Display VTT transcripts in audio/video players (#7418)

* Add transcript_ids property to file set model(s)

* Add transcript_ids to file set form

* Copy changes from @kirkkwang's draft PR, WIP: Implement Transcription Support in Hyrax #7380

https://github.com/samvera/hyrax/pull/7380/changes

* Support transcriptions for ActiveFedora works

* Fix broken specs

* Normalize VTT file set language values to 2-letter language codes.

* Simplify iiif_manifest_presenter_spec#annotates_content

* Add comments and rubocop fixes

* Add language field to file set form and move transcript ids form hint into simple form

* Add VTT transcripts to default audio/video partials

* Fix "l is not a function" error when displaying VTT transcripts in Annotations tab of Clover IIIF viewer.

"l" is supposed to refer to the <ItemStyled> object in Annotation (https://github.com/samvera-labs/clover-iiif/blob/main/src/components/Viewer/InformationPanel/Annotation/Item.styled.tsx). In the minified version of this file, this should actually be uppercase L, not lowercase l.

`l("span",{style:{backgroundImage` was changed to `L("span",{style:{backgroundImage`
`return l(n9,{dir:P,"data-format"` was changed to `return L(n9,{dir:P,"data-format"`

There appears to be an unnecessary call to `l()` in the following switch/case statement:
```
case "text/vtt":
  return l(HQ, {
    inlineCues: k,
    label: A,
    vttUri: ((D = y[0]) == null ? void 0 : D.id) || void 0
  });
```

In fact, `return l(HQ,{inlineCues:k` can be changed to `return HQ({inlineCues:k` since HQ is a function in the minified file.

* Fix "l is not a function" error for other file types. The implementation for other file types seems unfinished and doesn't work.

* Use the file set presenter to render the file set edit form instead of the solr document. Do not make language a required field for a file set.

* Allow transcriptions_controller to serve file types other than VTT. Fix some formatting and specs.

* Fix syntax error in clover.js

* Increase the height of Clover IIIF viewer so that thumbnails for works with multiple files are visible.

* Rename "transcriptions" to "transcripts" for consistency

* Account for ActiveTriples::Resource in a transcript's language field. Remove tests for code that was already tested/covered elsewhere.

* Remove unused fallback label, since a file set should always have a title or label

* Make language optional again in IIIF manifest annotations, especially since it's not required by the viewer.

* Delete a stray, unnecessary comment

* Rename .valid_transcripts to .available_transcripts

* Update file_set_form_helper_spec.rb

* Fixes for Koppie: use file_ids_ssim.first instead of original_file_id

* Fix and refactor file_set_form_helper_spec to use top-level context blocks (instead of describe)

* Rubocop fixes and add comment

* Refactor file set form

* Use cached parent for FileSetFormHelper instead of running an extra query. Rename form_transcript_ids_select_for to transcript_ids_select_options

* Add vtt file metadata to file_set_form_helper_spec

* When searching for available transcripts. use an fq filter to capture both FileSets and Hyrax::FileSets.

* Fix whitespace and add some clarifying comments.

* Authorize transcript before streaming the file

* Remove test for ActiveTriples::Resource in presenters/hyrax/displays_transcripts.

It was created for a hypothetical edge case in which someone might configure a controlled vocabulary field (e.g. Hyrax::ControlledVocabularies::Language) in an AF-based app. Removed since it doesn't refer to any existing use case in Hyrax.

* Make `transcript_ids` property conditional on Hyrax.config.file_set_include_metadata. Add `transcript_ids` to default flexible metadata profiles.

* Update metadata profiles to fix allinson and koppie builds

* Move the transcripts_ids form field to views/records/edit_fields so it can be optional for apps with flexible metadata.

* Move transcript_ids into the file set metadata schema

* Update clover.js to v.3.6.0, which officially fixes samvera-labs/clover-iiif#305

---------

Co-authored-by: Rob Kaufman <rob@notch8.com>

Author: eltiffster

.dassie/config/metadata_profiles/m3_profile.yaml

@@ -942,3 +942,17 @@ properties:
     view:
       render_as: faceted
       html_dl: true
+  transcript_ids:
+    available_on:
+      class:
+        - Hyrax::FileSet
+    indexing:
+      - transcript_ids_ssim
+    data_type: array
+    display_label:
+      default: Transcripts
+    form:
+      display: true
+      primary: false
+    range: http://www.w3.org/2001/XMLSchema#string
+    property_uri: http://vocabulary.samvera.org/ns#transcriptIds

.koppie/config/metadata_profiles/m3_profile.yaml

@@ -914,3 +914,17 @@ properties:
     view:
       render_as: faceted
       html_dl: true
+  transcript_ids:
+    available_on:
+      class:
+        - FileSet
+    indexing:
+      - transcript_ids_ssim
+    data_type: array
+    display_label:
+      default: Transcripts
+    form:
+      display: true
+      primary: false
+    range: http://www.w3.org/2001/XMLSchema#string
+    property_uri: http://vocabulary.samvera.org/ns#transcriptIds

app/assets/stylesheets/hyrax/_viewer.scss

@@ -11,9 +11,17 @@
   left: 0;
 }
 
+// Universal Viewer:
 // Show various footer, header, left and right panels on larger screens only
 @media only screen and (min-width: 640px) {
   .viewer-wrapper {
     height: 640px;
   }
+}
+
+// Clover IIIF Viewer:
+// Increase the height so that thumbnails below the player are visible
+// (for works with multiple audio/video files)
+.viewer-wrapper.clover-viewer-wrapper {
+  height: 775px;
 }

app/controllers/concerns/hyrax/works_controller_behavior.rb

@@ -139,12 +139,15 @@ def inspect_work
     end
 
     def manifest
-      headers['Access-Control-Allow-Origin'] = '*'
+      locale = params[:locale] || current_user&.preferred_locale || I18n.default_locale
+      I18n.with_locale(locale) do
+        headers['Access-Control-Allow-Origin'] = '*'
 
-      json = iiif_manifest_builder.manifest_for(presenter: iiif_manifest_presenter)
+        json = iiif_manifest_builder.manifest_for(presenter: iiif_manifest_presenter)
 
-      respond_to do |wants|
-        wants.any { render json: json }
+        respond_to do |wants|
+          wants.any { render json: json }
+        end
       end
     end
 

app/controllers/hyrax/transcripts_controller.rb

@@ -0,0 +1,31 @@
+# frozen_string_literal: true
+
+module Hyrax
+  class TranscriptsController < DownloadsController
+    def show
+      # Using the extracted text from the index is blocked
+      # by https://github.com/samvera/hyrax/issues/7410, so we
+      # need to get the original file instead.
+      file_metadata = find_file_metadata(file_set: Hyrax.query_service.find_by(id: params.require(:id)))
+      file = Hyrax.storage_adapter.find_by(id: file_metadata.file_identifier)
+
+      prepare_file_headers_valkyrie(metadata: file_metadata, file: file)
+      response.headers['Access-Control-Allow-Origin'] = '*'
+      send_file file.disk_path, data_options(file_metadata)
+    end
+
+    private
+
+    def disposition
+      'inline'
+    end
+
+    def data_options(file_metadata)
+      {
+        type: "#{file_metadata.mime_type}; charset=utf-8",
+        filename: file_metadata.original_filename,
+        disposition: disposition
+      }
+    end
+  end
+end

app/forms/concerns/hyrax/transcripts_behavior.rb

@@ -0,0 +1,40 @@
+# frozen_string_literal: true
+module Hyrax
+  module TranscriptsBehavior
+    extend ActiveSupport::Concern
+
+    class_methods do
+      def available_transcripts(parent:, current_ability:)
+         member_ids = Hyrax.custom_queries.find_child_file_set_ids(resource: parent)
+         Hyrax::SolrQueryService.new
+                                # Cast Valkyrie::IDs to strings
+                                .with_ids(ids: member_ids.map(&:to_s).to_a)
+                                .accessible_by(ability: current_ability, action: :edit)
+                                .solr_documents(
+                                  # Using "has_model_ssim:*FileSet" in fq will return both FileSet
+                                  # and Hyrax::FileSet documents. In test mode, Koppie and Sirenia 
+                                  # index file sets with has_model_ssim:Hyrax::FileSet. 
+                                  # But in dev mode, they index file sets with 
+                                  # has_model_ssim:FileSet instead. This query covers both cases.
+                                  fq: [mime_type_filter_query.to_s, "has_model_ssim:*FileSet"],
+                                  fl: "id,title_tesim",
+                                  rows: 1000
+                                )
+      end
+
+      private
+
+      def mime_type_filter_query
+        valid_mime_types.map { |type| "mime_type_ssi:\"#{type}\"" }.join(" OR ")
+      end
+
+      # According to IIIF, .srt and .ttml are also acceptable but may
+      # not be supported by viewers. Clover and Ramp are confirmed to work
+      # with .vtt. (https://iiif.io/api/cookbook/recipe/0219-using-caption-file/).
+      # When Hyrax supports Ramp, we may want to add "text/plain" (.srt) to this list.
+      def valid_mime_types
+        ["text/vtt"]
+      end
+    end
+  end
+end

app/forms/hyrax/forms/file_set_edit_form.rb

@@ -12,7 +12,7 @@ class FileSetEditForm
 
     self.terms = [:resource_type, :title, :creator, :contributor, :description,
                   :keyword, :license, :publisher, :date_created, :subject, :language,
-                  :identifier, :based_near, :related_url,
+                  :identifier, :based_near, :related_url, :transcript_ids,
                   :visibility_during_embargo, :visibility_after_embargo, :embargo_release_date,
                   :visibility_during_lease, :visibility_after_lease, :lease_expiration_date,
                   :visibility]

app/forms/hyrax/forms/file_set_form.rb

@@ -14,6 +14,7 @@ class FileSetForm < Hyrax::Forms::ResourceForm
       include Hyrax::ContainedInWorksBehavior
       include Hyrax::LeaseabilityBehavior
       include Hyrax::PermissionBehavior
+      include Hyrax::TranscriptsBehavior
 
       property :representative_id, type: Valkyrie::Types::String, writeable: false
       property :thumbnail_id, type: Valkyrie::Types::String, writeable: false

app/helpers/hyrax/file_set_form_helper.rb

@@ -0,0 +1,24 @@
+# frozen_string_literal: true
+
+module Hyrax
+  module FileSetFormHelper
+    def render_transcript_ids_field?(file_set)
+      return unless file_set.persisted?
+      return if @parent.nil?
+      case file_set
+      when ActiveFedora::Base
+        file_set.video? || file_set.audio?
+      when Valkyrie::Resource
+        service = Hyrax::FileSetTypeService.new(file_set: file_set)
+        service.video? || service.audio?
+      end
+    end
+
+    def transcript_ids_select_options
+      options = Forms::FileSetForm.available_transcripts(parent: @parent, current_ability: current_ability)
+      options.each_with_object({}) do |doc, hash|
+        hash[doc.title_or_label] = doc.id.to_s
+      end
+    end
+  end
+end

app/helpers/hyrax/hyrax_helper_behavior.rb

@@ -21,6 +21,7 @@ module HyraxHelperBehavior
     include Hyrax::FacetsHelper
     include Hyrax::AttributesHelper
     include Hyrax::WorksHelper
+    include Hyrax::FileSetFormHelper
 
     ##
     # @return [Array<String>] the list of admin sets available for creating works for this user