Tier 1 returns 0 results when Tier 0 returns 2547 results

[maximusunc]

Given the following query, Tier 1 returns 0 results fairly quickly. Maybe an overzealous branch pruning?

curl -X POST https://dev.retriever.biothings.io/query \
  -H 'accept: application/json' \
  -H 'Content-Type: application/json' \
  -d '{
    "message": {
      "query_graph": {
        "nodes": {
          "SN": { "categories": ["biolink:ChemicalEntity"] },
          "ON": {
            "ids": ["MONDO:0005148"],
            "categories": ["biolink:DiseaseOrPhenotypicFeature"]
          },
          "f": { "categories": ["biolink:Disease"] }
        },
        "edges": {
          "edge_0": {
            "subject": "SN",
            "object": "f",
            "predicates": ["biolink:treats_or_applied_or_studied_to_treat"]
          },
          "edge_1": {
            "subject": "f",
            "object": "ON",
            "predicates": ["biolink:has_phenotype"]
          },
          "edge_2": {
            "subject": "ON",
            "object": "f",
            "predicates": ["biolink:has_phenotype"]
          }
        }
      }
    },
    "parameters": {
      "tiers": [1]
    }
  }' > retriever_response.json

[tokebe]

Hmm, doesn't look like a pruning issue...looks like either Tier 1 isn't returning everything Tier 0 is, or the QGX is failing to reconcile partials correctly.

[tokebe]

Ah, turns out dgraph shouldn't be returning results. The results it returns are incorrectly bound. @everaldorodrigo looks like a cascaded_or issue or something?

[maximusunc]

How are things incorrectly bound? I'm seeing lots of self-edges, type-2 diabetes mellitus - has_phenotype -> type-2 diabetes mellitus, but that's still a valid edge. Does Tier 1 explicitly disallow that?