Prefix for mesh not conform to OBO context

[anitacaron]

The prefix should be mesh instead of MESH for mondo.sssom.tsv and mesh_chebi_biomappings.sssom.tsv. For reference, OBO Context at Bioregistry.

SSSOM-API can't expand or compress properly when it retrieves the' mesh' mappings, resulting in None.

Related to #29

[matentzn]

The prefix does not have to conform to obo context if it is included in the mappings. No curie_map in there?

[hrshdhgd]

If this is a result of my latest PR:

In umls-ingest: The priority sequence for the curies=> Converter for now is:

1. OBO converter
2. Bioregistry Converter
3. SSSOM default converter.

So mesh is the prefix in the curie_map: See here

[anitacaron]

SSSOM-API only depends on the OBO Context.

[hrshdhgd]

I'm a little confused. The prefix mesh is as expected (as per your first statement). @matentzn why am I tagged here?

[anitacaron]

Sorry, I'm referring to the previous mappings available here mondo.sssom.tsv, mesh_chebi_biomappings.sssom.tsv and gene_mappings.sssom.tsv. Now I checked the registry, and there are two others added yesterday.

[hrshdhgd]

Yes, those were added by me via #44

[anitacaron]

Yes, these new mappings are good. One more reason to update the others. I updated the issue to be more specific.

[matentzn]

I moved this discussion here for now: mapping-commons/sssom-api#85

@anitacaron we can attack this from two sides, but it must be from both:

1. Weaken the assumption of obo context (issue above)
2. Standardising as much as possible towards it.

@glass-ships can you make sure that all curies you output in gene mappings and chebi mesh are standardised uses curies.Converter.standardize() using the obo_converter()?

[glass-ships]

sure, I can try.