Open mcourtot opened 3 years ago
It's best-practise to merge all imports when creating a release artifact (such as gecko.owl
), so that there are no owl:import
elements left, and the release artifact is self-contained. If you don't do that, you have two problems:
gecko.owl
they also need to connect to the Internet (or use a catalog file) to find the import files; network connections are never 100% reliableYou can see that all the BFO and IAO terms we need (and all the other imported terms) have been merged into gecko.owl
.
Imports are updated/refreshed at build-time according to recipes in the Makefile
, for example:
https://github.com/IHCC-cohorts/GECKO/blob/master/Makefile#L110
See also:
hm - when I look at GECKO released version I do not see the BFO top-level though.
If you want to see the top-level of BFO (entity, continuant, independent continuant, generically dependent continuant, occurrent, etc.), then we will put it in to GECKO. The current top-level of GECKO is based on the subset of BFO terms used in COB, which were selected to be intuitive to a broad audience with minimal jargon. We've used this same top-level structure since the first draft of GECKO. It will, of course, merge cleanly with any other ontology based on BFO.
Thanks for the extra explanation. I had never noticed the top-level being different - I got used to preferred roots in OLS :) Happy to keep COB if that's what you recommend.
I didn't see the BFO or IAO import at https://raw.githubusercontent.com/IHCC-cohorts/GECKO/master/gecko.owl