pulibrary / pulfalight

This is an implementation of the Princeton University Library Finding Aids (PULFA) service using ArcLight
Other
7 stars 1 forks source link

dash vs dot in older figgy records not showing online content #954

Closed ccleeton closed 2 years ago

ccleeton commented 2 years ago

Many collections that previously had a . in the collection number and now have a - are not showing up as online in the finding aids. Example: https://findingaids.princeton.edu/catalog/AC107-06_c340 https://figgy.princeton.edu/catalog/055754ce-789d-4e71-aea5-e76dd4c042bb

Is there a way to fix this for the whole finding aids?

Resolution

Find all source metadata identifiers which have a period in them in Figgy, convert them to dashes.

phoebeno commented 2 years ago

This also in ACLU: https://findingaids.princeton.edu/catalog/MC001-03-05_c07334 https://figgy.princeton.edu/catalog/c494d2c4-b223-4400-b7b3-d2280157bf88

https://findingaids.princeton.edu/catalog/MC001-03-04_c1521 https://figgy.princeton.edu/catalog/8ea88671-9110-447e-b611-187d0a8282c7

hackartisan commented 2 years ago

Were all uses of . changed to -? Or is there a list? I'm not sure what the scale of this issue is.

@ccleeton I just changed the example you provided using the form in figgy, which causes the online content to come through. Could you take a look and see if it looks good still in figgy?

tpendragon commented 2 years ago

@ccleeton @phoebeno @faithc This got marked sudden priority. Presumably it's okay to push this to our next maintenance cycle - which would mean we'd look at bulk updating these the week of the 17th?

phoebeno commented 2 years ago

I think CC and FC are both out at the moment; the week of the 17th is not too far away for me. I would love to see the inheritance bug fixed first, as we have ACLU researchers coming in Jan. and the restrictions are not inheriting correctly to children (issue 924)

faithc commented 2 years ago

I'd defer to @ccleeton and @apferrar on this one, though addressing it the week of the 17th sounds doable to me.

@phoebeno - we're waiting for Regine to finish with the data cleanup before implementing this logic for accessrestrict notes since we identified data issues in MSS finding aids, correct? Unless it makes sense or is possible to apply it to repos aside from MSS or just ACLU for now?

tpendragon commented 2 years ago

@faithc The week of the 17th is a maintenance week, it's just easier to do these kind of between-cycle tickets during those times. The timing for the next work cycle for pulfalight hasn't changed.

faithc commented 2 years ago

Right, sorry. I confused myself with the lingo. Thanks for clarifying.

ccleeton commented 2 years ago

Sorry for the delay. Waiting for the maintenance week works for me.

tpendragon commented 2 years ago

All uses of . were changed to -

tpendragon commented 2 years ago

So the ID in Pulfalight was changed in the URL, but the actual call numbers still have periods in them.

For example: https://aspace.princeton.edu/staff/resources/2161#tree::archival_object_733628 The call number for this is AC107.07_c897. If you put that in Figgy it works fine.

In fact, if you use the "-" version (which is only in the pulfalight URL) in Figgy then I think DAO updating wouldn't work, since it'd try to find the archival object and not find it (there's no "AC107-07_c897").

So I think there's a few things to do here:

tpendragon commented 2 years ago

On the bright side I think there's only 15 resources in Figgy that have the "dash instead of period":

figgy_production=# select id, metadata->'source_metadata_identifier'->>0 FROM orm_resources WHERE metadata @> '{"source_metadata_identifier": []}' AND metadata->'source_metadata_identifier'->>0 LIKE '%-%\_%' AND metadata->'source_metadata_identifier'->>0 NOT LIKE '%.%';
                  id                  |       ?column?
--------------------------------------+-----------------------
 db638171-5856-451d-a323-cf2a9414217b | AC105-03_c3166
 8017bdb4-a9e7-4cce-9474-39c90e1fbf96 | AC105-04_c149
 055754ce-789d-4e71-aea5-e76dd4c042bb | AC107-06_c340
 1cbeb7b2-c379-4657-9102-df8ec6943bfb | RCPXG-5830371-2_c0528
 b2f92225-3835-4445-94f2-d244400daf1c | AC104-03_c2
 d40d3ede-7eef-405e-955f-8103b6eff6fc | AC198-03_c1536
 06b00c73-a596-4ecf-93cb-da052c8fa875 | AC198-04_c3932
 06746e41-aa46-41e2-8994-8109051a9af2 | AC104-03_c6158
 88a81495-718d-41e9-9534-d19c5ddc4e91 | AC107-14_c1223
 ad0046c2-f183-41ff-8277-2e668df453ad | AC198-02_c610
 150e0667-9a69-46b8-8aea-66d3c61aa707 | AC107-10_c3187
 ec23d20b-a52d-4687-9ab6-f3f42d6b0f97 | AC105-02_c147
 c667b46f-c069-4f58-959a-99a2a967855f | AC198-03_c698
 86f45dd2-1e75-4a0d-a381-4a411ac39641 | AC105-04_c4644
 2fb9377d-f556-4d94-a2a7-e4d490eac9e1 | AC104-03_c1338
(15 rows)
hackartisan commented 2 years ago

In fact, if you use the "-" version (which is only in the pulfalight URL) in Figgy then I think DAO updating wouldn't work, since it'd try to find the archival object and not find it (there's no "AC107-07_c897").

Figgy isn't searching by ID?

hackartisan commented 2 years ago

Many collections that previously had a . in the collection number and now have a -

So I read this ticket initially as saying that IDs had been changed in aspace. I think what @tpendragon is saying is that's not the case, that the only change was made in pulfalight for the purposes of having url-safe characters in the url. It sounds like we will keep the url that way but not use that form for anything else.

I think an ADR would be helpful here.

tpendragon commented 2 years ago

Records are fixed, ADR is in, and viewer queries appropriately.