problems with query processing

[balhoff]

This query is being test in the Translator standup:

{
  "message": {
    "query_graph": {
      "nodes": {
        "n0": {
          "categories": [
            "biolink:ChemicalSubstance"
          ],
          "name": "Chemical Substance"
        },
        "n1": {
          "name": "EGFR",
          "ids": [
            "NCBIGene:1956"
          ],
          "categories": [
            "biolink:Gene"
          ]
        }
      },
      "edges": {
        "e0": {
          "subject": "n0",
          "object": "n1",
          "predicates": [
            "biolink:decreases_abundance_of",
            "biolink:decreases_activity_of",
            "biolink:decreases_expression_of",
            "biolink:decreases_synthesis_of",
            "biolink:increases_degradation_of",
            "biolink:disrupts",
            "biolink:negatively_regulates_entity_to_entity"
          ]
        }
      }
    }
  }
}

I feel like something has gone very wrong, because the results of submitting this query look odd, both the result bindings and the returned query graph, which looks like this:

"query_graph": {
      "nodes": {
        "n0": {},
        "n1": {}
      },
      "edges": {
        "e0": {
          "predicate": [
            "biolink:related_to"
          ],
          "subject": "n0",
          "object": "n1"
        }
      }
    }

[balhoff]

We realized the above result comes from submitting that TRAPI 1.1 query to the 1.0 endpoint. If it is submitted to the 1.1 endpoint, it returns no results, probably because of two problems:

• Use of NCBIGene prefix instead of the NCBIGENE prefix specified in Biolink
• Use of non-existent biolink:negatively_regulates_entity_to_entity predicate

[balhoff]

I removed the bug label for now. I think we should leave open and make sure we follow up on the biolink issues.