Elasticsearch: Sort the Documents on the index value of the search string in a text field

Question

I have Elasticsearch data like this-

PUT /text/_doc/1
{
  "name": "pdf1",
  "text":"For the past six weeks. The unemployment crisis has unfolded so suddenly and rapidly."
}
PUT /text/_doc/2
{
  "name": "pdf2",
  "text":"The unemployment crisis has unfolded so suddenly and rapidly."
}

In this example I am making a full text search, I am searching for all the documents that have "unemployment" sub-string in the "text" field. And in the end i want all the documents sorted in the ascending order of the index value of "unemployment" string in the "text" field. For eg - the sub-string "unemployment" comes first in the doc2 at index "4" so i want this document to be returned first in the results.

GET /text/_search?pretty
{
  "query": {
    "match": {
      "text": "unemployment"
    }
  }
}

I have tried few things like term_vector, here is the mapping that i used but it didn't help.

PUT text/_mapping
{
    "properties": {
        "name" : {
          "type" : "text",
          "fields" : {
            "keyword" : {
              "type" : "keyword"
            }
          }
        },
        "text" : {
          "type" : "text",
          "term_vector": "with_positions_offsets"
        }
      }
}

Can anyone please help me in making the right mapping and search Query?

Thanks in Advance!

Answer 1

Try this query

GET text/_search
{
  "query": {
    "function_score": {
      "query": {
        "match": {
          "text": "unemployment"
        }
      },
      "functions": [
        {
          "script_score": {
            "script": {
              "source": """
                def docval = doc['text.keyword'].value;
                def length = docval.length();
                def index = (float) docval.indexOf('unemployment');

                // the sooner the word appears the better so 'invert' the 'index'
                return index > -1 ? (1 / index) : 0;
              """
            }
          }
        }
      ],
      "boost_mode": "sum"
    }
  }
}

using the auto-generated mapping

{
  "text" : {
    "mappings" : {
      "properties" : {
        "name" : {
          "type" : "text",
          "fields" : {
            "keyword" : {
              "type" : "keyword",
              "ignore_above" : 256
            }
          }
        },
        "text" : {
          "type" : "text",
          "fields" : {
            "keyword" : {
              "type" : "keyword",
              "ignore_above" : 256
            }
          }
        }
      }
    }
  }
}

Note that this is case-sensitive so it'd be reasonable to have a lowercase-normalized keyword field too and then access it in the script score script. This might get you on the right path.