Closed kshefchek closed 5 years ago
@kshefchek I think I was being conservative when adding routes to avoid routes that never returned any associations (and I must have missed a few in the process).
But I saw @DoctorBud mention that certain pages fail to load because it expects routes to be symmetrical.
Do you recommend adding all possible enumeration of routes to the API and then have the API respond with an empty JSON if that route is not yet implemented?
Do you recommend adding all possible enumeration of routes to the API and then have the API respond with an empty JSON if that route is not yet implemented?
I don't think we should add routes if we know it will not return any data (in the current db). But I may be misunderstanding. I know there are some nuances with sequence features being treated as models.
Okay, just making sure. Couple of months back I had pruned many of the routes that never returned any associations.
After doing the comparison between routes.yaml
and facets for association_type
, it looks like the following routes are missing:
variant/<id>/models
model/<id>/cases
case/<id>/models
case/<id>/diseases
disease/<id>/cases
case/<id>/variants
variant/<id>/cases
case/<id>/genotypes
genotype/<id>/cases
case/<id>/genes
gene/<id>/cases
case/<id>/phenotypes
phenotype/<id>/cases
@kshefchek Do you want me to take a stab at adding these routes or have you started working on this already?
go for it!
With the association types and routes.yaml we should be able to determine what routes have yet to be added to biolink from the monarch solr:
https://solr.monarchinitiative.org/solr/golr/select/?facet.field=association_type&q=*%3A*&facet=true&facet.method=enum&facet.limit=3000&rows=0&wt=json&facet.mincount=1&indent=on&facet.sort=count
For exmaple: variant/:id/models, and all case endpoints that would be needed to create our UDP pages, eg: http://monarchinitiative.org/case/:MONARCH:c000017
Note that gene_temporal is bad naming and should be merged with gene interactions (colocalization data)