🎁 Add conditional PDF split or PDF.js

This is a preliminary commit to add conditional logic for either
rendering PDF.js or IIIF viewer for split pages.

There is a particular case where QA might appear to fail:

1. Set the Tenant to use UV
2. Create a work with a PDF (this will split the page)
3. Go to the work's page, you'll see the UV
4. Set the Tenant to use PDF.js
5. Refresh the work page, and you should see PDF.js but instead will see
   the UV.

To work around this (and realistically how most people will experience this):

1. Set the Tenant to use UV
2. Create a work with a PDF
3. Go to the Dashboard
4. Click the work's URL
5. You should see a UV for the PDF
6. Close the work's page
7. Change the tenant to use PDF.js
8. Go to the Dashboard
9. Click the work's URL

The conjecture is that there's some iframe and/or turbolinks caching.
Given that this is an edge case regarding toggling on and off a tenant
feature, we think the work around is adequate.

Related to:

- scientist-softserv/palni-palci#666
- scientist-softserv/palni-palci#675
- scientist-softserv/palni-palci#704
- scientist-softserv/palni-palci#705
- #162

Co-authored-by: Shana Moore <[email protected]>
Co-authored-by: Kirk Wang <[email protected]>

Gemfile.lock

@@ -43,17 +43,15 @@ GIT
 
 GIT
  remote: https://github.com/scientist-softserv/iiif_print.git
- revision: 8fdf56e151b5adb7e6aab1cd29d39b74554ed48a
+ revision: 50b2659f4584ac85f468c7fcec56398f6221a412
  branch: main
  specs:
  iiif_print (1.0.0)
- blacklight_iiif_search (~> 1.0)
- dry-monads (~> 1.4.0)
- hyrax (>= 2.5, < 4.0)
+ blacklight_iiif_search (>= 1.0, < 3.0)
+ derivative-rodeo (~> 0.5)
+ hyrax (>= 2.5, < 6)
  nokogiri (>= 1.13.2)
- rails (~> 5.0)
  rdf-vocab (~> 3.0)
- reform-rails (= 0.2.3)
 
 GIT
  remote: https://github.com/tawan/active-elastic-job.git
@@ -313,6 +311,15 @@ GEM
  declarative-option (0.1.0)
  deprecation (1.1.0)
  activesupport
+ derivative-rodeo (0.5.2)
+ activesupport (>= 5)
+ aws-sdk-s3
+ aws-sdk-sqs
+ httparty
+ marcel
+ mime-types
+ mini_magick
+ nokogiri
  devise (4.8.1)
  bcrypt (~> 3.0)
  orm_adapter (~> 0.1)
@@ -469,6 +476,9 @@ GEM
  hiredis (0.6.3)
  htmlentities (4.3.4)
  http_logger (0.7.0)
+ httparty (0.21.0)
+ mini_mime (>= 1.0.0)
+ multi_xml (>= 0.5.2)
  httpclient (2.8.3)
  hydra-access-controls (11.0.7)
  active-fedora (>= 10.0.0)

app/helpers/iiif_print_helper.rb

@@ -0,0 +1,5 @@
+# frozen_string_literal: true
+
+module IiifPrintHelper
+ include IiifPrint::IiifPrintHelperBehavior
+end

app/models/concerns/pdf_behavior.rb

@@ -25,6 +25,11 @@ module PdfBehavior
  end
 
  after_initialize :set_default_show_pdf_viewer, :set_default_show_pdf_download_button
+
+ include IiifPrint.model_configuration(
+ pdf_split_child_model: GenericWork,
+ pdf_splitter_service: IiifPrint::TenantConfig::PdfSplitter
+ )
  end
 
  private

app/presenters/hyku/work_show_presenter.rb

@@ -11,6 +11,12 @@ class WorkShowPresenter < Hyrax::WorkShowPresenter
  include Hyrax::IiifAv::DisplaysIiifAv
  Hyrax::MemberPresenterFactory.file_presenter_class = Hyrax::IiifAv::IiifFileSetPresenter
 
+ ##
+ # NOTE: IIIF Print prepends a IiifPrint::WorkShowPresenterDecorator to Hyrax::WorkShowPresenter
+ # However, with the above `include Hyrax::IiifAv::DisplaysIiifAv` we obliterate that logic. So
+ # we need to re-introduce that logic.
+ prepend IiifPrint::TenantConfig::WorkShowPresenterDecorator
+
  delegate :title_or_label, :extent, :show_pdf_viewer, :show_pdf_download_button, to: :solr_document
 
  # OVERRIDE Hyrax v2.9.0 here to make featured collections work
@@ -58,15 +64,6 @@ def user_can_feature_collection?
  end
  # End Featured Collections Methods
 
- # @return [Boolean] render a IIIF viewer
- def iiif_viewer?
- Hyrax.config.iiif_image_server? &&
- representative_id.present? &&
- representative_presenter.present? &&
- iiif_media? &&
- members_include_viewable?
- end
-
  def video_embed_viewer?
  extract_video_embed_presence
  end
@@ -96,16 +93,6 @@ def extract_video_embed_presence
  solr_document[:video_embed_tesim]&.all?(&:present?)
  end
 
- def iiif_media?(presenter: representative_presenter)
- presenter.image? || presenter.video? || presenter.audio?
- end
-
- def members_include_viewable?
- file_set_presenters.any? do |presenter|
- iiif_media?(presenter: presenter) && current_ability.can?(:read, presenter.id)
- end
- end
-
  def extract_from_identifier(rgx)
  if solr_document['identifier_tesim'].present?
  ref = solr_document['identifier_tesim'].map do |str|

app/services/iiif_print/tenant_config.rb

@@ -0,0 +1,202 @@
+# frozen_string_literal: true
+
+# rubocop:disable Metrics/LineLength
+module IiifPrint
+ ##
+ # This module encapsulates the logic for whether or not we'll use the IIIF Print services for the
+ # current tenant/account. The IIIF Print services does the following:
+ #
+ # - Skipping IIIF Print based derivative generation
+ # - Skipping PDF Splitting
+ # - Ignoring showing PDFs in the UV
+ #
+ # @note I am specifically isolating as much of this code into one module as possible, so that it
+ # it is hopefully easier to understand the configuration requirements and scope to this
+ # change. At some point, this might make sense to bring into IIIF Print directly.
+ #
+ # @see https://github.com/scientist-softserv/palni-palci/issues/656 palni-palci#656
+ # @see https://github.com/scientist-softserv/palni-palci/issues/657 palni-palci#657
+ # @see https://github.com/scientist-softserv/palni-palci/issues/658 palni-palci#658
+ # @see https://github.com/scientist-softserv/palni-palci/issues/659 palni-palci#659
+ module TenantConfig
+ ##
+ # When we were not planning on calling the underlying IiifPrint service but did due to some kind
+ # of faulty programming logic.
+ #
+ # @note This is raised as a guard to say "Hey, you thought you weren't using IIIF Print but your
+ # code's logic paths say otherwise."
+ class LeakyAbstractionError < StandardError
+ def initialize(klass:, method_name:)
+ super("Called #{klass}##{method_name} when we had said that #{klass} was not valid because we weren't using IIIF Print")
+ end
+ end
+
+ ##
+ # If the default PDF viewer (PDF.js) is enabled, this method returns false,
+ # meaning the application should not use IIIF Print. If the default viewer is
+ # disabled, this method returns true, meaning the application should use IIIF Print.
+ def self.use_iiif_print?
+ !::Flipflop.default_pdf_viewer?
+ end
+
+ ##
+ # This class implements the interface of the Hyrax::DerivativeService. It is responsible for
+ # negotiating whether or not the DerivativeService is "on" for the current tenant.
+ #
+ # @see https://github.com/samvera/hyrax/blob/08ef6c9a4fac489972eea9be53403e173f4ffb29/app/services/hyrax/derivative_service.rb Hyrax::DerivativeService
+ class DerivativeService
+ ##
+ # This allows you to specify the IIIF derivative service to use when the tenant has chosen to
+ # use IIIF Print for processing PDFs.
+ #
+ # If you are using the DerivativeRodeo, you'd specify something else.
+ class_attribute :iiif_service_class, default: ::IiifPrint::PluggableDerivativeService
+
+ def initialize(file_set)
+ @file_set = file_set
+ end
+
+ delegate :use_iiif_print?, to: TenantConfig
+
+ def valid?
+ return false unless use_iiif_print?
+
+ iiif_print_service_instance.valid?
+ end
+
+ %i[create_derivatives cleanup_derivatives].each do |method_name|
+ define_method(method_name) do |*args|
+ raise LeakyAbstractionError.new(klass: self.class, method_name: method_name) unless use_iiif_print?
+
+ iiif_print_service_instance.public_send(method_name, *args)
+ end
+ end
+
+ ##
+ # @api private
+ #
+ # @note Public to ease testing.
+ def iiif_print_service_instance
+ @iiif_print_service_instance ||= iiif_service_class.new(@file_set)
+ end
+ end
+
+ ##
+ # This is the pdf_splitter_service that will be used. If the tenant does not allow PDF splitting
+ # we will return an empty array.
+ #
+ # @example
+ #
+ # class MyWork
+ # include IiifPrint.model_configuration(
+ # pdf_split_child_model: Attachment,
+ # pdf_splitter_service: IiifPrint::TenantConfig::PdfSplitter,
+ # derivative_service_plugins: [ IiifPrint::TextExtractionDerivativeService ])
+ # end
+ #
+ # @see https://github.com/scientist-softserv/iiif_print/blob/9e7837ce4bd08bf8fff9126455d0e0e2602f6018/lib/iiif_print.rb#L86-L138 Documentation for configuring
+ # @see https://github.com/scientist-softserv/adventist-dl/blob/d7676bdac2c672f09b28086d7145b68306978950/app/models/image.rb#L14-L20 Example implementation
+ module PdfSplitter
+ mattr_accessor :iiif_print_splitter
+ self.iiif_print_splitter = ::IiifPrint::SplitPdfs::PagesToJpgsSplitter
+
+ ##
+ # @api public
+ def self.call(*args)
+ return [] unless TenantConfig.use_iiif_print?
+
+ iiif_print_splitter.call(*args)
+ end
+ end
+
+ ##
+ # @see https://github.com/scientist-softserv/iiif_print/blob/9e7837ce4bd08bf8fff9126455d0e0e2602f6018/lib/iiif_print/split_pdfs/child_work_creation_from_pdf_service.rb#L10-L46 Interface of FileSetActor#service
+ module SkipSplittingPdfService
+ ##
+ # @return [Symbol] Always :tenant_does_not_split_pdfs
+ def self.conditionally_enqueue(*_args)
+ :tenant_does_not_split_pdfs
+ end
+ end
+
+ ##
+ # This decorator should ensure that we don't call model configured :pdf_splitter_service as
+ # documented in {TenantConfig::PdfSplitter} and the IIIF Print gem. It avoids the potentially
+ # expensive conditionally enqueue logic of the super class.
+ #
+ # Why not make an `app/actors/hyrax/actors/file_set_actor_decorator.rb`? It would be lost in that
+ # it is decorating the decoration of the IIIF Print gem. Beside, in bringing this here, we have
+ # a relatively singular place for all of the configurations.
+ module FileSetActorDecorator
+ ##
+ # @see https://github.com/scientist-softserv/iiif_print/blob/9e7837ce4bd08bf8fff9126455d0e0e2602f6018/app/actors/iiif_print/actors/file_set_actor_decorator.rb#L33-L35 Method we're overriding
+ def service
+ return TenantConfig::SkipSplittingPdfService unless TenantConfig.use_iiif_print?
+
+ super
+ end
+ end
+
+ ##
+ # OVERRIDE IiifPrint::WorkShowPresenterDecorator
+ # OVERRIDE Hyrax::WorkShowPresenter
+ #
+ # In IiifPrint we overrided #members_include_viewable_image? to query for both file sets and
+ # child works. (Child works being the pages split off of a PDF)
+ #
+ # In Hyrax::WorkShowPresenter we're only looking at the underlying file_sets. But IiifPrint
+ # needs to look at multiple places.
+ module WorkShowPresenterDecorator
+ ##
+ # @return [Array<Symbol>] predicate methods (e.g. ending in "?") that reflect the types
+ # of files we want to consider for showing in the IIIF Viewer.
+ def iiif_media_predicates
+ if TenantConfig.use_iiif_print?
+ %i[image? audio? video? pdf?]
+ else
+ %i[image? audio? video?]
+ end
+ end
+
+ def iiif_media?(presenter: representative_presenter)
+ iiif_media_predicates.any? { |predicate| presenter.try(predicate) || presenter.try(:solr_document).try(predicate) }
+ end
+
+ ##
+ # @return [Boolean] render a IIIF viewer
+ #
+ # OVERRIDE Hyrax::WorkShowPresenter; this override introduces behavior to handle over-rides.
+ def iiif_viewer?
+ Hyrax.config.iiif_image_server? &&
+ representative_id.present? &&
+ representative_presenter.present? &&
+ iiif_media? &&
+ members_include_iiif_viewable?
+ end
+
+ def members_include_iiif_viewable?
+ iiif_presentable_member_presenters.any? do |presenter|
+ iiif_media?(presenter: presenter) && current_ability.can?(:read, presenter.id)
+ end
+ end
+
+ ##
+ # @return [Array<Object>] An array of presenter objects
+ #
+ # In a non-IIIF Print using scenario, we use the file_set_presenters value; that is for
+ # objects that are very specifically file_sets.
+ #
+ # In a IIIF Print using scenario, we use the ill-named 'file_set_ids_ssim', because a
+ # long-standing decision is that this field will have both file_set IDs and child work IDs.
+ def iiif_presentable_member_presenters
+ if TenantConfig.use_iiif_print?
+ presentable_member_ids = Array.wrap(solr_document.try(:file_set_ids) || solr_document.try(:[], 'file_set_ids_ssim'))
+ member_presenters_for(presentable_member_ids)
+ else
+ file_set_presenters
+ end
+ end
+ end
+ end
+end
+# rubocop:enable Metrics/LineLength

config/application.rb

@@ -60,5 +60,21 @@ class Application < Rails::Application
 
  Object.include(AccountSwitch)
  end
+
+ config.after_initialize do
+ ##
+ # The first "#valid?" service is the one that we'll use for generating derivatives.
+ Hyrax::DerivativeService.services = [
+ IiifPrint::TenantConfig::DerivativeService,
+ Hyrax::FileSetDerivativesService
+ ]
+
+ ##
+ # This needs to be in the after initialize so that the IiifPrint gem can do it's decoration.
+ #
+ # @see https://github.com/scientist-softserv/iiif_print/blob/9e7837ce4bd08bf8fff9126455d0e0e2602f6018/lib/iiif_print/engine.rb#L54 Where we do the override.
+ Hyrax::Actors::FileSetActor.prepend(IiifPrint::TenantConfig::FileSetActorDecorator)
+ Hyrax::WorkShowPresenter.prepend(IiifPrint::TenantConfig::WorkShowPresenterDecorator)
+ end
  end
 end

db/migrate/20231128191136_add_model_details_to_iiif_print_pending_relationships.iiif_print.rb

@@ -0,0 +1,8 @@
+# This migration comes from iiif_print (originally 20231110163052)
+class AddModelDetailsToIiifPrintPendingRelationships < ActiveRecord::Migration[5.2]
+ def change
+ add_column :iiif_print_pending_relationships, :parent_model, :string
+ add_column :iiif_print_pending_relationships, :child_model, :string
+ add_column :iiif_print_pending_relationships, :file_id, :string
+ end
+end

db/schema.rb

@@ -10,7 +10,7 @@
 #
 # It's strongly recommended that you check this file into your version control system.
 
-ActiveRecord::Schema.define(version: 2023_06_08_153601) do
+ActiveRecord::Schema.define(version: 2023_11_28_191136) do
 
  # These are extensions that must be enabled in order to support this database
  enable_extension "plpgsql"
@@ -382,6 +382,9 @@
  t.string "child_order", null: false
  t.datetime "created_at", null: false
  t.datetime "updated_at", null: false
+ t.string "parent_model"
+ t.string "child_model"
+ t.string "file_id"
  t.index ["parent_id"], name: "index_iiif_print_pending_relationships_on_parent_id"
  end
 

docker-compose.bundle.yml

@@ -0,0 +1,9 @@
+# Copy file to docker-compose.override.yml to override docker-compose.yml
+# Only use for local development
+version: '3.8'
+services:
+ web:
+ command: sh -l -c "bundle && bundle exec puma -v -b tcp://0.0.0.0:3000"
+ worker:
+ command: sh -l -c "bundle && bundle exec sidekiq"
+

spec/presenters/concerns/hyrax/iiif_av/displays_content_decorator_spec.rb

@@ -0,0 +1,15 @@
+# frozen_string_literal: true
+
+require 'spec_helper'
+
+RSpec.describe Hyrax::IiifAv::DisplaysContentDecorator do
+ # We're prepending the DisplaysContentDecorator to the Hyrax::IiifAv::DisplaysContent
+ describe Hyrax::IiifAv::DisplaysContent do
+ describe '.public_instance_methods' do
+ subject { Hyrax::IiifAv::DisplaysContent.public_instance_methods }
+
+ it { is_expected.to include(:solr_document) }
+ it { is_expected.to include(:current_ability) }
+ end
+ end
+end