Closed cmungall closed 2 years ago
All of them? isnt it better to only load the "post-boomer" mappings? In this sense, ubergraph already loads them from Mondo for example.
I would suggest that if it really makes sense for Ubergraph to load mappings, it:
But before that happens it should be discussed whether these mappings are not better shipped by the ontology release file. What use case did you have in mind here? Having Ubergraph be a mapping provider as well as an ontology browser?
I did have in mind a single triplestore source of truth. It does make sense for each ontology to distribute the skos inside, but some ontologies will never do this. But maybe it's better to focus UG on being high quality manually vettted cocoon.
not urgent as there is not much there yet
the pipeline would be to iterate over sssom files using sssom-py to make ttl (maybe this will be done centrally at MC) then load