Azure托管Elastic Cloud中如何维护多份同义词文件作为token filter

步骤1：准备同义词文件 #

按Elasticsearch标准格式准备同义词文件，比如synonyms_brand.txt、synonyms_category.txt，每行定义一组同义词：

# synonyms_brand.txt
apple, iphone, macbook
google, pixel, android

# synonyms_category.txt
laptop, notebook
smartphone, mobile phone

步骤2：上传同义词文件到集群 #

Azure托管的Elastic Cloud操作逻辑与原生Elastic Cloud一致，可通过Kibana Dev Tools或curl命令上传：

用Kibana Dev Tools上传

PUT _synonyms/synonyms_brand.txt
{
  "synonyms": [
    "apple, iphone, macbook",
    "google, pixel, android"
  ]
}

PUT _synonyms/synonyms_category.txt
{
  "synonyms": [
    "laptop, notebook",
    "smartphone, mobile phone"
  ]
}

用curl命令上传本地文件

# 替换为你的集群地址、认证信息和本地文件路径
curl -X PUT "https://your-elastic-cluster.azure.elastic-cloud.com/_synonyms/synonyms_brand.txt?pretty" \
  -u "username:password" \
  -H "Content-Type: text/plain" \
  --data-binary @./synonyms_brand.txt

步骤3：配置索引的分词器与同义词Filter #

创建新索引或更新已有索引的设置，在analysis.filter中定义多个同义词Filter，分别绑定不同的同义词文件：

创建新索引

PUT /product_catalog
{
  "settings": {
    "analysis": {
      "filter": {
        "synonym_brand": {
          "type": "synonym",
          "synonyms_path": "_synonyms/synonyms_brand.txt"
        },
        "synonym_category": {
          "type": "synonym",
          "synonyms_path": "_synonyms/synonyms_category.txt"
        }
      },
      "analyzer": {
        "product_search_analyzer": {
          "tokenizer": "standard",
          "filter": [
            "lowercase",
            "synonym_brand",
            "synonym_category"
          ]
        }
      }
    }
  },
  "mappings": {
    "properties": {
      "product_name": {
        "type": "text",
        "analyzer": "product_search_analyzer"
      }
    }
  }
}

更新已有索引

已有索引需先关闭，更新设置后再重新打开：

# 关闭索引
POST /product_catalog/_close

# 更新分词器配置
PUT /product_catalog/_settings
{
  "analysis": {
    "filter": {
      "synonym_brand": {
        "type": "synonym",
        "synonyms_path": "_synonyms/synonyms_brand.txt"
      },
      "synonym_category": {
        "type": "synonym",
        "synonyms_path": "_synonyms/synonyms_category.txt"
      }
    },
    "analyzer": {
      "product_search_analyzer": {
        "tokenizer": "standard",
        "filter": [
          "lowercase",
          "synonym_brand",
          "synonym_category"
        ]
      }
    }
  }
}

# 重新打开索引
POST /product_catalog/_open

步骤4：验证同义词效果 #

用_analyze API测试分词器是否生效：

POST /product_catalog/_analyze
{
  "analyzer": "product_search_analyzer",
  "text": "iphone"
}

返回结果中应包含apple、macbook等同义词token。

关键注意事项 #

• 同义词文件必须存储在_synonyms/路径下，这是Elastic Cloud托管环境的专用全局路径，所有索引均可引用。
• 修改同义词文件后，无需重启索引，调用重载API即可生效：

  POST _synonyms/synonyms_brand.txt/_reload
  

• 确保文件名唯一，避免不同业务的同义词文件冲突。

内容的提问来源于stack exchange，提问作者Deery