在elasticsearch-py中提供"script“参数

Question

我试图在curl中复制下面的elasticsearch-py突击队

curl -XGET "localhost:9200/my-exact-knn-index/_search?pretty" -H 'Content-Type: application/json' -d'
{
  "query": {
    "script_score": {
      "query" : {
        "bool" : {
          "filter" : {
            "range" : {
              "my-price" : {
                "gte": 0
              }
            }
          }
        }
      },
      "script": {
        "source": "\n          double value = dotProduct(params.query_vector, 'my-product-vector');\n          return sigmoid(1, Math.E, -value); \n        ",
        "params": {
          "query_vector": [-0.5, 90.0, -10, 14.8, -156.0]
        }
      }
    }
  }
}

但是我很难理解如何为index.search()调用提供“脚本”部分。我尝试在query参数和body参数(现在不推荐)中提供它，但是它给了我一个BadRequestError

query_body = {
    "script_score": {
      "script": {
        "source": "\n          double value = dotProduct(params.query_vector, '\''my-product-vector'\'');\n          return sigmoid(1, Math.E, -value); \n        ",
        "params": {
          "query_vector": [-0.5, 90.0, -10, 14.8, -156.0]
        }
      }
    }
  }


search_results = client.search(index="my-exact-knn-index", query=query_body)

>>> BadRequestError: BadRequestError(400, 'illegal_argument_exception', 'Required [query]')

知道该怎么做吗？

Answer 1

就快到了，您只需要一个query部分：

query_body = {
    "script_score": {
      "query": {                   <----- add this
          ....
      },
      "script": {
        "source": "\n          double value = dotProduct(params.query_vector, '\''my-product-vector'\'');\n          return sigmoid(1, Math.E, -value); \n        ",
        "params": {
          "query_vector": [-0.5, 90.0, -10, 14.8, -156.0]
        }
      }
    }
  }
}

Answer 2

我做了以下几件事，让它发挥作用：

query_body = {
    "script_score": {
      "query" : {
        "bool" : {
          "filter" : {
            "range" : {
              "my-price" : {
                "gte": 0
              }
            }
          }
        }
      },
      "script": {
        "source": "\n          double value = dotProduct(params.query_vector, 'my-product-vector');\n          return sigmoid(1, Math.E, -value); \n        ",
        "params": {
          "query_vector": [-0.5, 90.0, -10, 14.8, -156.0]
        }
      }
    }
}

search_results = client.search(index="my-exact-knn-index", query=query_body)

所以query不得不紧跟在script_score之后。