nodes and query-paths endpoints return node data in different formats

[dhimmel]

@vincerubinetti mentioned that this is an annoyance for the frontend.

From https://search-api.het.io/v1/nodes/

        {
            "id": 0,
            "identifier": "128239",
            "identifier_type": "int",
            "name": "IQGAP3",
            "data": {
                "url": "http://identifiers.org/ncbigene/128239",
                "source": "Entrez Gene",
                "license": "CC0 1.0",
                "chromosome": "1",
                "description": "IQ motif containing GTPase activating protein 3"
            },
            "metanode": "Gene"
        },

From https://search-api.het.io/v1/query-paths/?source=11545&target=33324&metapath=SEcCrCrC:

    "nodes": {
        "11545": {
            "neo4j_id": 11545,
            "node_label": "SideEffect",
            "data": {
                "name": "Nasal itching",
                "source": "UMLS via SIDER 4.1",
                "identifier": "C0850060",
                "url": "http://identifiers.org/umls/C0850060",
                "license": "CC BY-NC-SA 4.0"
            },
            "metanode": "Side Effect"
        },

[vincerubinetti]

As we discussed in person, the best idea might be to put every property that will always be there in the top level, and any property that is only sometimes there under data. This would probably be the most clear and least arbitrary split. The other option is to put the more "significant" (subjectively) properties -- like perhaps name, metanode, and source -- in the top level, and the less important ones under data.

Also, I recommend changing the name of data since it's pretty ambiguous and nondescript, and it also conflicts with some of the nomenclature in D3 and elsewhere in the code. If we do the always/sometimes split mentioned above, I recommend renaming data to optional_fields or extra_fields or something like that.

We can get to this at some other point in time since it's a relatively minor issue.

[vincerubinetti]

@dhimmel seems like this can be closed?

[dhimmel]

Well we addressed the name of data, but we haven't unified the format of the nodes results to be the same from the /nodes and /query-paths endpoints.