After split, Show pages' items list includes the filesets of child works

[ShanaLMoore]

Summary

Per Katharine,

I tested with my own 6 page PDF.

https://adl.s2.adventistdigitallibrary.org/concern/generic_works/538a6bcf-e2b7-429e-b15a-f19530a9a7d9?locale=en

BUGS:
The splitting worked, but we have double the number of child works showing in the items list. On Production we usually only see the child works with page numbers attached. However, on staging, there are a string of JPEGs in the item list.

All child works are showing in the UV, so the work is duplicated in the UV.

Acceptance Criteria

• Feature parity with production
• After split, the items section should only show the child works with page number attached.
• After split, the parent work should not also show the child work's filesets in the items section.
• No duplication in the UV

Screenshots or Video

STAGING

PRODUCTION

Testing Instructions

Ensure that feature flipper is set to universal viewer. Find a work with child works.

• Universal viewer shows each page only once.
• Universal viewer search works from either parent or child works.
• Both child works and child work filesets will show in items list.

Select edit and save to reindex the parent work.

• Only child works show in items list... child work's filesets no longer appear.

Create a new work with a PDF to split.

• Universal viewer shows each page only once.
• Universal viewer search works from either parent or child works.
• Both child works and child work filesets show in items list.

Notes

related issue: samvera/hyrax#6767

[ShanaLMoore]

Thoughts on an override to the indexers? Will it have negative consequences?

# hyku_indexing.rb and maybe pcdm indexer?
        solr_doc['member_ids_ssim'] = filter_member_ids(object.member_ids.map(&:to_s)) 
        solr_doc['file_set_ids_ssim'] = filter_fileset_ids(solr_doc['member_ids_ssim'])

private

  def filter_member_ids(member_ids)
    member_ids.reject do |id|
      member = Hyrax.query_service.find_by(id: Valkyrie::ID.new(id))
      member.is_a?(Hyrax::FileSet)
    end
  end

def filter_fileset_ids(member_ids)
  member_ids.select do |id|
    member = Hyrax.query_service.find_by(id: Valkyrie::ID.new(id))
    member.is_a?(Hyrax::FileSet)
  end
end

update 1: this didn't fully solve the problem. I still see the child filesets in the items section but I don't see them in the UV. I think it's because #list_of_item_ids_to_display also needs to filter them out for the view

update 2:

# hyrax work_show_presenter.rb, ugly... but it works? 

# rejects Hyrax::FileSets unless it's a pdf. 

    def authorized_item_ids(filter_unreadable: Flipflop.hide_private_items?)
      @member_item_list_ids ||= begin
        ids = filter_unreadable ? ordered_ids.reject { |id| !current_ability.can?(:read, id) } : ordered_ids
        ids.reject do |id|
          member = Hyrax.query_service.find_by(id: id)
          if member.is_a?(Hyrax::FileSet)
            original_file = member.original_file
            original_file&.pdf? ? false : true
          else
            false
          end
        end
      end
    end

=> 🎉

[ShanaLMoore]

Per LaRita, koppie and dassie work as expected.

[laritakr]

Appears to be due to IiifPrint's indexing that includes descendents for OCR purposes.

[laritakr]

QA Results: ✅ Passed

Tested with work https://adl.s2.adventistdigitallibrary.org/concern/generic_works/9854b255-38ee-4cdd-95cf-004b9f8d324a?locale=en

Screenshots

UV does not duplicate pages

Before a reindex, child work filesets are included in items

After edit/save (to reindex) child work filesets are not included in items

[KatharineV]

Working as expected when tested in ADL staging. Sample work: https://adl.s2.adventistdigitallibrary.org/concern/journal_articles/lake_union_herald_2023_05_18_19_how_shall_we_go?locale=en

Viewer shows the PDF once, and the items list has only page number child works.