elasticsearch Sankt和St的同义词

我正在尝试让同义词在我现有的设置中工作。目前我有以下设置：

PUT city
{
    "settings": {
        "analysis": {
            "analyzer": {
                "autocomplete": {
                    "tokenizer": "autocomplete",
                    "filter": [
                        "lowercase",
                        "my_synonym_filter",
                        "german_normalization",
                        "my_ascii_folding"
                    ]
                },
                "autocomplete_search": {
                    "tokenizer": "lowercase",
                    "filter": [
                        "lowercase",
                        "my_synonym_filter",
                        "german_normalization",
                        "my_ascii_folding"
                    ]
                }
            },
                  "filter": {
                     "my_ascii_folding": {
                     "type": "asciifolding",
                     "preserve_original": true
            },
                  "my_synonym_filter": {
                  "type": "synonym",
                  "ignore_case": "true",
                  "synonyms": [
                     "sankt, st => sankt"
                  ]
            }
          },
            "tokenizer": {
                "autocomplete": {
                    "type": "edge_ngram",
                    "min_gram": 1,
                    "max_gram": 15,
                    "token_chars": [
                        "letter",
                        "digit",
                        "symbol"
                    ]
                }
            }
        }
    },
    "mappings": {
        "city": {
            "properties": {
                "name": {
                    "type": "text",
                    "analyzer": "autocomplete",
                    "search_analyzer": "autocomplete_search"
                }
            }
        }
    }
}

在这个城市索引中，我有这样的文档：

圣沃尔夫冈或圣沃尔夫冈等。对我来说，圣与圣是同义词。因此，如果我搜索Sankt，两个文档都应该出现

我创建了一个新过滤器，并将该过滤器添加到我的自动完成分析器中：

现在很好。但我面临的问题如下：

很明显，st后面的点目前没有被分析和搜索。但对于同义词，点很重要

第二个问题是，如果我搜索sankt，同义词是st，它会给我所有以st开头的文档，比如斯图加特。这也是因为没有使用圆点

你知道我怎样才能做到吗？如果你需要更多的信息，请告诉我

更新：

讨论后，我在我的设置中进行了以下更改：

PUT city
{
    "settings": {
        "analysis": {
            "analyzer": {
                "autocomplete": {
                    "tokenizer": "autocomplete",
                    "filter": [
                        "lowercase",
                        "my_synonym_filter",
                        "german_normalization",
                        "my_ascii_folding"
                    ]
                },
                "autocomplete_search": {
                    "tokenizer": "lowercase",
                    "filter": [
                        "lowercase",
                        "my_synonym_filter",
                        "german_normalization",
                        "my_ascii_folding"
                    ]
                }
            },
                  "filter": {
                     "my_ascii_folding": {
                     "type": "asciifolding",
                     "preserve_original": true
            },
                  "my_synonym_filter": {
                  "type": "synonym",
                  "ignore_case": "true",
                  "synonyms": [
                     "sankt, st => sankt"
                  ]
            }
          },
            "tokenizer": {
                "autocomplete": {
                    "type": "edge_ngram",
                    "min_gram": 1,
                    "max_gram": 15,
                    "token_chars": [
                        "letter",
                        "digit",
                        "symbol"
                    ]
                }
            }
        }
    },
    "mappings": {
        "city": {
            "properties": {
                "name": {
                    "type": "text",
                    "analyzer": "autocomplete",
                    "search_analyzer": "autocomplete_search"
                }
            }
        }
    }
}

将边缘内存标记器更改为标准标记器

添加了Edengram过滤器，并将此过滤器添加到我的分析器中

从我的分析器中删除了过滤器德语标准化和我的ascii折叠，以简化测试

PUT city
{
  "settings": {
    "analysis": {
      "analyzer": {
        "autocomplete": {
          "tokenizer": "autocomplete",
          "filter": [
            "lowercase",
            "my_synonym_filter",
            "edge_filter"
          ]
        },
        "autocomplete_search": {
          "tokenizer": "autocomplete",
          "filter": [
            "my_synonym_filter",
            "lowercase"
          ]
        }
      },
      "filter": {
        "edge_filter": {
          "type": "edgeNGram",
          "min_gram": 1,
          "max_gram": 15
        },
        "my_synonym_filter": {
          "type": "synonym",
          "ignore_case": "true",
          "synonyms": [
            "sankt, st => sankt"
          ]
        }
      },
      "tokenizer": {
        "autocomplete": {
          "type": "standard"
        }
      }
    }
  },
  "mappings": {
    "city": {
      "properties": {
        "name": {
          "type": "text",
          "analyzer": "autocomplete",
          "search_analyzer": "autocomplete_search"
        }
      }
    }
  }
}

我将这3个文档添加到索引中：

"name":"Sankt Wolfgang",
"name":"Stuttgart",
"name":"St. Wolfgang"

PUT city
{
  "settings": {
    "analysis": {
      "analyzer": {
        "autocomplete": {
          "tokenizer": "standard",
          "filter": [
            "lowercase",
            "my_synonym_filter",
            "edge_filter"
          ]
        },
        "autocomplete_search": {
          "tokenizer": "standard",
          "filter": [
            "lowercase",
            "my_synonym_filter"
          ]
        }
      },
      "filter": {
        "edge_filter": {
          "type": "edgeNGram",
          "min_gram": 1,
          "max_gram": 15
        },
        "my_synonym_filter": {
          "type": "synonym",
          "ignore_case": "true",
          "synonyms": [
            "sankt, st. => sankt"
          ]
        }
      }
    }
  },
  "mappings": {
    "city": {
      "properties": {
        "name": {
          "type": "text",
          "analyzer": "autocomplete",
          "search_analyzer": "autocomplete_search"
        }
      }
    }
  }
}

查询字符串-结果

st      ->    "St. Wolfgang", "Stuttgart"
st.     ->    "St. Wolfgang", "Sankt Wolfgang"
sankt   ->    "St. Wolfgang", "Sankt Wolfgang"

---

这对我来说很有效。这里的要点是确保

将同义词筛选器放在小写筛选器之后 将edge-n-gram过滤器放在末端 仅在索引时使用edge-n-gram 因此，我们创建索引：

"name":"Sankt Wolfgang",
"name":"Stuttgart",
"name":"St. Wolfgang"

PUT city
{
  "settings": {
    "analysis": {
      "analyzer": {
        "autocomplete": {
          "tokenizer": "standard",
          "filter": [
            "lowercase",
            "my_synonym_filter",
            "edge_filter"
          ]
        },
        "autocomplete_search": {
          "tokenizer": "standard",
          "filter": [
            "lowercase",
            "my_synonym_filter"
          ]
        }
      },
      "filter": {
        "edge_filter": {
          "type": "edgeNGram",
          "min_gram": 1,
          "max_gram": 15
        },
        "my_synonym_filter": {
          "type": "synonym",
          "ignore_case": "true",
          "synonyms": [
            "sankt, st. => sankt"
          ]
        }
      }
    }
  },
  "mappings": {
    "city": {
      "properties": {
        "name": {
          "type": "text",
          "analyzer": "autocomplete",
          "search_analyzer": "autocomplete_search"
        }
      }
    }
  }
}

然后我们索引数据：

PUT city/city/1
{
  "name":"St. Wolfgang"
}
PUT city/city/2
{
  "name":"Stuttgart"
}
PUT city/city/3
{
  "name":"Sankt Wolfgang"
}

最后，搜索st或sankt将只返回文档1和3，而不是2

POST city/_search?q=name:st
POST city/_search?q=name:sankt

---

您是否可以尝试更改sankt的同义词，st=>sankt，即st将被索引为sankt，因此搜索sankt将返回sankt，搜索st也应仅匹配sankt。你能试一试吗？@Val我不能通过更改同义词来获取任何文档。真奇怪。你还有什么其他办法可以让它工作吗？你能用同义词标记过滤器更新你的设置吗？这样我就可以复制它了？哦，你实际上还需要在搜索时间分析器中添加同义词标记过滤器，这样键入st的人也可以在引擎盖下搜索sankt。@Val我编辑了设置。这正是我所用的。是的，我在搜索分析器中添加了同义词过滤器。如果您还需要索引或其他任何数据，请让我知道。谢谢你的回答。我复制了你的索引。但对我来说，当我将同义词改为sankt时，sankt只给出doc1和doc3，st=>sankt不带点。而对于查询，圣斯图加特也应该是一个匹配。对于圣斯图加特来说，这已经不是一个选择了。让我们把这个颠倒过来。你能为每个搜索输入制定出你想要的输出吗？当然。我用查询字符串和结果更新了我的问题。你能用标记化的空白代替标准的空白吗？太棒了，很高兴我们找到了答案！