RWO URI dropped from subject when mapping from Cocina to MODS

[arcadiafalcone]

See https://purl.stanford.edu/bh354nc2071.mods

Desired output for MODS:

<subject>
  <name type="personal">
    <namePart>Kleinschmidt, Angeline</namePart>
    <nameIdentifier type="wikidata">https://www.wikidata.org/wiki/Q99810910</nameIdentifier>
  </name>
</subject>

[lwrubel]

@arcadiafalcone I understand that in this batch of records, the identifier.uri should be used for the nameIdentifier value and identifier.source.code becomes the type attribute. Is it required to have a identifier.source.code when a identifier.uri is used or is the type attribute optional? I'm looking back at #538 where it says "One of source.code, source.value, or source.uri should be required if there is identifier.value or identifier.code, but not required for identifier.uri."?

Are there other scenarios that we should transform here? I don't think we're expecting to see identifier.code for example; we're looking for identifier.value only at the moment.

[ndushay]

It might be good to do a comparison between the current name-identifier mapping code and the subject-name-identifier mapping code and see how they are the same and how they are different ... and decide what is desired.

For developers, the best way to express what is desired is in specs. Let us know how we can help.

[ndushay]

Given that so far this has only happened in one collection, and that collection has been addressed, I would say this is

• something that needs to be done
• NOT something that is uber-urgent ... it is more important to get it right than to get it done fast
• tricky, down-in-the-weed-details-of-metadata-mapping for the infra team developers -- it could be intricate enough vis-a-vis the mapping code to make it more tractable for JLitt or myself than for other team members
• possibly something that would benefit from Arcadia and Naomi pairing on determining what is happening now and specs for what we want (or maybe not -- maybe Arcadia can run a bunch of wrinkles through on qa or stage and figure out where we are now, and articulate in specs where we want to be).

whew!

[arcadiafalcone]

This project, and specifically the use of Wikidata URIs in this project, is the subject of a paper two SUL staff are writing. The paper doesn't have a hard deadline, but it would be good to get this done (at least to the point of the URI appearing in the MODS) sooner rather than later. I think the minimum logic to handle this use case is "if no identifier.value, use identifier.uri instead."

[ndushay]

@arcadiafalcone what we need to proceed here is some specs from you ... because the subject-name-identifier code is not the same as the contributor-name-identifier code, and it's unclear to us what exactly we need to change.

[arcadiafalcone]

There's an existing spec for it at https://github.com/sul-dlss/cocina-models/blob/06c90b69dc3183b643b236f948d658cb0966167e/spec/cocina/models/mapping/descriptive/mods/subject_name_spec.rb#L1370 which is not working in the Cocina>MODS direction. If there's another spec you need, let me know.

[mjgiarlo]

@arcadiafalcone @ndushay Hmmm, I may be misunderstanding. That spec looks like it's set up and not skipped, and when I run it, I see:

$ rspec spec/cocina/models/mapping/descriptive/mods/subject_name_spec.rb:1370
MODS subject name <--> cocina mappings
  Name subject with RWO identifier
    behaves like MODS cocina mapping
      when mapping from MODS (to cocina)
        mods snippet(s) produce valid MODS
        cocina hash produces valid cocina Description
        MODS maps to expected cocina
        notifier receives warning and/or error messages as specified
      when mapping to MODS (from cocina)
        cocina Description maps to expected MODS
        cocina Description maps to normalized MODS
      when mapping from roundtrip MODS (to cocina)
        roundtrip MODS maps to expected cocina
        roundtrip cocina maps to roundtrip MODS
        roundtrip cocina maps to normalized roundtrip MODS

Finished in 0.07003 seconds (files took 1.82 seconds to load)
9 examples, 0 failures

Is the data in the spec wrong or not rigged up correctly?

[mjgiarlo]

@arcadiafalcone I looked at this last month and had a question ☝🏻

[mjgiarlo]

I think we should close this until it's clear what needs fixing. Feel free to re-open then!