Is it possible to get ids of matching documents along with keywords in response from SKG

[Pratik--Patel]

Following is the sample response given in README page.

{ "data": [
  {
    "type": "jobtitle",
    "values": [
      {
        "id": "",
        "name": "Data Scientist",
        "relatedness": 0.989,
        "popularity": 86.0,
        "foreground_popularity": 86.0,
        "background_popularity": 142.0,
        "compare": [
          {
            "type": "skills.v3",
            "values": [
              {
                "id": "",
                "name": "Machine Learning",
                "relatedness": 0.97286,
                "popularity": 54.0,
                "foreground_popularity": 54.0,
                "background_popularity": 356.0
              },
              {
                "id": "",
                "name": "Predictive Modelling",
                "relatedness": 0.94565,
                "popularity": 27.0,
                "foreground_popularity": 27.0,
                "background_popularity": 384.0
              },
              {
                "id": "",
                "name": "Artificial Neural Networks",
                "relatedness": 0.94416,
                "popularity": 10.0,
                "foreground_popularity": 10.0,
                "background_popularity": 57.0
              },
              {
                "id": "",
                "name": "Apache Hadoop",
                "relatedness": 0.94274,
                "popularity": 50.0,
                "foreground_popularity": 50.0,
                "background_popularity": 1418.0
              },
              {
                "id": "",
                "name": "Java (Programming Language)",
                "relatedness": 0.76606,
                "popularity": 37.0,
                "foreground_popularity": 37.0,
                "background_popularity": 17442.0
              }
            ]
          }
        ]
      }
    ]
  }
 ]
}

For each word in the response, is it possible to get ids of documents having that word?

             {
                "document_ids": ["id1", "id2", "id3"...],
                "name": "Machine Learning",
                "relatedness": 0.97286,
                "popularity": 54.0,
                "foreground_popularity": 54.0,
                "background_popularity": 356.0
              }

If it is possible to do so then it would be a really powerful feature because then I can traverse from word_set1 to document_set1 and from document_set1 to word_set2 to document_set2 and so on AND keep track of the traversal (because it would be captured in the JSON response). I can then create a graph of words where link between the words represents set of documents and know which documents are in common.
Is it possible to do something like this currently?

Thanks for the great plugin BTW!

[SimmsJeason]

I try to get in the http://localhost:8983/solr/knowledge-graph/rel
but the error -->{"error":{"msg":"KnowledgeGraphHandler requires POST data","code":400}}

[scottbouma]

Are you just pasting that URL into a browser window? That is a "GET" call, rather than a POST. You might want to read up on REST API calls and how to use them. Maybe read this: https://mlsdev.com/blog/81-a-beginner-s-tutorial-for-understanding-restful-api