oai.yosemite validation results

[christinklez]

oai.yosemite issues

• [ ] coverage & spatial: currently mapping <dcterms:spatial> to both coverage and spatial. Please map to only spatial. (Note however that we have some messy spatial and coverage mapping logic in solr_updater.py : https://cdlib.slack.com/archives/CQ534NT47/p1699395989918789?thread_ts=1699393088.260719&cid=CQ534NT47)
• [ ] spatial: should be in an array (currently a string).

See examples: #26989 (2023-10-25); #27631 (2023-10-25)

---

likely addressed at upper levels

Verify whether these are addressed by the base oai mapper updates

• [ ] source: map source -- [note: this may get addressed in the base oai mapper]

[amywieliczka]

Made the following updates - not yet pushed up, but making a note here regarding some of the details and will push up momentarily.

• mapping: is_shown_by and is_shown_at both filtered identifiers for identifiers with "npsgallery.nps.gov" in the string (along with other unique filters for each), but this subdomain has changed to "nps.gov/npgallery", updated filter accordingly
• validating: updated is_shown_at, is_shown_by, and identifier validation logic to ignore this specific subdomain discrepancy
• validating: updated coverage and spatial validation logic to report out exclusively lossy discrepancies, instances where not all solr values exist in rikolti, but not additive discrepancies, instances where all solr values exist in rikolti as well as new values
• validating: records from collection 26989 all had the same value for is_shown_at: "https://npgallery.nps.gov/AssetDetail/3331008b-6cb0-4583-85f7-b3a93f7e074d", but in Rikolti, their is_shown_at value is unique to the record. the is_shown_at_improvements validator and the after_validation function check to make sure this new is_shown_at url is at the proper URL location and contains the proper object ID

Spatial is now a list, and source seems to be mapped - I didn't see any validation report errors for either of those. I did not remove the mapping to coverage due to the complexities of the solr_updater. Every collection has spatial and coverage values that are identical. We can revisit removing one or the other post MVP.

[amywieliczka]

Is closed by #696