Rank 字段类型 #

下表列出了 Easysearch 支持的所有 rank 字段类型。

字段数据类型	描述
`rank_feature`	提升或降低文档的相关性得分。
`rank_features`	提升或降低文档的相关性得分。用于特征列表稀疏的情况。

Rank feature 和 rank features 字段只能使用 rank feature 查询进行查询。它们不支持聚合或排序。

Rank feature 字段类型 #

Rank feature 字段类型使用正浮点值来提升或降低文档在 rank_feature 查询中的相关性得分。默认情况下，该值会提升相关性得分。要降低相关性得分，请将可选参数 positive_score_impact 设置为 false。

示例 #

创建一个包含 rank feature 字段的映射：

PUT chessplayers
{
  "mappings": {
    "properties": {
      "name": {
        "type": "text"
      },
      "rating": {
        "type": "rank_feature"
      },
      "age": {
        "type": "rank_feature",
        "positive_score_impact": false
      }
    }
  }
}

索引三个文档，其中包含一个提升得分的 rank_feature 字段（rating）和一个降低得分的 rank_feature 字段（age）：

PUT chessplayers/_doc/1
{
  "name": "John Doe",
  "rating": 2554,
  "age": 75
}

PUT chessplayers/_doc/2
{
  "name": "Kwaku Mensah",
  "rating": 2067,
  "age": 10
}

PUT chessplayers/_doc/3
{
  "name": "Nikki Wolf",
  "rating": 1864,
  "age": 22
}

Rank feature 查询 #

使用 rank feature 查询，您可以按评分、年龄或同时按评分和年龄对选手进行排名。如果按评分排名，评分较高的选手将获得更高的相关性得分。如果按年龄排名，年龄较小的选手将获得更高的相关性得分。

使用 rank feature 查询根据年龄和评分搜索选手：

GET chessplayers/_search
{
  "query": {
    "bool": {
      "should": [
        {
          "rank_feature": {
            "field": "rating"
          }
        },
        {
          "rank_feature": {
            "field": "age"
          }
        }
      ]
    }
  }
}

当同时按年龄和评分排名时，年龄较小和评分较高的选手得分更好：

{
  "took": 2,
  "timed_out": false,
  "_shards": {
    "total": 1,
    "successful": 1,
    "skipped": 0,
    "failed": 0
  },
  "hits": {
    "total": {
      "value": 3,
      "relation": "eq"
    },
    "max_score": 1.2093145,
    "hits": [
      {
        "_index": "chessplayers",
        "_type": "_doc",
        "_id": "2",
        "_score": 1.2093145,
        "_source": {
          "name": "Kwaku Mensah",
          "rating": 1967,
          "age": 10
        }
      },
      {
        "_index": "chessplayers",
        "_type": "_doc",
        "_id": "3",
        "_score": 1.0150313,
        "_source": {
          "name": "Nikki Wolf",
          "rating": 1864,
          "age": 22
        }
      },
      {
        "_index": "chessplayers",
        "_type": "_doc",
        "_id": "1",
        "_score": 0.8098284,
        "_source": {
          "name": "John Doe",
          "rating": 2554,
          "age": 75
        }
      }
    ]
  }
}

Rank features 字段类型 #

Rank features 字段类型与 rank feature 字段类型类似，但它更适合用于稀疏特征列表。Rank features 字段可以索引数值特征向量，这些向量后续用于在 rank_feature 查询中提升或降低文档的相关性得分。

示例 #

创建一个包含 rank features 字段的映射：

PUT testindex1
{
  "mappings": {
    "properties": {
      "correlations": {
        "type": "rank_features"
      }
    }
  }
}

要索引包含 rank features 字段的文档，请使用带有字符串键和正浮点值的哈希映射：

PUT testindex1/_doc/1
{
  "correlations": {
    "young kids": 1,
    "older kids": 15,
    "teens": 25.9
  }
}

PUT testindex1/_doc/2
{
  "correlations": {
    "teens": 10,
    "adults": 95.7
  }
}

使用 rank feature 查询检索文档：

GET testindex1/_search
{
  "query": {
    "rank_feature": {
      "field": "correlations.teens"
    }
  }
}

响应按相关性得分排序：

{
  "took": 123,
  "timed_out": false,
  "_shards": {
    "total": 1,
    "successful": 1,
    "skipped": 0,
    "failed": 0
  },
  "hits": {
    "total": {
      "value": 2,
      "relation": "eq"
    },
    "max_score": 0.6258503,
    "hits": [
      {
        "_index": "testindex1",
        "_type": "_doc",
        "_id": "1",
        "_score": 0.6258503,
        "_source": {
          "correlations": {
            "young kids": 1,
            "older kids": 15,
            "teens": 25.9
          }
        }
      },
      {
        "_index": "testindex1",
        "_type": "_doc",
        "_id": "2",
        "_score": 0.39263803,
        "_source": {
          "correlations": {
            "teens": 10,
            "adults": 95.7
          }
        }
      }
    ]
  }
}

Rank feature 和 rank features 字段使用前 9 个有效位来保证精度，导致大约 0.4% 的相对误差。值的存储相对精度为 2^−8 = 0.00390625。