scientist-softserv / adventist_knapsack

Apache License 2.0
1 stars 0 forks source link

UV not rendering after split #178

Closed KatharineV closed 1 month ago

KatharineV commented 2 months ago

Works uploaded after Knapsack are not rendering in the UV after the pages are split. We expect all split documents to render in the UV.

In this screenshot from SDAPI prod, you see the first split page shows in the Items list, but the document doesn't render in PDF.js (see scientist-softserv/adventist_knapsack#180) or the UV.

https://sdapi.b2.adventistdigitallibrary.org/concern/journal_articles/dialogue_2023_2_29_31_a_pioneer_of_adventist_church_school_curriculum_alma Image

ShanaLMoore commented 2 months ago

Local: This may be the issue - I created a Generic Work and attached a multi page pdf. The pdfs are attached but the UV displays an error. Yarn did not fix it. I think Hyku/adventist_dev is missing the pdfjs directory in its public folder.

Image

ShanaLMoore commented 2 months ago

PR started:

https://github.com/samvera/hyku/pull/2185

TODO:

Update Adventist Knapsack submodule once it gets merged into HYKU/adventist_dev

ShanaLMoore commented 2 months ago

You can see the UV trying to load after the changes in the above PR. I will turn PDF.js off, reprocess the works and report back if they render in the UV.

Image

UPDATE:

The UV still won't render due to errors in the manifest.

Image

ShanaLMoore commented 2 months ago

QA: Pass ✅

Image

KatharineV commented 2 months ago

Team, I'm having a heck of a time trying to test this ticket. I can't get a new work into either ADL or SDAPI production tenants, because of bugs that are blocking single upload and bulkrax imports. So I can't see if a new work will split and render in the UV as expected. I can confirm that uploads from about a week ago or earlier this week are either a) not showing in the UV or _any _viewer,__ despite splitting, or b) not splitting at all and just stopped at the PDF.js stage.

On a staging site, I managed to use Bulkrax to bring in a work via OAI, and it has not split yet (see link below). It's been a couple hours since import. No splitting, no UV...so I still can't test. I feel pretty stuck. It seems like this ticket isn't fixed, based on older works, but I can't confirm with a new upload until other blockers are cleared...unless you have a suggestion for me.

https://testing.s2.adventistdigitallibrary.org/concern/published_works/20121633_to_brother_j_n_andrews_and_sister_h_n_smith?locale=en

I imported the above work using this importer, which brought in a second work which HAS split, but incorrectly. There are issues with the child works and how they render in the viewer. The UV is rendering, but: Child works in the items list appear incomplete--thumbnails and metadata are weird. In the viewer, the page numbers are appearing out of order and don't follow the page order of the original PDF (see printed page numbers and compare viewer to download of original PDF). The last page of the PDF is entirely missing. That's printed page 32, and it is a single page, not a double page spread. Double page spread of printed pages 28 & 29 appear to be missing too.

Link: https://testing.s2.adventistdigitallibrary.org/concern/published_works/20121629_the_first_day_of_the_week_not_the_sabbath_of_the_lord

Child works look weird in Items list: Image

KatharineV commented 2 months ago

Related to this ticket, I found a work that has child works visible via the Dashboard, but they don't show on the parent record in the items list. The work is not loading in the UV, the child works seem lost, and even PDF.js isn't rendering right now (it's been spinning for a few minutes).

https://adl.b2.adventistdigitallibrary.org/concern/published_works/22266530_cardinal_2017

No child works on the work page:

Image

Child works do show on the Dashboard:

Image

Here's one of them: https://adl.b2.adventistdigitallibrary.org/concern/published_works/6c63102e-2e93-40bc-84df-e7fcd46ee127?locale=en

ShanaLMoore commented 2 months ago

Hi @KatharineV Is this still an issue? ref: https://adl.b2.adventistdigitallibrary.org/concern/published_works/22266530_cardinal_2017

This morning I've been testing with https://april-test.b2.adventistdigitallibrary.org/concern/images/b1e2291b-edfa-421e-87ff-7d49d28ec3f5?locale=en and here are my results. (this is a manually created work with files attached. UV feature turned on. jobs complete)

Image

KatharineV commented 2 months ago

Confirmed that the work that was stuck before is rendering in the UV now. https://adl.b2.adventistdigitallibrary.org/concern/published_works/22266530_cardinal_2017

Hooray!

Image

I also just uploaded a set of works to SDAPI prod using a CSV/files importer and the PDFs split and are displaying as expected in the UV:

https://sdapi.b2.adventistdigitallibrary.org/concern/journal_articles/dynamic_steward_2023_04_06_18_19_proposal_of_partnership_between_treasury_a

KatharineV commented 2 months ago

Team, I found a work today where the UV isn't rendering anything on ADL prod: https://adl.b2.adventistdigitallibrary.org/concern/images/20000208_illustration_of_harmon_hall. The attached files are images, so I expect to see the UV, not PDF.js. I see nothing. Is this a one-off issue or possibly related to the UV bug, post-Knapsack? Thanks for checking! @ShanaLMoore I hope you don't mind me tagging you. This ticket moved to the Done category, so I want to make sure this note is seen.

Here's how the work is displaying for me right now:

Image

ShanaLMoore commented 2 months ago

Got it. We'll take a look.

ShanaLMoore commented 2 months ago

@KatharineV Do you know how this record was created? The first tiff file isn't actually present or downloadable. Can we delete that file and/or try recreating this record?

KatharineV commented 2 months ago

Yes, this is a work that my predecessor uploaded using Bulkrax and the OAI feed (adl:image set, I believe) back in December 2021. It would be part of the earliest batch uploads. Other works from those loads do show in the UV, which is why I'm suspicious this might be an issue with the work, not Knapsack.

I see your edits now. Let me try updating that single record and TIFF!