ES支持中文&&全拼&&拼音首字母搜索
2019-06-13 本文已影响0人
箭飞天
环境
Ubuntu18.04ES 6.6.1
搜索模板
新建模板，便于后续创建索引，直接使用模板，省事方便
PUT _template/goods
{
"index_patterns":"goods*",
"settings": {
"index.number_of_replicas": "1",
"index.number_of_shards": "5",
"index.translog.flush_threshold_size": "512mb",
"index.translog.sync_interval": "60s",
"index.codec": "best_compression",
"analysis": {
"filter": {
"edge_ngram_filter": {
"type": "edge_ngram",
"min_gram": 1,
"max_gram": 50
},
"simple_pinyin_filter": {
"type": "pinyin",
"keep_first_letter": true,
"keep_separate_first_letter": false,
"keep_full_pinyin": false,
"keep_original": false,
"limit_first_letter_length": 50,
"lowercase": true
},
"full_pinyin_filter": {
"type": "pinyin",
"keep_first_letter": false,
"keep_separate_first_letter": false,
"keep_full_pinyin": true,
"none_chinese_pinyin_tokenize": true,
"keep_original": false,
"limit_first_letter_length": 50,
"lowercase": true
}
},
"char_filter": {
"charconvert": {
"type": "mapping",
"mappings_path": "char_filter_text.txt"
}
},
"tokenizer": {
"ik_max_word": {
"type": "ik_max_word",
"use_smart": true
}
},
"analyzer": {
"ngramIndexAnalyzer": {
"type": "custom",
"tokenizer": "keyword",
"filter": [
"edge_ngram_filter",
"lowercase"
],
"char_filter": [
"charconvert"
]
},
"ngramSearchAnalyzer": {
"type": "custom",
"tokenizer": "keyword",
"filter": [
"lowercase"
],
"char_filter": [
"charconvert"
]
},
"ikIndexAnalyzer": {
"type": "custom",
"tokenizer": "ik_max_word",
"char_filter": [
"charconvert"
]
},
"ikSearchAnalyzer": {
"type": "custom",
"tokenizer": "ik_max_word",
"char_filter": [
"charconvert"
]
},
"simplePinyinIndexAnalyzer": {
"tokenizer": "keyword",
"filter": [
"simple_pinyin_filter",
"edge_ngram_filter",
"lowercase"
]
},
"simplePinyinSearchAnalyzer": {
"tokenizer": "keyword",
"filter": [
"simple_pinyin_filter",
"lowercase"
]
},
"fullPinyinIndexAnalyzer": {
"tokenizer": "keyword",
"filter": [
"full_pinyin_filter",
"edge_ngram_filter",
"lowercase"
]
},
"fullPinyinSearchAnalyzer": {
"tokenizer": "keyword",
"filter": [
"full_pinyin_filter",
"lowercase"
]
}
}
}
}
}
新建
char_filter_text.txt存于ES_HOME/config目录下
新建索引
goods_v1 将使用上述模板（goods*），可直接使用模板中定义的 analyzer
DELETE goods_v1
PUT /goods_v1
{
"mappings": {
"doc":{
"properties" : {
"id":{
"type":"long"
},
"name" : {
"type": "text",
"analyzer": "ikIndexAnalyzer",
"fields": {
"ngram": {
"type": "text",
"analyzer": "ngramIndexAnalyzer"
},
"SPY": {
"type": "text",
"analyzer": "simplePinyinIndexAnalyzer"
},
"FPY": {
"type": "text",
"analyzer": "fullPinyinIndexAnalyzer"
}
}
},
"update_time" : {
"type":"date"
},
"deleted" : {
"type":"boolean"
}
}
}
}
}
PUT /goods_v1/_alias/goods
{
"is_write_index":true
}
filter&&结果id倒序
本搜索只做简单过滤，并且按id倒序返回，不需要计算相似度
GET goods/_search
{
"query": {
"constant_score": {
"filter": {
"bool": {
"filter": [
{
"term": {
"deleted": {
"value": false,
"boost": 1
}
}
},
{
"bool": {
"should": [
{
"match": {
"name.ngram": {
"query": "水果",
"operator": "OR",
"analyzer": "ngramSearchAnalyzer",
"prefix_length": 0,
"max_expansions": 50,
"fuzzy_transpositions": true,
"lenient": false,
"zero_terms_query": "NONE",
"auto_generate_synonyms_phrase_query": true,
"boost": 5
}
}
},
{
"term": {
"name.SPY": {
"value": "水果",
"boost": 1
}
}
},
{
"wildcard": {
"name.SPY": {
"wildcard": "*水果*",
"boost": 0.8
}
}
},
{
"match_phrase": {
"name.FPY": {
"query": "水果",
"analyzer": "fullPinyinSearchAnalyzer",
"slop": 0,
"zero_terms_query": "NONE",
"boost": 1
}
}
},
{
"match": {
"name": {
"query": "水果",
"operator": "OR",
"analyzer": "ikSearchAnalyzer",
"prefix_length": 0,
"max_expansions": 50,
"minimum_should_match": "100%",
"fuzzy_transpositions": true,
"lenient": false,
"zero_terms_query": "NONE",
"auto_generate_synonyms_phrase_query": true,
"boost": 1
}
}
}
],
"adjust_pure_negative": true,
"boost": 1
}
}
],
"adjust_pure_negative": true,
"boost": 1
}
},
"boost": 1
}
},
"sort": [
{
"id": {
"order": "desc"
}
}
]
}
根据score返回结果
多个query结果，取相似度score最大的返回
GET goods/_search
{
"query": {
"bool": {
"must": [
{
"dis_max": {
"tie_breaker": 0,
"queries": [
{
"match": {
"name.ngram": {
"query": "水果",
"operator": "OR",
"analyzer": "ngramSearchAnalyzer",
"prefix_length": 0,
"max_expansions": 50,
"fuzzy_transpositions": true,
"lenient": false,
"zero_terms_query": "NONE",
"auto_generate_synonyms_phrase_query": true,
"boost": 5
}
}
},
{
"term": {
"name.SPY": {
"value": "水果",
"boost": 1
}
}
},
{
"wildcard": {
"name.SPY": {
"wildcard": "*水果*",
"boost": 0.8
}
}
},
{
"match_phrase": {
"name.FPY": {
"query": "水果",
"analyzer": "fullPinyinSearchAnalyzer",
"slop": 0,
"zero_terms_query": "NONE",
"boost": 1
}
}
},
{
"match": {
"name": {
"query": "水果",
"operator": "OR",
"analyzer": "ikSearchAnalyzer",
"prefix_length": 0,
"max_expansions": 50,
"minimum_should_match": "100%",
"fuzzy_transpositions": true,
"lenient": false,
"zero_terms_query": "NONE",
"auto_generate_synonyms_phrase_query": true,
"boost": 1
}
}
}
],
"boost": 1
}
}
],
"filter": [
{
"term": {
"deleted": {
"value": false,
"boost": 1
}
}
}
],
"adjust_pure_negative": true,
"boost": 1
}
}
}
部分相关java代码
public class SearchRequest {
public int page;
public int size;
}
public class GoodsRequest extends SearchRequest {
public String keyword;
}
private BoolQueryBuilder addFilter(QueryBuilder baseQueryBuilder, GoodsRequest request) {
BoolQueryBuilder boolQueryBuilder = getBaseFilterQueryBuilder(request);
if (baseQueryBuilder != null) {
boolQueryBuilder.must(baseQueryBuilder);
}
return boolQueryBuilder;
}
private BoolQueryBuilder getBaseFilterQueryBuilder() {
BoolQueryBuilder boolQueryBuilder = new BoolQueryBuilder();
boolQueryBuilder.filter(QueryBuilders.termQuery("deleted", false));
return boolQueryBuilder;
}
private BoolQueryBuilder getKeywordQueryBuilder(String keyword) {
if (StringUtils.isEmpty(keyword)) {
return null;
}
BoolQueryBuilder keywordQueryBuilder = QueryBuilders.boolQuery();
QueryBuilder ngramSearchBuilder = QueryBuilders
.matchQuery("name.ngram", keyword)
.analyzer("ngramSearchAnalyzer")
.boost(5f);
keywordQueryBuilder.should(ngramSearchBuilder);
TermQueryBuilder simplePinYinQueryBuilder = QueryBuilders.termQuery("name.SPY", keyword);
keywordQueryBuilder.should(simplePinYinQueryBuilder);
QueryBuilder simplePinYinContainQueryBuilder = QueryBuilders
.wildcardQuery("name.SPY", "*" + keyword + "*")
.boost(0.8f);
keywordQueryBuilder.should(simplePinYinContainQueryBuilder);
QueryBuilder fullPingYinQueryBuilder = QueryBuilders
.matchPhraseQuery("name.FPY", keyword)
.analyzer("fullPinyinSearchAnalyzer");
keywordQueryBuilder.should(fullPingYinQueryBuilder);
QueryBuilder containSearchBuilder = QueryBuilders
.matchQuery("name", keyword)
.analyzer("ikSearchAnalyzer")
.minimumShouldMatch("100%");
keywordQueryBuilder.should(containSearchBuilder);
return keywordQueryBuilder;
}
// 返回满足搜索条件的结果，按id倒序
private SearchSourceBuilder getSearchBuilder(GoodsRequest request) {
BoolQueryBuilder boolQueryBuilder =getBaseFilterQueryBuilder();
BoolQueryBuilder keywordQueryBuilder = getKeywordQueryBuilder(request.keyword);
if (keywordQueryBuilder != null) {
boolQueryBuilder.filter(keywordQueryBuilder);
}
SearchSourceBuilder searchSourceBuilder = new SearchSourceBuilder();
searchSourceBuilder.query(QueryBuilders.constantScoreQuery(boolQueryBuilder));
searchSourceBuilder.sort(SortBuilders.fieldSort("id").order(SortOrder.DESC));
return searchSourceBuilder;
}
// 按相关度返回结果
private SearchSourceBuilder getSearchBuilderWithScore() {
//使用dis_max直接取多个query中，分数最高的那一部分结果返回
DisMaxQueryBuilder disMaxQueryBuilder = QueryBuilders.disMaxQuery();
QueryBuilder ngramSearchBuilder = QueryBuilders
.matchQuery("name.ngram", request.keyword)
.analyzer("ngramSearchAnalyzer")
.boost(5f);
disMaxQueryBuilder.add(ngramSearchBuilder);
TermQueryBuilder simplePinYinQueryBuilder = QueryBuilders.termQuery("name.SPY", request.keyword);
disMaxQueryBuilder.add(simplePinYinQueryBuilder);
QueryBuilder simplePinYinContainQueryBuilder = QueryBuilders
.wildcardQuery("name.SPY", "*" + request.keyword + "*")
.boost(0.8f);
disMaxQueryBuilder.add(simplePinYinContainQueryBuilder);
QueryBuilder fullPingYinQueryBuilder = QueryBuilders
.matchPhraseQuery("name.FPY", request.keyword)
.analyzer("fullPinyinSearchAnalyzer");
disMaxQueryBuilder.add(fullPingYinQueryBuilder);
QueryBuilder containSearchBuilder = QueryBuilders
.matchQuery("name", request.keyword)
.analyzer("ikSearchAnalyzer")
.minimumShouldMatch("100%");
disMaxQueryBuilder.add(containSearchBuilder);
SearchSourceBuilder searchSourceBuilder = new SearchSourceBuilder();
searchSourceBuilder.query(addFilter(disMaxQueryBuilder, request));
return searchSourceBuilder;
}
遇到问题
---未完待续---
