hyrax 5.1.0.pre.beta1 → 5.2.0

data/Dockerfile

@@ -1,50 +1,98 @@
-ARG ALPINE_VERSION=3.21
-ARG RUBY_VERSION=3.3.6
-
-FROM ruby:$RUBY_VERSION-alpine$ALPINE_VERSION AS hyrax-base
-
-ARG DATABASE_APK_PACKAGE="postgresql-dev"
-ARG EXTRA_APK_PACKAGES="git"
-ARG RUBYGEMS_VERSION=""
-
-RUN addgroup -S --gid 101 app && \
-  adduser -S -G app -u 1001 -s /bin/sh -h /app app
-
-RUN apk --no-cache upgrade && \
-  apk --no-cache add acl \
-  build-base \
-  curl \
-  gcompat \
-  imagemagick \
-  imagemagick-heic \
-  imagemagick-jpeg \
-  imagemagick-jxl \
-  imagemagick-pdf \
-  imagemagick-svg \
-  imagemagick-tiff \
-  imagemagick-webp \
-  jemalloc \
-  ruby-grpc \
-  tzdata \
-  nodejs \
-  yarn \
-  zip \
-  $DATABASE_APK_PACKAGE \
-  $EXTRA_APK_PACKAGES
+ARG DEBIAN_VERSION=bookworm
+ARG RUBY_VERSION=3.3
+
+FROM ruby:$RUBY_VERSION-$DEBIAN_VERSION AS hyrax-base
+
+RUN apt-get update && \
+    curl -sL "https://deb.nodesource.com/setup_20.x" | bash - && \
+    apt-get install -y --no-install-recommends \
+    acl \
+    build-essential \
+    curl \
+    exiftool \
+    ffmpeg \
+    ghostscript \
+    git \
+    less \
+    libgsf-1-dev \
+    libimagequant-dev \
+    libjemalloc2 \
+    libjpeg62-turbo-dev \
+    libopenjp2-7-dev \
+    libopenjp2-tools \
+    libpng-dev \
+    libpoppler-cpp-dev \
+    libpoppler-dev \
+    libpoppler-glib-dev \
+    libpoppler-private-dev \
+    libpoppler-qt5-dev \
+    libreoffice \
+    libreoffice-l10n-uk \
+    librsvg2-dev \
+    libtiff-dev \
+    libvips-dev \
+    libvips-tools \
+    libwebp-dev \
+    libxml2-dev \
+    lsof \
+    mediainfo \
+    netcat-openbsd \
+    nodejs \
+    perl \
+    poppler-utils \
+    postgresql-client \
+    rsync \
+    ruby-grpc \
+    screen \
+    tesseract-ocr \
+    tzdata \
+    vim \
+    zip \
+    && \
+    npm install --global yarn && \
+    apt-get clean && \
+    rm -rf /var/lib/apt/lists/* && \
+    ln -s /usr/lib/*-linux-gnu/libjemalloc.so.2 /usr/lib/libjemalloc.so.2 && \
+    echo "******** Packages Installed *********"
+
+RUN bash -x -c "\
+    if [ $(dpkg --print-architecture) = 'amd64' ]; then \
+      wget https://github.com/ImageMagick/ImageMagick/releases/download/7.1.1-47/ImageMagick-82572af-gcc-x86_64.AppImage -O magick \
+        && chmod a+x magick \
+        && ./magick --appimage-extract \
+        && mv squashfs-root/usr/etc/ImageMagick*  /etc \
+        && rm -rf squashfs-root/usr/share/doc \
+        && cp -rv squashfs-root/usr/*  /usr/local \
+        && rm -rf magick squashfs-root ; \
+    else \
+      wget https://github.com/ImageMagick/ImageMagick/archive/refs/tags/7.1.1-47.tar.gz \
+        && tar xf 7.1.1-47.tar.gz \
+        && cd ImageMagick* \
+        && ./configure \
+        && make install \
+        && ldconfig /usr/local/lib \
+        && cd $OLDPWD \
+        && rm -rf 7.1.1-47.tar.gz ImageMagick* ; \
+    fi \
+    && identify -version"
 
 RUN setfacl -d -m o::rwx /usr/local/bundle && \
-  gem update --silent --system $RUBYGEMS_VERSION
+    gem update --silent --system
 
-USER app
+RUN useradd -m -u 1001 -U -s /bin/bash --home-dir /app app && \
+    mkdir -p /app/samvera/hyrax-webapp && \
+    chown -R app:app /app && \
+    echo "export PATH=/app/samvera/hyrax-webapp/bin:${PATH}" >> /etc/bash.bashrc
 
-RUN mkdir -p /app/samvera/hyrax-webapp
+USER app
 WORKDIR /app/samvera/hyrax-webapp
 
-COPY --chown=1001:101 ./bin/*.sh /app/samvera/
+COPY --chown=1001 ./bin/*.sh /app/samvera/
 ENV PATH="/app/samvera:$PATH" \
     RAILS_ROOT="/app/samvera/hyrax-webapp" \
     RAILS_SERVE_STATIC_FILES="1" \
-    LD_PRELOAD="/usr/local/lib/libjemalloc.so.2"
+    LD_PRELOAD="/usr/lib/libjemalloc.so.2" \
+    MALLOC_CONF="dirty_decay_ms:1000,narenas:2,background_thread:true"
 
 ENTRYPOINT ["hyrax-entrypoint.sh"]
 CMD ["bundle", "exec", "puma", "-v", "-b", "tcp://0.0.0.0:3000"]
@@ -55,22 +103,23 @@ FROM hyrax-base AS hyrax
 ARG APP_PATH=.
 ARG BUNDLE_WITHOUT="development test"
 
-ONBUILD COPY --chown=1001:101 $APP_PATH /app/samvera/hyrax-webapp
+ONBUILD COPY --chown=1001 $APP_PATH /app/samvera/hyrax-webapp
 ONBUILD RUN bundle install --jobs "$(nproc)"
 ONBUILD RUN RAILS_ENV=production SECRET_KEY_BASE=`bin/rake secret` DATABASE_URL='nulldb://nulldb' bundle exec rake assets:precompile
+ARG BUILD_GITSHA
+ARG BUILD_TIMESTAMP
+ENV BUILD_GITSHA=$BUILD_GITSHA \
+    BUILD_TIMESTAMP=$BUILD_TIMESTAMP
 
 
 FROM hyrax-base AS hyrax-worker-base
-
 USER root
-RUN apk --no-cache add bash \
-  ffmpeg \
-  mediainfo \
-  openjdk17-jre \
-  perl
-USER app
 
-RUN mkdir -p /app/fits && \
+RUN apt update && \
+    apt install -y --no-install-recommends default-jre-headless && \
+    apt-get clean && \
+    rm -rf /var/lib/apt/lists/* && \
+    mkdir -p /app/fits && \
     cd /app/fits && \
     wget https://github.com/harvard-lts/fits/releases/download/1.6.0/fits-1.6.0.zip -O fits.zip && \
     unzip fits.zip && \
@@ -90,6 +139,10 @@ ARG BUNDLE_WITHOUT="development test"
 ONBUILD COPY --chown=1001:101 $APP_PATH /app/samvera/hyrax-webapp
 ONBUILD RUN bundle install --jobs "$(nproc)"
 ONBUILD RUN RAILS_ENV=production SECRET_KEY_BASE=`bin/rake secret` DATABASE_URL='nulldb://nulldb' bundle exec rake assets:precompile
+ARG BUILD_GITSHA
+ARG BUILD_TIMESTAMP
+ENV BUILD_GITSHA=$BUILD_GITSHA \
+    BUILD_TIMESTAMP=$BUILD_TIMESTAMP
 
 
 FROM hyrax-worker-base AS hyrax-engine-dev
@@ -98,8 +151,9 @@ USER app
 ARG BUNDLE_WITHOUT=
 ENV HYRAX_ENGINE_PATH=/app/samvera/hyrax-engine
 
-COPY --chown=1001:101 .dassie /app/samvera/hyrax-webapp
-COPY --chown=1001:101 . /app/samvera/hyrax-engine
+COPY --chown=1001 .dassie /app/samvera/hyrax-webapp
+COPY --chown=1001 .koppie /app/samvera/hyrax-koppie
+COPY --chown=1001 . /app/samvera/hyrax-engine
 
 RUN bundle -v && \
   BUNDLE_GEMFILE=Gemfile.dassie bundle install --jobs "$(nproc)" && yarn && \
@@ -107,4 +161,8 @@ RUN bundle -v && \
   yarn cache clean
 
 ENTRYPOINT ["dev-entrypoint.sh"]
-CMD ["bundle", "exec", "puma", "-v", "-b", "tcp://0.0.0.0:3000"]
+CMD ["bundle", "exec", "puma", "-v", "-b", "tcp://0.0.0.0:3000"]
+ARG BUILD_GITSHA
+ARG BUILD_TIMESTAMP
+ENV BUILD_GITSHA=$BUILD_GITSHA \
+    BUILD_TIMESTAMP=$BUILD_TIMESTAMP

data/Gemfile

@@ -20,5 +20,6 @@ group :development, :test do
   gem 'rspec'
   gem 'ruby-prof', require: false
   gem 'semaphore_test_boosters'
-  gem "simplecov", require: false
+  gem 'simplecov', require: false
+  gem 'timecop'
 end

data/app/controllers/concerns/hyrax/valkyrie_downloads_controller_behavior.rb

@@ -102,6 +102,7 @@ module Hyrax
     def find_file_metadata(file_set:, use: :original_file, mime_type: nil)
       if mime_type.nil?
         use = :thumbnail_file if use == :thumbnail
+        use = :original_file unless Hyrax::FileMetadata::Use.keys.include?(use)
         use = Hyrax::FileMetadata::Use.uri_for(use: use)
         results = Hyrax.custom_queries.find_many_file_metadata_by_use(resource: file_set, use: use)
       else

data/app/controllers/concerns/hyrax/works_controller_behavior.rb

@@ -359,8 +359,9 @@ module Hyrax
       # intersection, we get the files they added via BrowseEverything
       # that they have not removed from the upload widget.
       uploaded_files = params.fetch(:uploaded_files, [])
+      params.permit(selected_files: [:expires, :file_name, :url])
       selected_files = params.fetch(:selected_files, {}).values
-                             .map { |f| f.permit(:expires, :file_name, :url, {}) }
+
       browse_everything_urls = uploaded_files &
                                selected_files.map { |f| f[:url] }
 

data/app/controllers/hyrax/admin/analytics/work_reports_controller.rb

@@ -14,9 +14,9 @@ module Hyrax
           @works_count = @accessible_works.count
           @top_works = paginate(top_works_list, rows: 10)
           @top_file_set_downloads = paginate(top_files_list, rows: 10)
-
-          @pageviews = Hyrax::Analytics.daily_events('work-view'), @downloads = Hyrax::Analytics.daily_events('file-set-download') if current_user.ability.admin?
-
+          # rubocop:disable Style/ParallelAssignment
+          @pageviews, @downloads = Hyrax::Analytics.daily_events('work-view'), Hyrax::Analytics.daily_events('file-set-download') if current_user.ability.admin?
+          # rubocop:enable Style/ParallelAssignment
           respond_to do |format|
             format.html
             format.csv { export_data }
@@ -34,7 +34,7 @@ module Hyrax
           end
         end
 
-  private
+        private
 
         def accessible_works
           models = Hyrax::ModelRegistry.work_rdf_representations.map { |m| "\"#{m}\"" }

data/app/controllers/hyrax/file_sets_controller.rb

@@ -9,6 +9,10 @@ module Hyrax
 
     before_action :authenticate_user!, except: [:show, :citation, :stats]
     load_and_authorize_resource class: Hyrax.config.file_set_class
+    # If Hyrax.config.file_set_class is set to ::FileSet, the load above will force-cast
+    # the instance as a ::FileSet, even if it is a Hyrax::FileSet. Re-cast it back to
+    # prevent method errors and nil objects later
+    before_action :cast_file_set
     before_action :build_breadcrumbs, only: [:show, :edit, :stats]
     before_action do
       blacklight_config.track_search_session = false
@@ -159,6 +163,13 @@ module Hyrax
       attrs
     end
 
+    def cast_file_set
+      return unless @file_set.class == ::FileSet
+      # We can tell if a Hyrax::FileSet was improperly cast because this AF method will
+      # return nil since its parent is not a ActiveFedora work.
+      @file_set = @file_set.valkyrie_resource if @file_set.respond_to?(:parent) && @file_set.parent&.id.nil?
+    end
+
     def parent(file_set: curation_concern)
       @parent ||=
         case file_set

data/app/helpers/hyrax/hyrax_helper_behavior.rb

@@ -68,10 +68,10 @@ module Hyrax
       mailbox = UserMailbox.new(user)
       unread_notifications = mailbox.unread_count
       link_to(hyrax.notifications_path,
-              'aria-label' => mailbox.label(params[:locale]),
+              'aria-description' => mailbox.label(params[:locale]),
               class: 'notify-number nav-link') do
         capture do
-          concat tag.span('', class: 'fa fa-bell')
+          concat tag.span('', class: 'fa fa-bell', 'aria-label': t('hyrax.admin.sidebar.notifications'))
           concat "\n"
           concat tag.span(unread_notifications,
                              class: count_classes_for(unread_notifications))

data/app/helpers/hyrax/trophy_helper.rb

@@ -15,7 +15,7 @@ module Hyrax
       args[:data]['remove-text'] = args[:remove_text]
 
       args[:data][:url] = hyrax.trophy_work_path(id)
-      link_to '#', id: 'action-highlight-work', class: args[:class], data: args[:data] do
+      link_to '#', id: 'action-highlight-work', role: args[:role], class: args[:class], data: args[:data] do
         yield(text)
       end
     end

data/app/jobs/concerns/hyrax/queued_job_behavior.rb

@@ -0,0 +1,22 @@
+# frozen_string_literal: true
+module Hyrax
+  # Grants the user's edit access on the provided FileSet
+  module QueuedJobBehavior
+    extend ActiveSupport::Concern
+
+    included do
+      queue_as Hyrax.config.ingest_queue_name
+      cattr_accessor :requeue_frequency
+    end
+
+    private
+
+    def redis_queue
+      Valkyrie::IndexingAdapter.find(:redis_queue)
+    end
+
+    def requeue(**args)
+      self.class.set(wait_until: (self.class.requeue_frequency || 5.minutes).from_now).perform_later(**args)
+    end
+  end
+end

data/app/jobs/hyrax/propagate_change_depositor_job.rb

@@ -26,7 +26,7 @@ module Hyrax
     def apply_depositor_metadata(resource, depositor)
       depositor_id = depositor.respond_to?(:user_key) ? depositor.user_key : depositor
       resource.depositor = depositor_id if resource.respond_to? :depositor=
-      Hyrax::AccessControlList.new(resource: resource).grant(:edit).to(::User.find_by_user_key(depositor_id)).save
+      resource.permission_manager.acl.grant(:edit).to(::User.find_by_user_key(depositor_id)).save
     end
   end
 end

data/app/jobs/hyrax/queued_delete_job.rb

@@ -0,0 +1,11 @@
+# frozen_string_literal: true
+module Hyrax
+  class QueuedDeleteJob < ApplicationJob
+    include QueuedJobBehavior
+
+    def perform(size: 200)
+      redis_queue.delete_queue(size: size)
+      requeue(size: size)
+    end
+  end
+end

data/app/jobs/hyrax/queued_indexing_job.rb

@@ -0,0 +1,11 @@
+# frozen_string_literal: true
+module Hyrax
+  class QueuedIndexingJob < ApplicationJob
+    include QueuedJobBehavior
+
+    def perform(size: 200)
+      redis_queue.index_queue(size: size)
+      requeue(size: size)
+    end
+  end
+end

data/app/jobs/migrate_files_to_valkyrie_job.rb

@@ -2,6 +2,11 @@
 # Responsible for conditionally enqueuing the file and thumbnail migration
 # logic of an ActiveFedora object.
 class MigrateFilesToValkyrieJob < Hyrax::ApplicationJob
+  # Define a logger for this job
+  def logger
+    FileUtils.mkdir_p(Hyrax.config.working_path)
+    @logger ||= Logger.new(Hyrax.config.working_path.join('migrate_files_to_valkyrie_job.log'))
+  end
   ##
   #
   # @param resource [Hyrax::FileSet]
@@ -33,6 +38,7 @@ class MigrateFilesToValkyrieJob < Hyrax::ApplicationJob
     # @todo should we trigger a job if the member is a child work?
     paths = Hyrax::DerivativePath.derivatives_for_reference(resource)
     paths.each do |path|
+      next unless File.size?(path) # skip blank files
       container = container_for(path)
       mime_type = Marcel::MimeType.for(extension: File.extname(path))
       directives = { url: path, container: container, mime_type: mime_type }
@@ -50,27 +56,33 @@ class MigrateFilesToValkyrieJob < Hyrax::ApplicationJob
 
     files = Hyrax.custom_queries.find_many_file_metadata_by_ids(ids: resource.file_ids)
     files.each do |file|
-      # If it doesn't start with fedora, we've likely already migrated it.
-      next unless /^fedora:/.match?(file.file_identifier.to_s)
-      resource.file_ids.delete(file.id)
+      begin
+        # If it doesn't start with fedora, we've likely already migrated it.
+        next unless /^fedora:/.match?(file.file_identifier.to_s)
+        resource.file_ids.delete(file.id)
 
-      Tempfile.create do |tempfile|
-        tempfile.binmode
-        tempfile.write(URI.open(file.file_identifier.to_s.gsub("fedora:", "http:")).read)
-        tempfile.rewind
+        Tempfile.create do |tempfile|
+          tempfile.binmode
+          tempfile.write(URI.open(file.file_identifier.to_s.gsub("fedora:", "http:")).read)
+          tempfile.rewind
 
-        # valkyrie_file = Hyrax.storage_adapter.upload(resource: resource, file: tempfile, original_filename: file.original_filename)
-        valkyrie_file = Hyrax::ValkyrieUpload.file(
-          filename: resource.label,
-          file_set: resource,
-          io: tempfile,
-          use: file.pcdm_use.select {|use| Hyrax::FileMetadata::Use.use_list.include?(use)},
-          user: User.find_or_initialize_by(User.user_key_field => resource.depositor),
-          mime_type: file.mime_type,
-          skip_derivatives: true
-        )
-        valkyrie_file = copy_attributes(valkyrie_file:, original_file: file)
-        Hyrax.persister.save(resource: valkyrie_file)
+          # valkyrie_file = Hyrax.storage_adapter.upload(resource: resource, file: tempfile, original_filename: file.original_filename)
+          valkyrie_file = Hyrax::ValkyrieUpload.file(
+            filename: resource.label,
+            file_set: resource,
+            io: tempfile,
+            use: file.pcdm_use.select {|use| Hyrax::FileMetadata::Use.use_list.include?(use)},
+            user: User.find_or_initialize_by(User.user_key_field => resource.depositor),
+            mime_type: file.mime_type,
+            skip_derivatives: true
+          )
+          valkyrie_file = copy_attributes(valkyrie_file:, original_file: file)
+          Hyrax.persister.save(resource: valkyrie_file)
+        end
+      rescue StandardError => e
+        # Log errors specific to file migration
+        logger.error("Error migrating file #{file.id} for resource #{resource.id}: #{e.message}")
+        logger.error(e.backtrace.join("\n"))
       end
     end
     # reindex the file set after migrating files to include characterization info
@@ -93,9 +105,9 @@ class MigrateFilesToValkyrieJob < Hyrax::ApplicationJob
   #
   # @param filename [String] the name of the derivative file: i.e. 'x-thumbnail.jpg'
   # @return [String]
-  def container_for(filename)
+  def container_for(path)
     # we want the portion between the '-' and the '.'
-    file_blob = File.basename(filename, '.*').split('-').last
+    file_blob = File.basename(path, '.*').split('-').last
 
     case file_blob
     when 'thumbnail'

data/app/jobs/migrate_sipity_entity_job.rb

@@ -0,0 +1,21 @@
+# frozen_string_literal: true
+
+# migrates a resource's sipity entity so it can be found
+class MigrateSipityEntityJob < ApplicationJob
+  # input [String] id of a migrated resource
+  def perform(id:)
+    resource = Hyrax.query_service.find_by(id: id)
+    new_gid = Hyrax::GlobalID(resource).to_s
+    return if Sipity::Entity.find_by(proxy_for_global_id: new_gid)
+
+    work = resource.internal_resource.constantize.find(id)
+    original_gid = Hyrax::GlobalID(work).to_s
+    return if new_gid == original_gid
+    original_entity = Sipity::Entity.find_by(proxy_for_global_id: original_gid)
+    return if original_entity.nil?
+    original_entity.update(proxy_for_global_id: new_gid)
+  rescue ActiveFedora::ObjectNotFoundError
+    # this happens when the resource was never in Fedora so there is nothing to migrate.
+    # We don't want to retry the job so we don't raise an error.
+  end
+end

data/app/models/concerns/hyrax/ability.rb

@@ -404,8 +404,10 @@ module Hyrax
                                     .select(:source_id)
                                     .distinct
                                     .pluck(:source_id)
-
-      Hyrax.custom_queries.find_ids_by_model(model: Hyrax::AdministrativeSet, ids: ids).any?
+      return false if ids.empty?
+      Hyrax::SolrQueryService.new.with_ids(ids: ids).query_result(rows: 1000)['response']['docs'].any? do |doc|
+        (Hyrax::ModelRegistry.admin_set_rdf_representations & doc['has_model_ssim']).present?
+      end
     end
 
     def registered_user?

data/app/models/concerns/hyrax/solr_document_behavior.rb

@@ -82,9 +82,12 @@ module Hyrax
 
     # Method to return the model
     def hydra_model(classifier: nil)
+      # finds the model from the solr document
       model = first('has_model_ssim')&.safe_constantize
-      model = (first('has_model_ssim')&.+ 'Resource')&.safe_constantize if Hyrax.config.valkyrie_transition?
-      model || model_classifier(classifier).classifier(self).best_model
+      # this returns nil if it isn't a valid model
+      resource_model = (first('has_model_ssim')&.+ 'Resource')&.safe_constantize if Hyrax.config.valkyrie_transition?
+      # if valkyrie_transition is enabled, we generally want to use the resource model if it exists
+      resource_model || model || model_classifier(classifier).classifier(self).best_model
     end
 
     def depositor(default = '')

data/app/models/hyrax/file_metadata.rb

@@ -44,16 +44,31 @@ module Hyrax
 
       # @return [Array<RDF::URI>] list of all uses
       def use_list
-        [ORIGINAL_FILE,
-         THUMBNAIL_IMAGE,
-         EXTRACTED_TEXT,
-         INTERMEDIATE_FILE,
-         PRESERVATION_FILE,
-         SERVICE_FILE,
-         TRANSCRIPT]
+        [
+          EXTRACTED_TEXT,
+          INTERMEDIATE_FILE,
+          ORIGINAL_FILE,
+          PRESERVATION_FILE,
+          SERVICE_FILE,
+          THUMBNAIL_IMAGE,
+          TRANSCRIPT
+        ]
       end
       module_function :use_list
 
+      def keys
+        [
+          :extracted_file,
+          :intermediate_file,
+          :original_file,
+          :preservation_file,
+          :service_file,
+          :thumbnail_file,
+          :transcript_file
+        ]
+      end
+      module_function :keys
+
       ##
       # @param use [RDF::URI, Symbol]
       #

data/app/services/hyrax/analytics/ga4/base.rb

@@ -88,7 +88,7 @@ module Hyrax
         end
 
         def unwrap_metric(metric)
-          metric.metric_values.first.value.to_i
+          metric&.metric_values&.first&.value.to_i
         end
       end
     end

data/app/services/hyrax/analytics/ga4.rb

@@ -76,7 +76,11 @@ module Hyrax
           end
 
           def account_info
-            @account_info ||= JSON.parse(account_json_string)
+            @account_info ||= if account_json_string.is_a? Hash
+                                account_json_string
+                              else
+                                JSON.parse(account_json_string)
+                              end
           end
 
           KEYS.each do |key|

data/app/services/hyrax/change_depositor_service.rb

@@ -63,7 +63,7 @@ module Hyrax
     def self.apply_depositor_metadata(resource, depositor)
       depositor_id = depositor.respond_to?(:user_key) ? depositor.user_key : depositor
       resource.depositor = depositor_id if resource.respond_to? :depositor=
-      Hyrax::AccessControlList.new(resource: resource).grant(:edit).to(::User.find_by_user_key(depositor_id)).save
+      resource.permission_manager.acl.grant(:edit).to(::User.find_by_user_key(depositor_id)).save
     end
     private_class_method :apply_depositor_metadata
   end

data/app/services/hyrax/characterization/valkyrie_characterization_service.rb

@@ -107,13 +107,19 @@ class Hyrax::Characterization::ValkyrieCharacterizationService
   end
 
   # Assign values of the instance properties from the metadata mapping :prop => val
+  # @todo push exceptional per-property behavior into the mapping somehow?
   # @return [Hash]
   def apply_metadata(terms)
-    terms.each_pair do |term, value|
-      property = property_for(term)
-      next if property.nil?
-
-      Array(value).each { |v| append_property_value(property, v) }
+    values_by_property(terms).each do |property, values|
+      value = if property == :mime_type
+                values.last
+              elsif [:height, :width].include?(property)
+                # keep only the max height or width
+                values.map(&:to_i).max.to_s
+              else
+                values
+              end
+      metadata.public_send("#{property}=", value)
     end
   end
 
@@ -127,13 +133,15 @@ class Hyrax::Characterization::ValkyrieCharacterizationService
     end
   end
 
-  ##
-  # @todo push exceptional per-property behavior into the mapping somehow?
-  def append_property_value(property, value)
-    # We don't want multiple mime_types; this overwrites each time to accept last value
-    value = Array(metadata.public_send(property)) + [value] unless property == :mime_type
-    # We don't want multiple heights / widths, pick the max
-    value = value.map(&:to_i).max.to_s if property == :height || property == :width
-    metadata.public_send("#{property}=", value)
+  # Map each term to the corresponding property
+  # (multiple terms can map to the same property),
+  # and gather all values for each property
+  # @return [Hash]
+  def values_by_property(terms)
+    terms.each_with_object({}) do |(term, value), property_values|
+      next unless (property = property_for(term))
+
+      (property_values[property] ||= []).concat(Array(value))
+    end
   end
 end