You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I am inclined to keep it that way - this is very useful, and breaking this to be an object instead like:
primary_knowledge_source: {id: infores:123, version: 2024-01-01} would break a lot of good work out there. But I am also not opposed to considering this - I am more interested that the KGX format does not break then the model.
And alternative is a second slot, like primary_knowledge_source_version which is conceptually not so nice as it is disconnected from primary_knowledge_source, but it would do the trick.
we don't have a standard in place to do this consistently, but we can add one.
check out biolink:DatasetVersion https://biolink.github.io/biolink-model/DatasetVersion/ as a starting point. Probably to do this "right" though, we'd need to link up *knoweldge_source with this object instead or in addition to denormalizing it on the resulting edge itself.
Or we need to disambiguate the use of primary_knowledge_source separate from other kinds of providence metadata (e.g. if we only every use primary_knowledge_source as a shorthand, we don't need to worry about grouping it in an object with other related metadata and instead just add something like, or DatasetVersion itself, in addition to primary_knowledge_source without a formal connection between the two).
The text was updated successfully, but these errors were encountered:
I would like to slowly promote adopting a strong provenance model for sources in biolink. This issue is the first in a sequence.
We currently have
https://biolink.github.io/biolink-model/primary_knowledge_source/
which is described using an infores identifier.
I am inclined to keep it that way - this is very useful, and breaking this to be an object instead like:
primary_knowledge_source: {id: infores:123, version: 2024-01-01} would break a lot of good work out there. But I am also not opposed to considering this - I am more interested that the KGX format does not break then the model.
And alternative is a second slot, like
primary_knowledge_source_version
which is conceptually not so nice as it is disconnected fromprimary_knowledge_source
, but it would do the trick.Any other ideas?
from @sierra-moxon:
The text was updated successfully, but these errors were encountered: