Add more example local unique identifiers

[cthoyt]

Motivated by biopragmatics/bioregistry#1821

There are unfortunately only a handful (<20) of BARTOC records that annotate example local unique identifiers using the EXAMPLES field.

It would be great if more records had this information, which is an important signal that a resource is "real" and not just a stub. Additionally, having an explicit example is a requirement for annotating a resource in the Bioregistry/Semantic Farm, so many BARTOC records can't be imported without additional manual curation of this field.

I have considered using BARTOC's API annotations for many resources to try and automatically obtain an example local unique identifier, but this is going to take some time as I will have to understand each API schema first. BARTOC already implements doing this, so it would make a lot of sense to handle these queries in BARTOC itself, and add an explicit example when possible.

[nichtich]

Additional notes:

• We also need to decide on whether and where to put this field in strict JSKOS. There is item field example but this is for textual examples. Maybe keep EXAMPLES as extension specific to BARTOC
• Examples should also be used for automated testing whether an API is reachable
• There are encoding issues when an example identifier is a notation vs a full IRI.
• Wikidata identifier properties also contain examples, maybe these can be harvested as well

[cthoyt]

We also need to decide on whether and where to put this field in strict JSKOS. There is item field example but this is for textual examples. Maybe keep EXAMPLES as extension specific to BARTOC

An example IRI also works if there is a namespace defined, since it can just be stripped off the beginning

Examples should also be used for automated testing whether an API is reachable

Bioregistry/Semantic Farm implements something like this in https://github.com/biopragmatics/bioregistry/blob/main/src/bioregistry/health/check_providers.py with a cron job configured in https://github.com/biopragmatics/bioregistry/blob/main/.github/workflows/health.yml, results displayed at https://biopragmatics.github.io/bioregistry/health/.

Wikidata identifier properties also contain examples, maybe these can be harvested as well

Yes! We do this in Bioregistry/Semantic Farm with this code: https://github.com/biopragmatics/bioregistry/blob/main/src/bioregistry/external/wikidata/__init__.py. Please take a look there for inspiration

[cthoyt]

I wrote a script that produces BARTOC -> example LUID lists based on bioregistry->BARTOC mappings. I attached the script and the results (in JSON) in

https://gist.github.com/cthoyt/a11b51a8a186b1e3d6e47918f2a82165