scientist-softserv / adventist_knapsack

Apache License 2.0
2 stars 0 forks source link

Mappings - Collection: #573

Closed jeremyf closed 1 year ago

jeremyf commented 1 year ago

Summary

This issue refers to mapping spreadsheet line 24, Collection information. As of 2022-12-19, transfer from OAI_ADL to Hyku is changing the value of metadata across the fields in that line of the mapping spreadsheet. MODS and OAI_ADL have the correct collection, but somehow Hyku shows incorrect data. See this example: MODS, OAI_ADL, and Hyku, and MODS, OAI_ADL, Hyku. (The behavior on the ADL/Ebstore end was fixed by Eric Koester.) The examples linked here were reimported to Hyku and the metadata refreshed, but the incorrect values remained. Is there Hyku behavior at play here?

From https://docs.google.com/document/d/1mIOT23UAilSO77pAlXYSWJEHw3YK3BNNQVTKNzd41ao/edit#

Acceptance Criteria

Screenshots or Video

Testing Instructions

This one could be challenging to test because there are two problems: 1) Collections being associated with the wrong parent collection and 2) Collections not being associated with the right parent collection.

We have https://adl.s2.adventistdigitallibrary.org/importers/32?locale=en#collection-entries as an example importer:

As structured in Bulkrax, we don't appear to break existing relationships during import. This means that any incorrect associations will remain after we run the importer. For this and only this test, I have manually removed the bad relationships (e.g. part of the "genre:periodical" and "The Southern Watchmen" parent collection.) In part because it's unclear how they got there in the first place, and I want to see if they "re-appear" (or if some other things reappear during the test).

I ran the following Ruby code in the console to tidy up the collections:

switch!(Account.first) # This is the adl.s2 tenant
collection = Collection.find("6b27e5d2-28e5-49b0-aed6-305b0af3c795")
collection.member_of_collections = []
collection.save

From the above information we'll want to seek guidance from Katharine regarding the state of relationships on the staging server, and what we'll want to do for a clean run.

Do we eradicate and start some ingests over? Or do we run a script to break relationships for objects we ingest/reingest?

Notes

KatharineV commented 1 year ago

See also this recent import to staging: https://adl.s2.adventistdigitallibrary.org/collections/6b27e5d2-28e5-49b0-aed6-305b0af3c795?locale=en

In this example, the metadata field reads the correct "part of" value, but the parent collection remains incorrect.

jeremyf commented 1 year ago

@KatharineV could you point me to the importer that generated this entry?

I need to sleuth the date that this run, the application state, and also understand the raw metadata of this record.

KatharineV commented 1 year ago

@jeremyf I believe this is the importer: https://adl.s2.adventistdigitallibrary.org/importers/32?locale=en#collection-entries

jeremyf commented 1 year ago

Given that we're skipping auto-collection creation, I'm going to close this ticket. I think it's stale and there are more representative tickets in the queue.