首页 文章

模糊匹配得分高于完全匹配

提问于
浏览 次
0

我是ElasticSearch的新手，并试图配置Elasticsearch给我模糊匹配 . 在实现模糊搜索，自动完成过滤器和带状疱疹时，精确匹配似乎具有比部分匹配更低的分数 . 例如，如果查询是“Ring”，则它似乎与“Brass Ring”而不是“Ring”具有更高的匹配 .

谁能帮我吗？

以下是我制作索引的方法：

itemindex = es.indices.create(
        index='mo-items-index-1',
        body={
        "settings": {
            "number_of_shards": 1,
            "analysis": {
                "filter": {
                    "autocomplete_filter": {
                        "type":     "edge_ngram",
                        "min_gram": 1,
                        "max_gram": 20
                    },
                    "custom_shingle": {
                        "type": "shingle",
                        "min_shingle_size": 2,
                        "max_shingle_size": 3,
                        "output_unigrams": True

                    },
                    "my_char_filter": {
                        "type": "pattern_replace",
                        "pattern": " ",
                        "replacement": ""
                    }
                },
                "analyzer": {
                    "autocomplete": {
                        "type":      "custom",
                        "tokenizer": "standard",
                        "filter": [
                            "lowercase",
                            "custom_shingle",
                            "autocomplete_filter",
                            "my_char_filter"
                        ]
                    }
                }
            }
        },
        "mappings": {
        "my_type": {
            "properties": {
                "item_id": {
                    "type":     "string",
                    "analyzer": "autocomplete",
            "search_analyzer": "standard"

                },
            "item_name": {
                    "type":     "string",
                    "analyzer": "autocomplete",
            "search_analyzer": "standard"

                }
            }
        }
    }
       },
        # Will ignore 400 errors, remove to ensure you're prompted
       ignore=400
    )

以下是我查询术语的方法：

res2 = es.search(index="mo-items-index-1", size=200, body={"query": {"multi_match": {
        "fields": [
            "item_name", "item_id"], "query": userQuery, "fuzziness": "AUTO"}}, "highlight": {

        "fields": {
            "item_name": {},
            "item_id": {}

        }
    }, })
elasticsearch boto3 elasticsearch-5

1 回答

  • 1

    有一个非常简单的方法可以"boost"完全匹配的得分：使用 bool 查询将使用您已经存在的查询和 term 一个 should 语句：

    "query": {
    "bool": {
      "should": [
        {
          "multi_match": {
            "fields": [
              "item_name",
              "item_id"
            ],
            "query": "Ring",
            "fuzziness": "AUTO"
          }
        },
        {
          "term": {
            "item_name.keyword": {
              "value": "Ring"
            }
          }
        }
      ]
    }
  }

    而且你还需要在字段中添加一个 keyword 类型的子字段，以支持完美匹配：

    "mappings": {
    "my_type": {
      "properties": {
        "item_id": {
          "type": "string",
          "analyzer": "autocomplete",
          "search_analyzer": "standard"
        },
        "item_name": {
          "type": "string",
          "analyzer": "autocomplete",
          "search_analyzer": "standard",
          "fields": {
            "keyword": {
              "type": "keyword"
            }
          }
        }
      }
    }
  }
    回复于

相关问题