There seems to be a validation issue in the latest version of ARAX in master:
https://arax.ncats.io/?r=468575
See result 100:
This appears to be an edge inference event. the support graph of this edge suggests that ARAX is inferring a "treats" edge based on the "applied_to_treat" and the "in_clinical_trials_for" edges from Retriever I think.
But there's only one edge source, which is an aggregator_knowledge_source of arax. But it has an upstream_source_ids of infores:retriever. but there is no infores:retriever source record.
It may be that the only fix needed is that this sort of edge needs to have one more edge with infores:retriever as the primary? hmm, no I don't think that's right.
Maybe what we need is this:
{
"resource_id": "infores:obie",
"resource_role": "primary_knowledge_source",
"upstream_resource_ids": [
"infores:retriever"
]
},
{
"resource_id": "infores:retriever",
"resource_role": "supporting_data_source",
}
But I'm quite uncertain about this.
Maybe the thing to do is to corner Matt Brush next week and see if we can understand what we should put here.
Hoping for help from @saramsey
This grew out of the incorrect bug report #2802
There seems to be a validation issue in the latest version of ARAX in
master:https://arax.ncats.io/?r=468575
See result 100:
This appears to be an edge inference event. the support graph of this edge suggests that ARAX is inferring a "treats" edge based on the "applied_to_treat" and the "in_clinical_trials_for" edges from Retriever I think.
But there's only one edge source, which is an aggregator_knowledge_source of arax. But it has an upstream_source_ids of infores:retriever. but there is no infores:retriever source record.
It may be that the only fix needed is that this sort of edge needs to have one more edge with infores:retriever as the primary? hmm, no I don't think that's right.
Maybe what we need is this:
But I'm quite uncertain about this.
Maybe the thing to do is to corner Matt Brush next week and see if we can understand what we should put here.
Hoping for help from @saramsey
This grew out of the incorrect bug report #2802