能换行 }
_bulk 批量添加文档
POST /article-index/_doc/_bulk
{"index":{"_id":1}}
{"id":1,"title":"人工智能技术","summary":"ElasticSearch 实现分词全文检索 - ES、Kibana、IK安装","createDate":"2023-02-23"}
{"index":{"_id":2}}
{"id":2,"title":"人工智能软件 Chart GTP","summary":"太极生两仪,两仪生四象,四象生八卦","createDate":"2023-02-23"}
{"index":{"_id":3}}
{"id":3,"title":"Restful基本操作","summary":"ElasticSearch 实现分词全文检索 - Java SpringBoot ES 索引操作","createDate":"2023-02-23"}
{"index":{"_id":4}}
{"id":4,"title":"人工呼吸","summary":"ElasticSearch 实现分词全文检索 - 经纬度查询","createDate":"2023-02-23"}
{"index":{"_id":5}}
{"id":5,"title":"SpringBoot 全文检索实战","summary":"ElasticSearch 实现分词全文检索 - SpringBoot 全文检索实战","createDate":"2023-02-23"}
查询数据
查询
GET /article-index/_doc/_search
{
"suggest": {
"my-suggest" : {
"prefix" : "人",
"completion" : {
"field" : "title.suggest"
}
}
}
}
返回值--自动带出人开头的关键字
{
"took" : 3,
"timed_out" : false,
"_shards" : {
"total" : 3,
"successful" : 3,
"skipped" : 0,
"failed" : 0
},
"hits" : {
"total" : {
"value" : 0,
"relation" : "eq"
},
"max_score" : null,
"hits" : [ ]
},
"suggest" : {
"my-suggest" : [
{
"text" : "人",
"offset" : 0,
"length" : 1,
"options" : [
{
"text" : "人工呼吸",
"_index" : "article-index",
"_type" : "_doc",
"_id" : "4",
"_score" : 1.0,
"_source" : {
"id" : 4,
"title" : "人工呼吸",
"summary" : "ElasticSearch 实现分词全文检索 - 经纬度查询",
"createDate" : "2023-02-23"
}
},
{
"text" : "人工智能技术",
"_index" : "article-index",
"_type" : "_doc",
"_id" : "1",
"_score" : 1.0,
"_source" : {
"id" : 1,
"title" : "人工智能技术",
"summary" : "ElasticSearch 实现分词全文检索 - ES、Kibana、IK安装",
"createDate" : "2023-02-23"
}
},
{
"text" : "人工智能软件 Chart GTP",
"_index" : "article-index",
"_type" : "_doc",
"_id" : "2",
"_score" : 1.0,
"_source" : {
"id" : 2,
"title" : "人工智能软件 Chart GTP",
"summary" : "太极生两仪,两仪生四象,四象生八卦",
"createDate" : "2023-02-23"
}
}
]
}
]
}
}
JAVA SpringBoot 实现 搜素关键字自动补全(suggest)
创建索引
/**
* 第一步:系统初始化,创建索引
* 如果索引不存在,创建,输出
*/
@Test
void createIndexTest() throws Exception {
boolean indexExists = elasticSearchUtil.indexExists(INDEX_NAME);
if (!indexExists) {
try {
createIndex(INDEX_NAME);
logger.info("索引【{}】,创建成功", INDEX_NAME);
//测试效果 可再次查询验证。
indexExists = elasticSearchUtil.indexExists(INDEX_NAME);
logger.info("索引【{}】, {}", INDEX_NAME, indexExists ? "验证存在" : "验证不存在");
} catch (Exception e) {
logger.error(e.getMessage(), e);
}
} else {
logger.info("索引【{}】已存在,无需创建", INDEX_NAME);
}
}
/**
* 创建索引
*
* @param indexName
* @throws Exception
*/
void createIndex(String indexName) throws Exception {
//准备索引的 settings
Settings.Builder settings = Settings.builder()
.put("number_of_shards", INDEX_NUMBER_OF_SHARDS) //分片数,可以使用常量
.put("number_of_replicas", esProperties.getReplicasNum()); //是否集群,需要多少副本,在配置文件中配置
//准备索引的结构 Mappings
XContentBuilder mappings = JsonXContent.contentBuilder()
.startObject()
.startObject("properties")
.startObject("id").field("type", "keyword").endObject()
.startObject("title").field("type", "text").field("analyzer", "ik_max_word")
.startObject("fields").startObject("suggest").field("type", "completion").field("analyzer", "ik_max_word").endObject().endObject()
.endObject() //对该字段进行分词
.startObject("summary").field("type", "text").field("analyzer", "ik_max_word").endObject() //对该字段进行分词
.startObject("createDate").field("type", "date").field("format", "yyyy-MM-dd HH:mm:ss").endObject()
.endObject()
.endObject();
CreateIndexResponse resp = elasticSearchUtil.createIndex(indexName, settings, mappings);
//输出
logger.info("CreateIndexResponse => {} ", resp.toString());
}
添加数据
/**
* 第二步:模拟后台管理员,在添加文章时,将要检查的字段内容,同步到ES中
*/
@Test
void addArticleTest() throws Exception {
Map<Integer, String> titleMap = new HashMap<>();
titleMap.put(1, "人工智能技术");
titleMap.put(2, "人工智能软件 Chart GTP");
titleMap.put(3, "Restful基本操作");
titleMap.put(4, "Java SpringBoot ES 索引操作");
titleMap.put(5, "Java SpringBoot ES 文档操作");
titleMap.put(6, "人工呼吸");
titleMap.put(7, "SpringBoot 全文检索实战");
Map<Integer, String> introMap = new HashMap<>();
introMap.put(1, "ElasticSearch 实现分词全文检索 - 概述");
introMap.put(2, "ElasticSearch 实现分词全文检索 - ES、Kibana、IK安装");
introMap.put(3, "ElasticSearch 实现分词全文检索 - Restful基本操作");
introMap.put(4, "ElasticSearch 实现分词全文检索 - Java SpringBoot ES 索引操作");
introMap.put(5, "ElasticSearch 实现分词全文检索 - Java SpringBoot ES 文档操作");
introMap.put(6, "ElasticSearch 实现分词全文检索 - 经纬度查询");
introMap.put(7, "ElasticSearch 实现分词全文检索 - SpringBoot 全文检索实战");
//内容
Map<Integer, String> contentMap = new HashMap<>();
contentMap.put(1, "【阿里云】尊敬的vipsoft:您有2台云服务器ECS配置升级成功。如有CPU、内存变更或0Mbps带宽升级,您需要在ECS控制台手动重启云服务器后才能生效。");
contentMap.put(2, "为更好地为您提供服务,温馨提醒:您本月有1次抽奖机会,赢取大额通用流量,月月抽月月领,点击掌厅链接 原URL:http://wap.js.10086.cn/Mq 快来试试你的运气吧,如本月已参与请忽略【江苏移动心级服务,让爱连接】");
contentMap.put(3, "国家反诈中心提醒:公检法机关会当面向涉案人员出示证件或法律文书,绝对不会通过网络给当事人发送通缉令、拘留证、逮捕证等法律文书,并要求转账汇款。\n" +
"切记:公检法机关不存在所谓“安全账户”,更不会让你远程转账汇款!");
contentMap.put(4, "【江苏省公安厅、江苏省通信管理局】温馨提示:近期利用苹果手机iMessage消息冒充熟人、冒充领导换号、添加新微信号等诈骗形式多发。如有收到类似短信,请您谨慎判断,苹果手机用户如无需要可关闭iMessage功能,以免上当受骗。");
contentMap.put(5, "多一点快乐,少一点懊恼,不管钞票有多少,只有天天开心就好,累了就睡觉,生活的甜苦,自己来调味。收到信息就要开心的笑");
contentMap.put(6, "黄金周好运每天交,我把祝福来送到:愿您生活步步高,彩票期期中,股票每天涨,生意年年旺,祝您新年新景象!");
contentMap.put(7, "【阿里云】当你手机响,那是我的问候;当你收到短信,那有我的心声;当你翻阅短信,那有我的牵挂;当你筹备关机时,记得我今天说过周末快乐!");
contentMap.put(8, "我刚去了一趟银行,取了无数的幸福黄金好运珠宝平安翡翠成功股票健康基金。嘘!别作声,统统的送给你,因为我想提“钱”祝你国庆节快乐!");
contentMap.put(9, "一个人的精彩,一个人的打拼,一个人的承载,一个人的舞蹈。光棍节送你祝福,不因你是光棍,只因你生活色彩。祝你:快乐打拼,生活出彩!");
contentMap.put(10, "爆竹响激情燃放,雪花舞祥风欢畅,烟火腾期待闪亮,感动涌心中激荡,心情美春节冲浪,愿景好心中珍藏,祝与福短信奉上:祝您身体健康,兔年吉祥!");
//模似7次 添加文章
for (int i = 1; i <= 7; i++) {
ArticleInfo article = new ArticleInfo();
article.setId(String.valueOf(i));
article.setTitle(titleMap.get(i));
article.setAuthor("VipSoft");
article.setSummary(introMap.get(i));
article.setContent(contentMap.get(i));
article.setCreateTime(new Date());
//将article 保存到 MySQL --- 省略
boolean flag = true; //保存数据到 MySQL 数据库成功
if (flag) {
//将需要查询的数据,赋给DTO,更新到 ES中
ArticleDTO articleDTO = new ArticleDTO();
BeanUtils.copyProperties(article, articleDTO);
String json = JSON.toJSONStringWithDateFormat(articleDTO, "yyyy-MM-dd HH:mm:ss"); //FastJson 将日期格式化
IndexResponse resp = elasticSearchUtil.createDoc(INDEX_NAME, articleDTO.getId(), json);
logger.info(" {}", resp.getResult().toString());
}
}
}
查询数据
/**
* 第三步:模拟用户搜索,输入关键词“人”,带出和人有关的关键词
*/
@Test
void earchTest() throws Exception {
List<String> resp = elasticSearchUtil.suggest(INDEX_NAME, "title.suggest", "人", 2);
//4. 获取到 _source 中的数据,并展示
for (String hit : resp) {
System.out.println(hit);
}
}
/**
* 自动补全 根据用户的输入联想到可能的词或者短语
*
* @param indexName 索引名称
* @param field 搜索条件字段
* @param keywords 搜索关键字
* @param size 匹配数量
* @return
* @throws Exception
*/
public List<String> suggest(String indexName, String field, String keywords, int size) throws Exception {
//定义返回
List<String> suggestList = new ArrayList<>();
//构建查询请求
SearchRequest searchRequest = new SearchRequest(indexName);
//通过查询构建器定义评分排序
SearchSourceBuilder searchSourceBuilder = new SearchSourceBuilder();
searchSourceBuilder.sort(new ScoreSortBuilder().order(SortOrder.DESC));
//构造搜索建议语句,搜索条件字段
CompletionSuggestionBuilder completionSuggestionBuilder = new CompletionSuggestionBuilder(field);
//搜索关键字
completionSuggestionBuilder.prefix(keywords);
//去除重复
completionSuggestionBuilder.skipDuplicates(true);
//匹配数量
completionSuggestionBuilder.size(size);
searchSourceBuilder.suggest(new SuggestBuilder().addSuggestion("my-suggest", completionSuggestionBuilder));
//czbk-suggest为返回的字段,所有返回将在czbk-suggest里面,可写死,sort按照评分排序
searchRequest.source(searchSourceBuilder);
//定义查找响应
SearchResponse suggestResponse = esClient.search(searchRequest, RequestOptions.DEFAULT);
//定义完成建议对象
CompletionSuggestion completionSuggestion = suggestResponse.getSuggest().getSuggestion("my-suggest");
List<CompletionSuggestion.Entry.Option> optionsList = completionSuggestion.getEntries().get(0).getOptions();
//从optionsList取出结果
if (!CollectionUtils.isEmpty(optionsList)) {
optionsList.forEach(item -> suggestList.add(item.getText().toString()));
}
return suggestList;
}