前言
最近在学习elasticsearch,想实现跟谷歌和百度类似的功能:下拉补全提示,如图所示:
准备
我使用的版本和依赖包,如下所示:
17
3.0.2
2022.0.0
2022.0.0.0-RC2
org.springframework.boot
spring-boot-starter-data-elasticsearch
co.elastic.clients
elasticsearch-java
8.7.1
devtools操作
首先,我们必须先创建index
并设置成想要设置成下拉补全提示的字段为completion
类型,如下所示:
PUT index_urls
{
"mappings": {
"properties": {
"suggest": {
"type": "completion",
"analyzer": "ik_smart"
}
}
}
}
这样我们创建了index_urls
,并且设置好了suggest
字段是completion
类型。
接着,我们往里面加点数据,并且使用completion suggester
进行查询,如下所示 :
// 自动补全查询
GET /index_urls/_search
{
"suggest": {
"title_suggest": {
"text": "n", // 关键字
"completion": {
"field": "suggest", // 补全查询的字段
"skip_duplicates": true, // 跳过重复的
"size": 10 // 获取前10条结果
}
}
}
}
查询结果如下所示:
{
"took": 1,
"timed_out": false,
"_shards": {
"total": 1,
"successful": 1,
"skipped": 0,
"failed": 0
},
"hits": {
"total": {
"value": 0,
"relation": "eq"
},
"max_score": null,
"hits": []
},
"suggest": {
"title_suggest": [
{
"text": "n",
"offset": 0,
"length": 1,
"options": [
{
"text": "npm、yarn的安装和设置淘宝镜像源_yarn设置淘宝镜像_cn_ljr的博客-CSDN博客",
"_index": "index_urls",
"_id": "64d0548bd8239d21cb9fe4e5",
"_score": 1,
"_source": {
"_class": "com.seaurl.searchservice.document.UrlIndex",
"id": "64d0548bd8239d21cb9fe4e5",
"uid": "6461e6ac25d966329b7d7642",
"categoryId": "64b89acfaf50e77e53ef89ef",
"categoryName": "户外运动",
"categoryParentId": "",
"categoryParentName": "",
"title": "npm、yarn的安装和设置淘宝镜像源_yarn设置淘宝镜像_cn_ljr的博客-CSDN博客",
"suggest": {
"input": [
"npm、yarn的安装和设置淘宝镜像源_yarn设置淘宝镜像_cn_ljr的博客-CSDN博客"
]
},
"url": "https://blog.csdn.net/cn_ljr/article/details/126319130",
"domain": "blog.csdn.net",
"description": "安装npm和yarn,配置npm和yarn的镜像源为淘宝镜像源_yarn设置淘宝镜像",
"favicon": "https://cdn.seaurl.com/space/url/blog.csdn.net/favicon.ico",
"env": "dev",
"createdDt": 1691374731421,
"updatedDt": 1691374731421
}
}
]
}
]
}
}
这样就完成了自动补全提示查询了,接下来我们使用java再实现一遍。
java操作
首先,我们创建实体类,并且定义completion
类型的字段,如下所示:
@Data
public class UrlIndex {
private String id;
private String uid;
// 分类id
private String categoryId;
private String categoryName;
private String categoryParentId;
private String categoryParentName;
private String title;
// 下拉提示建议使用的字段,注意:Completion类型字段,不能做筛选用
// @CompletionField(analyzer="ik_smart",searchAnalyzer="ik_smart", maxInputLength = 100)
private Completion suggest;
private String url;
private String domain;
private String description;
//如果没获取前端显示domain首字母
private String favicon;
private Date createdDt;//创建时间
private Date updatedDt;//更新时间
}
接着,我们往es中添加数据,并且把查询转化成java代码,如下所示:
private SearchResponse getEsSuggestionData(String filedName, String fieldValue) {
try {
String index = getEsIndexByEnv();
Suggester suggester = Suggester.of(s -> s
.suggesters("title-suggest", FieldSuggester.of(fs -> fs
.completion(cs -> cs.skipDuplicates(true)
.size(10)
.field(filedName)
)
))
.text(fieldValue)
);
return elasticsearchClient.search(s ->
s.index(index)
// 使用source可以返回固定字段
// .source()
.suggest(suggester)
, Map.class);
} catch (Exception ex) {
log.info("getEsSuggestionData error={}", ex.getMessage());
return null;
}
}
这样就完成了查询的功能,接下来,我们要把返回的数据拿出来,并转换之后返回给前端显示,如下所示:
@Override
public List searchSuggestionDataBy(String fieldName, String fieldValue) {
SearchResponse response = getEsSuggestionData(fieldName, fieldValue);
if (response != null) {
Map suggestMap = response.suggest();
if (suggestMap != null) {
List> titleSuggestionList = (List>) suggestMap.get("title-suggest");
if (titleSuggestionList != null) {
for (Object titleSuggestion : titleSuggestionList) {
if (titleSuggestion instanceof Suggestion) {
Suggestion suggestion = (Suggestion) titleSuggestion;
if (suggestion.isCompletion()) {
CompletionSuggest completionSuggest = suggestion.completion();
List options = completionSuggest.options();
List urlIndexDtos = new ArrayList<>();
for (co.elastic.clients.elasticsearch.core.search.CompletionSuggestOption option : options) {
Object obj = option.source();
ObjectMapper objectMapper = new ObjectMapper();
UrlIndexDto urlIndexDto = objectMapper.convertValue(obj, UrlIndexDto.class);
urlIndexDtos.add(urlIndexDto);
}
return urlIndexDtos;
}
}
}
}
}
}
return null;
}
这样就完成了查询并返回的操作,下面我们讲解下前端如何实现。
前端操作
我们使用react和antd来实现,先定义组件和相关方法,如下所示:
{
setInput(e)
debouncedOnChange(e);
}}>
} size={'large'}
onPressEnter={onSearch} />
相关方法:
const [input, setInput] = useState('')
const [options, setOptions] = useState([])
// 在组件外部定义防抖函数
const debouncedOnChange = _.debounce(async (e) => {
// 调用自动补全接口
const res = await dispatch(webSearchSuggestion({
q: e
}))
const resData = res.payload.data
if (resData && resData.length>0){
let list = []
resData.forEach(item =>{
console.log('item resData=', item)
list.push({
label: item.title,
value: item.title,
url: item.url
})
})
// 将数据设置到options中
setOptions(list)
}else{
setOptions([])
}
}, 500);
// 搜索框回车事件,调用查询接口(非自动补全接口)
async function onSelect(value, option){
dispatch(webSearch({
q: option.value,
pageNum: 1,
pageSize: 20
}))
}
我们使用了ladash的防抖函数来实现搜索。
注意
1、不是你在实体类中定义了Completion
类型的字段就ok了,要提前声明Completion
才管用,亲自踩过的坑!!!
总结
1、完成这些代码其实花费了一点时间的,主要是不知道如何把devtools的查询转换成java代码,并且我的版本是最新的,用的依赖包是上面提到的,所以写法上市面上没有找到过,完全靠自己搜索和chatgpt给的提示完成的。
引用
Completion Suggester — Java API Client
十.全文检索ElasticSearch经典入门-自动补全功能