Need to change google distance edge

[edeutsch]

The semantic validator that Richard is putting together is almost ready to deploy. He and I have been doing some testing and there are still a few things for him to fix, but I've successfully run some ARAX Responses through the validator and for the most part is works and our output is good. A few of the flagged problems are things that he needs to fix. But one notable error is:

    {
      "context": "Query Graph n00--['biolink:has_normalized_google_distance_with']->n01 Predicate",
      "name": "biolink:has_normalized_google_distance_with",
      "code": "error.unknown"
    }

flagging the fact that we just made up biolink:has_normalized_google_distance_with and it doesn't exist. But apparently there is now a predicate biolink:occurs_together_in_literature_with
https://biolink.github.io/biolink-model/docs/occurs_together_in_literature_with.html#relation-occurs_together_in_literature_with

So I'm thinking we should switch all occurrences in our code to that. Agreed?

But the we also need a proper way to encode the metric. We currently do:

  "message": {
    "knowledge_graph": {
      "edges": {
        "N1_0": {
          "attributes": [
            {
              "attribute_source": null,
              "attribute_type_id": "EDAM:data_2526",
              "attributes": null,
              "description": "\n        Normalized google distance is a metric based on edge subject/object node co-occurrence in abstracts of all [PubMed](https://pubmed.ncbi.nlm.nih.gov/) articles. \n        The formula can be found here on [wikipedia.](https://en.wikipedia.org/wiki/Normalized_Google_distance) \n        Where in this case f(x,y) is the number of PubMed abstracts both concepts apear in, f(x)/f(y) are the number of abstracts individual concepts apear in, and N is the number of pubmed articles times the average numbver of search terms per article (27 million * 20).\n        ",
              "original_attribute_name": "normalized_google_distance",
              "value": "0.775434002709891",
              "value_type_id": null,
              "value_url": "https://arax.ncats.io/api/rtx/v1/ui/#/PubmedMeshNgd"
            }
          ],
          "object": "CHEMBL.COMPOUND:CHEMBL2109588",
          "predicate": "biolink:has_normalized_google_distance_with",

          "subject": "CHEMBL.COMPOUND:CHEMBL112"
        },

I'm thinking that we need to change:

1. The predicate should change as described above
2. The attribute_type_id is EDAM:data_2526 which is just "text data": https://www.ebi.ac.uk/ols/ontologies/edam/terms?iri=http%3A%2F%2Fedamontology.org%2Fdata_2526&lang=en&viewMode=All&siblings=false . I'm thinking we can do better than that.
3. is the description accurate for what we're actually doing?

[chunyuma]

The predicate name has been changed from biolink:has_normalized_google_distance_with to biolink:occurs_together_in_literature_with used in all associated scripts in a new branch issue1925. All related tests have been passed. Now we need to wait the feedback for getting a better EDAM code to replace the original "EDAM:data_2526".

[edeutsch]

Great, thanks, @chunyuma !
If a replacement for EDAM:data_2526 seems like a long way off based on initial contact with Matt, Sierra, I would be fine to move ahead with just the predicate change. That will quiet an egregious error and only leave a minor warning that might take a time to resolve. @saramsey let us know whether you think Matt or Sierra can suggest a quick remedy or not.

[saramsey]

I've posted this question on the EPC GitHub project area:
NCATSTranslator/Evidence-Provenance-Confidence-Working-Group#8

and DM'd Matt and Sierra about it.

[saramsey]

Matt and Sierra have responded. Here is what Matt wrote:

Here was the hyperlink that Matt provided: here

Does this help clarify our path forward?

[saramsey]

Sierra asked me to explain how ARAX computes NGD. Here is what I wrote (hope I got it right?)

[dkoslicki]

@saramsey Can we have the link to the "s/s" google doc mentioned by Matt?

[edeutsch]

So what's the state of sunsetting has_normalized_google_distance_with? Tyler is turning up the heat on non-compliant TRAPI. This still appears to be a problem. Perusing the issue here suggests that maybe the fix has been sitting stale in a branch for months? Are we waiting for something to happen first before merging this branch? Is the ranker prepared to handle this change?

@chunyuma or anyone else: what stands in the way of merging this into master?

[dkoslicki]

@chunyuma , this is a simple name change, right? Or is there more to it than that? I don't see any reason why not to make it biolink compliant

[chunyuma]

Sorry @edeutsch @dkoslicki, I previously thought that we might need to wait for the final decision about whether we need to change the EDAM code. It seems like our final decision is to keep it as is, right?

I have just merged the branch issue1925 to the master branch. I'm so sorry for the delay.

[dkoslicki]

No worries at all @chunyuma ! I think this just fell through the cracks: nothing to be sorry about!

[edeutsch]

yes, no problem, no one has really been checking, but now they're going to start checking so it is good that we roll this out now. thank you!

I wonder if there are other branches that are similarly sitting stale that could be merged? Maybe we should do a branch and issue cleaning party at a mini-hackathon soon?

[edeutsch]

This is now rolled out in all the usual places where master is rolled out and seems to work. closing, thanks.