ElasticSearch 距离排序模型

• 验证按时间的增量是否能跟上去

  GET /shop/_search
  {
  "query": {
    "match_phrase": {
      "name": "和府捞面(正大乐城店)"
    }
  },
  "sort": [
    {
      "id": {
        "order": "asc"
      }
    }
  ]
  }
  

• 期望搜索出凯悦酒店，但是出现了 花悦庭果木烤鸭

GET /shop/_search
{
  "query": {
    "match": {
      "name": "凯悦"
    }
  }
}

GET /shop/_analyze
{
  "analyzer": "ik_smart", //分为 凯 悦 
  "text": "凯悦"  
}

GET /shop/_analyze
{
  "analyzer": "ik_max_word",  // 含有 悦 
  "text": "花悦庭果木烤鸭"  
}

• 带上距离字段

  GET /shop/_search
  {
  "query": {
    "match": {
      "name": "凯悦"
    }
  },
  "_source": "*", // 不加则只有 distance 字段展示在结果中
  "script_fields": {
    "distance": { // 新增定义
      "script": {
        "source": "haversin(lat,lon,doc['location'].lat,doc['location'].lon)", //半正失曲线?
        "lang": "expression",
        "params": {"lat":31.37,"lon":127.12}
      }
    }
  }
  }
  

• 使用距离排序

  GET /shop/_search
  {
  "query": {
    "match": {
      "name": "凯悦"
    }
  },
  "_source": "*", // 不加则只有 distance 字段展示在结果中
  "script_fields": {
    "distance": { // 新增定义
      "script": {
        "source": "haversin(lat,lon,doc['location'].lat,doc['location'].lon)", //半正失曲线?
        "lang": "expression",
        "params": {"lat":31.37,"lon":127.12}
      }
    }
  },
  "sort": [ // 导致"_score" : null, 
    {
      "_geo_distance": {
        "location": {
          "lat": 31.37,
          "lon": 127.12
        },
        "unit": "km",
        "distance_type": "arc" //球形
      }
    }
  ]
  }
  

• 使用function score解决排序模型 【搜索最终版本】

  GET /shop/_search
  {
  //"explain": true,  //最后再使用explain调优 
  "_source": "*",
  "script_fields": {
     "distance": { // 新增定义
      "script": {
        "source": "haversin(lat,lon,doc['location'].lat,doc['location'].lon)", //半正失曲线?
        "lang": "expression",
        "params": {"lat":31.23916171,"lon":121.48789949}
      }
    }
  },
  "query": {
    "function_score": {
      "query": {
        "bool": {
          "must": [
            {"match": {"name":{"query": "凯悦","boost": 0.1}}}, //boost文本再降权0.1，花悦庭果木烤鸭被拉到了第二名
            {"term":{"seller_disabled_flag": 0}}, //精确匹配
            {"term":{"category_id": 2}}
          ]
        }
      },
      "functions": [
        {
          "gauss":{
            "location":{
              "origin":"31.23916171,121.48789949",
              "scale":"100km", //衰减
              "offset":"0km", //起始点
              "decay":0.5 
            }
          },
          "weight": 9 //距离对于评价来说，跟重要些
        },
        {
          "field_value_factor": {
            "field": "remark_score" //1-5分,归一化到 0-1 之间，所以设置 weight：0.2
          },
          "weight": 0.2
        },
        {
          "field_value_factor": {
            "field": "seller_remark_score"  //1-5分,归一化到 0-1 之间，所以设置 weight：0.2，但是门店评分比商家评分更重要，所以设置 商家评分seller_remark_score weight : 0.1
          },
          "weight": 0.1
        }
      ],
      "score_mode": "sum",
      "boost_mode": "sum" //replace 不考虑match文本相关性，只考虑距离和评分。本项目搜索使用replace。
    }
  },
  "sort": [
    {
      "_score": { //最后按打分排序
        "order": "desc"
      }
    }
  ]
  }
  

  
  
  

• 重新定义门店索引结构，支持同义词

  PUT /shop
  {
  "settings": {
    "number_of_shards": 1, 
    "number_of_replicas": 0, //根据实际集群情况调整，此时是单节点，所以使用0
    "analysis":{ //支持同义词
      "filter":{
        "my_synonym_fiter":{ //随便取名
          "type":"synonym",
          "synonyms_path":"analysis-ik/synonyms.dic" //相对esnode/config/路径
        }
      },
      "analyzer":{
        "ik_syno":{ //自定义后要去修改下面 name 的分词器
          "type":"custom",
          "tokenizer":"ik_smart",
          "filter":["my_synonym_fiter"] 
        },
        "ik_syno_max":{
          "type":"custom",
          "tokenizer":"ik_max_word",
          "filter":["my_synonym_fiter"] 
        }
      }
    }
  },
  "mappings": {
    "properties": {
      "id":{"type": "integer"}, 
      "name":{"type": "text","analyzer": "ik_syno_max","search_analyzer": "ik_syno"}, //分词器的最佳实践，使用自定义的分词器支持同义词
      "tags":{"type": "text","analyzer": "whitespace","fielddata": true},//多个标签以空格分开；fielddata表示分词后数量的聚合，1变多，比较耗内存
      "location":{"type": "geo_point"},
      "remark_score":{"type": "double"},
      "price_per_man":{"type": "integer"},
      "category_id":{"type": "integer"},
      "category_name":{"type": "keyword"},//keyword 不分词，完全匹配
      "seller_id":{"type": "integer"},
      "seller_remark_score":{"type": "double"},
      "seller_disabled_flag":{"type": "integer"}
    }
  }
  }
  

• 验证同义词分析

  GET /shop/_analyze
  {
  "field": "name", 
  "text": "烤鸭"  //分成 烤鸭 能吃 的，同时分析"能吃的"，也能得到3个相同分析结果，但是分析"能吃"，仅能得到"能吃"
  }