Lucene CustomScoreQuery not enumerating long primitives

[ghost]

Elasticsearch version (bin/elasticsearch --version): Version: 5.6.1, Build: 667b497/2017-09-14T19:22:05.189Z, JVM: 1.8.0_144

Plugins installed: [custom plugin]

JVM version (java -version): 1.8

OS version (uname -a if on a Unix-like system): #39~16.04.1-Ubuntu SMP Wed Sep 13 09:02:42 UTC 2017 x86_64 x86_64 x86_64 GNU/Linux

Description of the problem including expected versus actual behavior:
I've written a custom Lucene CustomScoreQuery subclass that is expected to: (1) classify documents using a long field type as matched or not, (2) score the hit documents using another long field type. The problem I am facing is that the CustomScoreQuery's subquery is not iterating through the terms of the given field. If I change the field type from long to keyword for example, the getTermsEnum() function of the subquery is executed and a FilteredTermsEnum instance is returned, which is the expected behaviour. In addition, I've used the plugin successfully in Apache Solr, however the same does not imply to Elastic. The most important implementation details may be found here.

The index mapping I've defined is as follows:

{
  "images" : {
    "mappings" : {
      "descriptors" : {
        "properties" : {
          "image_descriptor_long" : {
            "type" : "long"
          },
          "image_descriptor_short" : {
            "type" : "long"
          }
        }
      }
    }
  }
}

[jimczi]

@dsafaric I answered on the forum directly, thanks for opening the discussion. I am going to close this issue because we reserve github for confirmed bugs and feature requests, we can continue the discussion on the forum.