如何在elasticsearch上的列表中有完全匹配时提高分数？

uqdfh47h  于 2023-11-17  发布在  ElasticSearch
关注(0)|答案(2)|浏览(225)

我对ElasticSearch很陌生，我下面有这个问题。
有这两个记录：

  1. POST test/_doc/1
  2.     {
  3.       "id": 1,
  4.       "authors": [
  5.         {
  6.           "name": "Test Name",
  7.           "url": "/url/1/"
  8.         }
  9.       ]
  10.     }
  12.     POST test/_doc/2
  13.     {
  14.       "id": 2,
  15.       "authors": [
  16.         {
  17.           "name": "Test Name",
  18.           "url": "/url/1/"
  19.         },
  20.             {
  21.           "name": "Another author",
  22.           "url": "/url/another/"
  23.         }
  24.       ]
  25.     }

字符串
这个查询：

  1. GET test/_search
  2.     {
  3.       "query": {
  4.         "function_score": {
  5.           "query": {
  6.             "bool": {
  7.               "should": [
  8.                 {
  9.                   "match_phrase": {
  10.                     "authors.name": {
  11.                       "_name": "exact match in authors",
  12.                       "query": "Test Name",
  13.                       "boost": 100,
  14.                       "slop": 1
  15.                     }
  16.                   }
  17.                 }
  18.               ]
  19.             }
  20.           }
  21.         }
  22.       }
  23.     }

型
为什么当有多个作者时，分数会降低？我如何才能使它更高或与只有一个作者的记录相同？

  1. {
  2.       ...
  3.       "hits": {
  4.         "max_score": 42.221836,
  5.         "hits": [
  6.           {
  7.             "_score": 42.221836,
  8.             "_source": {
  9.               "id": 1,
  10.               "authors": [
  11.                 {
  12.                   "name": "Test Name",
  13.                   "url": "/url/1/"
  14.                 }
  15.               ]
  16.             },
  17.             "matched_queries": [
  18.               "exact match in authors"
  19.             ]
  20.           },
  21.           {
  22.             "_score": 32.088596,
  23.             "_source": {
  24.               "id": 2,
  25.               "authors": [
  26.                 {
  27.                   "name": "Test Name",
  28.                   "url": "/url/1/"
  29.                 },
  30.                 {
  31.                   "name": "Another author",
  32.                   "url": "/url/another/"
  33.                 }
  34.               ]
  35.             },
  36.             "matched_queries": [
  37.               "exact match in authors"
  38.             ]
  39.           }
  40.         ]
  41.       }
  42.     }

型
我在文件上找不到任何关于这个的东西。
下面的详细信息只是为了确保stackoverflow不会显示以下错误：It looks like your post is mostly code; please add some more details.

elasticsearch

2条答案

按热度按时间
kqqjbcuj

kqqjbcuj1#

TLDR;

这是因为你的第二个文件有一个较长的字段。你可能不习惯看：

  • 恒定计分
  • 滤波器
  • 功能评分

去理解

这是什么意思？
Elasticsearch在处理一个文档数组时，会像这样存储它们：
最初：

  1. {
  2.   "authors": [
  3.     {
  4.       "name": "A0"
  5.     },
  6.         {
  7.       "name": "A1"
  8.     }
  9.   ]
  10. }

字符串
收件人：

  1. {
  2.   "authors.name": ["A0", "A1"]
  3. }

型
而文档得分的计算采用TF/IDF，但TF与文档长度有关。

  • 文档% 1 authors.name的长度为% 2
  • Doc 2 authors.name的长度为4

调查：

你可以使用API _explain：

  1. GET 77469343/_explain/1
  2. {
  3.   "query": {
  4.     "bool": {
  5.       "should": [
  6.         {
  7.           "match_phrase": {
  8.             "authors.name": {
  9.               "_name": "exact match in authors",
  10.               "query": "Test Name",
  11.               "boost": 100,
  12.               "slop": 1
  13.             }
  14.           }
  15.         }
  16.       ]
  17.     }
  18.   }
  19. }

型
这将给你给予以下结果：

文档1

  1. {
  2.   "_index": "77469343",
  3.   "_id": "1",
  4.   "matched": true,
  5.   "explanation": {
  6.     "value": 42.221836,
  7.     "description": """weight(authors.name:"test name"~1 in 0) [PerFieldSimilarity], result of:""",
  8.     "details": [
  9.       {
  10.         "value": 42.221836,
  11.         "description": "score(freq=1.0), computed as boost * idf * tf from:",
  12.         "details": [
  13.           {
  14.             "value": 220,
  15.             "description": "boost",
  16.             "details": []
  17.           },
  18.           {
  19.             "value": 0.36464313,
  20.             "description": "idf, sum of:",
  21.             "details": [
  22.               {
  23.                 "value": 0.18232156,
  24.                 "description": "idf, computed as log(1 + (N - n + 0.5) / (n + 0.5)) from:",
  25.                 "details": [
  26.                   {
  27.                     "value": 2,
  28.                     "description": "n, number of documents containing term",
  29.                     "details": []
  30.                   },
  31.                   {
  32.                     "value": 2,
  33.                     "description": "N, total number of documents with field",
  34.                     "details": []
  35.                   }
  36.                 ]
  37.               },
  38.               {
  39.                 "value": 0.18232156,
  40.                 "description": "idf, computed as log(1 + (N - n + 0.5) / (n + 0.5)) from:",
  41.                 "details": [
  42.                   {
  43.                     "value": 2,
  44.                     "description": "n, number of documents containing term",
  45.                     "details": []
  46.                   },
  47.                   {
  48.                     "value": 2,
  49.                     "description": "N, total number of documents with field",
  50.                     "details": []
  51.                   }
  52.                 ]
  53.               }
  54.             ]
  55.           },
  56.           {
  57.             "value": 0.5263158,
  58.             "description": "tf, computed as freq / (freq + k1 * (1 - b + b * dl / avgdl)) from:",
  59.             "details": [
  60.               {
  61.                 "value": 1,
  62.                 "description": "phraseFreq=1.0",
  63.                 "details": []
  64.               },
  65.               {
  66.                 "value": 1.2,
  67.                 "description": "k1, term saturation parameter",
  68.                 "details": []
  69.               },
  70.               {
  71.                 "value": 0.75,
  72.                 "description": "b, length normalization parameter",
  73.                 "details": []
  74.               },
  75.               {
  76.                 "value": 2,
  77.                 "description": "dl, length of field",
  78.                 "details": []
  79.               },
  80.               {
  81.                 "value": 3,
  82.                 "description": "avgdl, average length of field",
  83.                 "details": []
  84.               }
  85.             ]
  86.           }
  87.         ]
  88.       }
  89.     ]
  90.   }
  91. }

型

文档2

  1. {
  2.   "_index": "77469343",
  3.   "_id": "2",
  4.   "matched": true,
  5.   "explanation": {
  6.     "value": 32.088596,
  7.     "description": """weight(authors.name:"test name"~1 in 1) [PerFieldSimilarity], result of:""",
  8.     "details": [
  9.       {
  10.         "value": 32.088596,
  11.         "description": "score(freq=1.0), computed as boost * idf * tf from:",
  12.         "details": [
  13.           {
  14.             "value": 220,
  15.             "description": "boost",
  16.             "details": []
  17.           },
  18.           {
  19.             "value": 0.36464313,
  20.             "description": "idf, sum of:",
  21.             "details": [
  22.               {
  23.                 "value": 0.18232156,
  24.                 "description": "idf, computed as log(1 + (N - n + 0.5) / (n + 0.5)) from:",
  25.                 "details": [
  26.                   {
  27.                     "value": 2,
  28.                     "description": "n, number of documents containing term",
  29.                     "details": []
  30.                   },
  31.                   {
  32.                     "value": 2,
  33.                     "description": "N, total number of documents with field",
  34.                     "details": []
  35.                   }
  36.                 ]
  37.               },
  38.               {
  39.                 "value": 0.18232156,
  40.                 "description": "idf, computed as log(1 + (N - n + 0.5) / (n + 0.5)) from:",
  41.                 "details": [
  42.                   {
  43.                     "value": 2,
  44.                     "description": "n, number of documents containing term",
  45.                     "details": []
  46.                   },
  47.                   {
  48.                     "value": 2,
  49.                     "description": "N, total number of documents with field",
  50.                     "details": []
  51.                   }
  52.                 ]
  53.               }
  54.             ]
  55.           },
  56.           {
  57.             "value": 0.40000004,
  58.             "description": "tf, computed as freq / (freq + k1 * (1 - b + b * dl / avgdl)) from:",
  59.             "details": [
  60.               {
  61.                 "value": 1,
  62.                 "description": "phraseFreq=1.0",
  63.                 "details": []
  64.               },
  65.               {
  66.                 "value": 1.2,
  67.                 "description": "k1, term saturation parameter",
  68.                 "details": []
  69.               },
  70.               {
  71.                 "value": 0.75,
  72.                 "description": "b, length normalization parameter",
  73.                 "details": []
  74.               },
  75.               {
  76.                 "value": 4,
  77.                 "description": "dl, length of field",
  78.                 "details": []
  79.               },
  80.               {
  81.                 "value": 3,
  82.                 "description": "avgdl, average length of field",
  83.                 "details": []
  84.               }
  85.             ]
  86.           }
  87.         ]
  88.       }
  89.     ]
  90.   }
  91. }

型

修复

常量评分

如果你仍然想要一个分数，你可能想看看constant_score查询：

  1. GET 77469343/_search
  2. {
  3.   "query": {
  4.     "bool": {
  5.       "should": [
  6.         {
  7.           "constant_score": {
  8.             "filter": {
  9.               "match_phrase": {
  10.                 "authors.name": {
  11.                   "_name": "exact match in authors",
  12.                   "query": "Test Name",
  13.                   "boost": 100,
  14.                   "slop": 1
  15.                 }
  16.               }
  17.             },
  18.             "boost": 1.2
  19.           }
  20.         }
  21.       ]
  22.     }
  23.   }
  24. }

型

过滤而不是应该？

如果你使用过滤器，匹配的文档不会影响分数：

  1. GET 77469343/_search
  2. {
  3.   "query": {
  4.     "bool": {
  5.       "filter": [
  6.         {
  7.           "match_phrase": {
  8.             "authors.name": {
  9.               "_name": "exact match in authors",
  10.               "query": "Test Name",
  11.               "boost": 100,
  12.               "slop": 1
  13.             }
  14.           }
  15.         }
  16.       ]
  17.     }
  18.   }
  19. }

型

展开查看全部
赞(0）分享  回复(0）举报  2023-11-17
o4tp2gmn

o4tp2gmn2#

我尝试了@paulo解决方案，但它并不完全适合我，所以我最终添加了一个嵌套字段：

  1. "authors": {
  2.         "type": "nested",
  3.         "properties": {
  4.           "name": {
  5.             "type": "text",
  6.             "fields": {
  7.               "keyword": {
  8.                 "type": "keyword",
  9.                 "ignore_above": 256
  10.               }
  11.             }
  12.           },
  13.           "url": {
  14.             "type": "text",
  15.             "fields": {
  16.               "keyword": {
  17.                 "type": "keyword",
  18.                 "ignore_above": 256
  19.               }
  20.             }
  21.           }
  22.         }
  23.       },

字符串
并使用此查询：

  1. {
  2.     "nested": {
  3.         "path": "authors",
  4.         "_name": "exact match in authors",
  5.         "query": {
  6.             "bool": {
  7.                 "must": {
  8.                     "match_phrase": {
  9.                         "authors.name": {
  10.                             "query": "Test Name",
  11.                             "boost": 100,
  12.                             "slop": 1,
  13.                         }
  14.                     }
  15.                 }
  16.             }
  17.         },
  18.     }
  19. }

型
ElasticSearch文档：https://www.elastic.co/guide/en/elasticsearch/reference/current/nested.html
经过这些修改后，它工作得很好！

展开查看全部
赞(0）分享  回复(0）举报  2023-11-17
我来回答

相关问题

    查看更多

    热门标签

    更多
    JavaquerypythonNode开发语言requestUtil数据库Table后端算法LoggerMessageElementParser

    最新问答

    更多
    • 蜀ICP备13028337号-1 大数据知识库 https://www.saoniuhuo.com © All rights reserved
    • 本站内容来源互联网,如果侵犯您的权益请联系我们删除, 联系方式：448109455@qq.com