Support custom grounding workflows - #438

Thanks! Once this makes it into a release I’ll update my post accordingly.

PyOBO's goal is to provide downstream user access to the most up-to-date versions of the ontologies more than support local/custom ontology usage

Fair enough. After all an ontology has (hopefully!) more users than editors, so it’s understandable to cater more for the needs of the formers than the laters. :)

@gouttegd v0.12.10 is live (https://github.com/biopragmatics/pyobo/releases/tag/v0.12.10) and the tutorial is live on RTD (https://pyobo.readthedocs.io/en/latest/ner.html)

all that being said, I think the reason for the slowdown is 1) bioregistry takes a long time to load these days, I would like to improve its caching and 2) identifier standardization happens in the background, which is beyond what the other OBO parsers do (and takes some extra time)

I think the reason for the slowdown is 1) bioregistry takes a long time to load these days, I would like to improve its caching and 2) identifier standardization happens in the background, which is beyond what the other OBO parsers do (and takes some extra time)

For what it’s worth, I observed that obonet.read_obo (used under the hood by pyobo.from_obo_path) can read FBbt in about 3 to 4 seconds if I recall correctly, so indeed it seems that most of the total time taken by PyOBO to “read“ FBbt (~45 seconds) is spent doing post-parsing, PyOBO-specific stuff.