如何在Elasticsearch中对去HTML标签的结果实现高亮？

问题背景 #

已实现Elasticsearch搜索结果去除HTML标签，但对去标签后的内容执行高亮时，当前高亮字段仍保留原HTML内容，需调整实现目标效果。

现有映射与设置 #

{
  "settings": {
    "analysis": {
      "filter": {
        "my_pattern_replace_filter": {
          "type": "pattern_replace",
          "pattern": "\n",
          "replacement": ""
        }
      },
      "analyzer": {
        "my_analyzer": {
          "tokenizer": "standard",
          "filter": [
            "lowercase"
          ],
          "char_filter": [
            "html_strip"
          ]
        },
        "parsed_analyzer": {
          "type": "custom",
          "tokenizer": "keyword",
          "char_filter": [
            "html_strip"
          ],
          "filter": [
            "my_pattern_replace_filter"
          ]
        }
      }
    }
  },
  "mappings": {
    "properties": {
      "html": {
        "type": "text",
        "analyzer": "my_analyzer",
        "fields": {
          "raw": {
            "type": "text",
            "fielddata": true,
            "analyzer": "parsed_analyzer"
          }
        }
      }
    }
  }
}

当前搜索查询 #

POST idx_test/_search

{
  "script_fields": {
    "raw": {
      "script": "doc['html.raw']"
    }
  }, 
  "query": {
    "match": {
      "html": "more"
    }
  },
  "highlight": {
    "fields": {
      "*":{ "pre_tags" : ["<strong>"], "post_tags" : ["</strong>"] }
    }
  }
}

当前返回结果 #

"hits": [
    {
        "_index": "idx_test2",
        "_type": "_doc",
        "_id": "GijDsYMBjgX3UBaguGxc",
        "_score": 0.2876821,
        "fields": {
            "raw": [
                "Test More test"
            ]
        },
        "highlight": {
            "html": [
                "<html><body><h1=\"font-family: Arial\">Test</h1> <span><strong>More</strong> test</span></body></html>"
            ]
        }
    }
]

期望返回结果 #

"hits": [
    {
        "_index": "idx_test2",
        "_type": "_doc",
        "_id": "GijDsYMBjgX3UBaguGxc",
        "_score": 0.2876821,
        "fields": {
            "raw": [
                "Test <strong>More</strong> test"
            ]
        }
    }
]

解决方案 #

方案1：修改搜索查询，针对处理后字段高亮 #

直接在高亮配置中指定html.raw字段（已去除HTML标签），并调整参数返回完整高亮内容：

POST idx_test/_search
{
  "query": {
    "match": {
      "html": "more"
    }
  },
  "highlight": {
    "fields": {
      "html.raw": {
        "pre_tags": ["<strong>"],
        "post_tags": ["</strong>"],
        "number_of_fragments": 0
      }
    }
  }
}

返回结果示例（符合期望的高亮效果）：

"hits": [
    {
        "_index": "idx_test2",
        "_type": "_doc",
        "_id": "GijDsYMBjgX3UBaguGxc",
        "_score": 0.2876821,
        "highlight": {
            "html.raw": [
                "Test <strong>More</strong> test"
            ]
        }
    }
]

方案2：调整映射，让原字段高亮使用去标签分析器 #

在映射中为html字段配置高亮使用parsed_analyzer，这样无需指定具体字段即可实现去标签后高亮：

修改后的映射片段：

"html": {
  "type": "text",
  "analyzer": "my_analyzer",
  "highlight": {
    "analyzer": "parsed_analyzer"
  },
  "fields": {
    "raw": {
      "type": "text",
      "fielddata": true,
      "analyzer": "parsed_analyzer"
    }
  }
}

使用原搜索查询即可得到去标签后的高亮结果。

内容的提问来源于stack exchange，提问作者user17047040