Automate xrefs to FAIRsharing

[cthoyt]

Closes #67

This PR automatically generates lexical matches to FAIRsharing and adds them in the xref field

SSSOM would be a much better way to store these mappings, so this PR also has code to generate SSSOM, but it was suggested by nico just to put them in the ad-hoc xrefs list.

FYI it took longer to figure out how to make the YAML diff small than to actually make the contribution :/ there were still a few cases where I couldn't get the YAML serializer to be standardized - this suggests that there should be a standard linter and CI test for correct formatting (in a follow-up PR)

[matentzn]

@cthoyt. This is great, thanks. Two things:

1. Can you share a suggestion of a good linter/formatter I can install in a github action to keep the YAML standardised?
2. I dont know yet if I need to, but would you mind if I had to change the curie string to a URI for the xrefs field? I need some clarity from the maintainers here, because I dont see an obvious place to document the prefixes.

[sierra-moxon]

Thanks for the edits!! awesome.

Note: there is a yamllint check in the GH actions, I'd rather not have it auto-correct and commit the fixed file as part of the GH action.

Instead of using xref, why don't we add "exact_mappings" or some other kind of mapping slot depending on the semantics intended, to the InformationResource class:

information-resource-registry/src/information_resource_registry/schema/information_resource_registry.yaml

Line 34 in ca03a8d

InformationResource:

. Generating SSSOM would be nice, if we did that based on a schema design pattern (e.g., "exact_mappings", I could emulate in other LinkML schemas, this could be a generic tool.

[cthoyt]

@matentzn

1. Can you share a suggestion of a good linter/formatter I can install in a github action to keep the YAML standardised?

the way every other project I've done is always to have a custom, explicit YAML/JSON formatter script that's part of the repo

2. I dont know yet if I need to, but would you mind if I had to change the curie string to a URI for the xrefs field? I need some clarity from the maintainers here, because I dont see an obvious place to document the prefixes.

I don't mind, this doesn't make a difference. Any reasonable data consumption workflow uses the bioregistry to standardize to CURIEs after the fact anyway ;)

@sierra-moxon does LinkML not have support for SSSOM? Something to consider. Otherwise, I think having an exact mapping slot would make a lot more sense (also using CURIEs)

[sierra-moxon]

there is a SSSOM schema element mapping generator, yes: https://github.com/linkml/linkml/blob/main/linkml/generators/sssomgen.py

I would consider these value mappings.