hyrax 4.0.0 → 5.0.0.rc2

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

@@ -15,6 +15,11 @@ class Hyrax::Characterization::ValkyrieCharacterizationService
     new(metadata: metadata, file: file, **options).characterize
     saved = Hyrax.persister.save(resource: metadata)
     Hyrax.publisher.publish('file.metadata.updated', metadata: saved, user: user)
+
+    Hyrax.publisher.publish('file.characterized',
+                            file_set: Hyrax.query_service.find_by(id: saved.file_set_id),
+                            file_id: saved.id.to_s,
+                            path_hint: saved.file_identifier.to_s)
   end
 
   ##
@@ -27,25 +32,27 @@ class Hyrax::Characterization::ValkyrieCharacterizationService
   # @!attribute [rw] source
   #   @return [Valkyrie::StorageAdapter::StreamFile]
   # @!attribute [rw] tools
-  #   @todo clarify what this is for. it gets passed to {#characterize}
-  #     on the characterizer, but it's not clear how to customize
-  #     effectively.
+  #   can be :fits, :fits_servlet, :ffprobe or any other service added to HydraFileCharacterization
+  #   note that ffprope is faster but only works on AV files.
   #   @return [Symbol]
   attr_accessor :mapping, :metadata, :parser, :source, :tools
 
   ##
   # @api private
-  def initialize(metadata:, # rubocop:disable Metrics/ParameterLists
-                 file:,
-                 characterizer: Hydra::FileCharacterization,
-                 parser_mapping: Hydra::Works::Characterization.mapper,
-                 parser: Hydra::Works::Characterization::FitsDocument.new)
+  def initialize( # rubocop:disable Metrics/ParameterLists
+    metadata:,
+    file:,
+    characterizer: Hydra::FileCharacterization,
+    parser_mapping: Hydra::Works::Characterization.mapper,
+    parser: Hydra::Works::Characterization::FitsDocument.new,
+    ch12n_tool: :fits
+  )
     @characterizer = characterizer
     @metadata      = metadata
     @source        = file
     @mapping       = parser_mapping
     @parser        = parser
-    @tools         = :fits
+    @tools         = ch12n_tool
   end
 
   ##

data/app/services/hyrax/collections/collection_member_service.rb

@@ -106,7 +106,7 @@ module Hyrax
         def add_member(collection_id:, new_member:, user:)
           message = Hyrax::MultipleMembershipChecker.new(item: new_member).check(collection_ids: [collection_id], include_current_members: true)
           raise Hyrax::SingleMembershipError, message if message.present?
-          new_member.member_of_collection_ids << collection_id # only populate this direction
+          new_member.member_of_collection_ids += [collection_id] # only populate this direction
           new_member = Hyrax.persister.save(resource: new_member)
           publish_metadata_updated(new_member, user)
           new_member

data/app/services/hyrax/custom_queries/find_by_date_range.rb

@@ -0,0 +1,55 @@
+# frozen_string_literal: true
+module Hyrax
+  module CustomQueries
+    ##
+    # @see https://github.com/samvera/valkyrie/wiki/Queries#custom-queries
+    class FindByDateRange
+      def self.queries
+        [:find_by_date_range]
+      end
+
+      def initialize(query_service:)
+        @query_service = query_service
+      end
+
+      attr_reader :query_service
+      delegate :resource_factory, to: :query_service
+      delegate :orm_class, to: :resource_factory
+
+      ##
+      # @note this is an unoptimized default implementation of this custom
+      #   query. it's Hyrax's policy to provide such implementations of custom
+      #   queries in use for cross-compatibility of Valkyrie query services.
+      #   it's advisable to provide an optimized query for the specific adapter.
+      #
+      # @param models [Array]
+      # @param start_datetime [DateTime]
+      # @param end_datetime [DateTime]
+      def find_by_date_range(start_datetime:, end_datetime: nil, models: nil)
+        end_datetime = 1.second.since(Time.zone.now) if end_datetime.blank?
+        if models.present?
+          query_service.run_query(find_models_by_date_range_query, start_datetime.to_s, end_datetime.to_s, models)
+        else
+          query_service.run_query(find_by_date_range_query, start_datetime.to_s, end_datetime.to_s)
+        end
+      end
+
+      def find_models_by_date_range_query
+        <<-SQL
+          SELECT * FROM orm_resources
+          WHERE created_at >= ?
+          AND created_at <= ?
+          AND internal_resource IN (?);
+        SQL
+      end
+
+      def find_by_date_range_query
+        <<-SQL
+          SELECT * FROM orm_resources
+          WHERE created_at >= ?
+          AND created_at <= ?;
+        SQL
+      end
+    end
+  end
+end

data/app/services/hyrax/custom_queries/find_count_by.rb

@@ -0,0 +1,62 @@
+# frozen_string_literal: true
+module Hyrax
+  module CustomQueries
+    ##
+    # @see https://github.com/samvera/valkyrie/wiki/Queries#custom-queries
+    class FindCountBy
+      def self.queries
+        [:find_count_by]
+      end
+
+      def initialize(query_service:)
+        @query_service = query_service
+      end
+
+      attr_reader :query_service
+      delegate :resource_factory, to: :query_service
+      delegate :orm_class, to: :resource_factory
+
+      ##
+      # @note this is an unoptimized default implementation of this custom
+      #   query. it's Hyrax's policy to provide such implementations of custom
+      #   queries in use for cross-compatibility of Valkyrie query services.
+      #   it's advisable to provide an optimized query for the specific adapter.
+      #
+      # @param hash [Hash] the hash representation of the query
+      def find_count_by(hash = {}, models: nil)
+        return nil if models.empty? && hash.blank?
+
+        internal_array = ["{ #{hash.map { |k, v| "\"#{k}\": #{v}" }.join(', ')} }"] if hash.present?
+        if models.empty?
+          query_service.orm_class.count_by_sql(([find_count_by_properties_query] + internal_array))
+        elsif hash.blank?
+          query_service.orm_class.count_by_sql([find_count_by_models_query] + [models])
+        else
+          query_service.orm_class.count_by_sql(([find_count_by_properties_and_models_query] + internal_array + [models]))
+        end
+      end
+
+      def find_count_by_properties_and_models_query
+        <<-SQL
+          SELECT count(*) FROM orm_resources
+          WHERE metadata @> ?
+          AND internal_resource IN (?);
+        SQL
+      end
+
+      def find_count_by_models_query
+        <<-SQL
+          SELECT count(*) FROM orm_resources
+          WHERE internal_resource IN (?);
+        SQL
+      end
+
+      def find_count_by_properties_query
+        <<-SQL
+          SELECT count(*) FROM orm_resources
+          WHERE metadata @> ?;
+        SQL
+      end
+    end
+  end
+end

data/app/services/hyrax/custom_queries/find_file_metadata.rb

@@ -67,7 +67,7 @@ module Hyrax
         return [] unless resource.try(:file_ids)
 
         results = find_many_file_metadata_by_ids(ids: resource.file_ids)
-        results.select { |fm| fm.type.include?(use) }
+        results.select { |fm| fm.pcdm_use.include?(use) }
       end
     end
   end

data/app/services/hyrax/custom_queries/find_models_by_access.rb

@@ -0,0 +1,59 @@
+# frozen_string_literal: true
+module Hyrax
+  module CustomQueries
+    ##
+    # @see https://github.com/samvera/valkyrie/wiki/Queries#custom-queries
+    class FindModelsByAccess
+      def self.queries
+        [:find_models_by_access]
+      end
+
+      def initialize(query_service:)
+        @query_service = query_service
+      end
+
+      attr_reader :query_service
+      delegate :resource_factory, to: :query_service
+      delegate :orm_class, to: :resource_factory
+
+      ##
+      # @note this is an unoptimized default implementation of this custom
+      #   query. it's Hyrax's policy to provide such implementations of custom
+      #   queries in use for cross-compatibility of Valkyrie query services.
+      #   it's advisable to provide an optimized query for the specific adapter.
+      #
+      # @param model [Class]
+      # @param ids [Enumerable<#to_s>, Symbol]
+      #
+      def find_models_by_access(mode:, models: nil, agent:, group: nil)
+        agent = "group/#{agent}" if group.present?
+        internal_array = "{\"permissions\": [{\"mode\": \"#{mode}\", \"agent\": \"#{agent}\"}]}"
+        if models.present?
+          query_service.run_query(find_models_by_access_query, internal_array, models)
+        else
+          query_service.run_query(find_by_access_query, internal_array)
+        end
+      end
+
+      def find_models_by_access_query
+        <<-SQL
+          SELECT * FROM orm_resources
+          WHERE id IN (
+            SELECT uuid(metadata::json#>'{access_to,0}'->>'id') FROM orm_resources
+            WHERE metadata @> ?
+          ) AND internal_resource IN (?);
+        SQL
+      end
+
+      def find_by_access_query
+        <<-SQL
+          SELECT * FROM orm_resources
+          WHERE id IN (
+            SELECT uuid(metadata::json#>'{access_to,0}'->>'id') FROM orm_resources
+            WHERE metadata @> ?
+          );
+        SQL
+      end
+    end
+  end
+end

data/app/services/hyrax/derivative_bucketed_storage.rb

@@ -0,0 +1,25 @@
+# frozen_string_literal: true
+
+# this class overrides the Valkyrie::Storage::Disk::BucketStorage class so that file paths match
+module Hyrax
+  class DerivativeBucketedStorage
+    attr_reader :base_path
+
+    def initialize(base_path:)
+      @base_path = base_path
+    end
+
+    # rubocop:disable Lint/UnusedMethodArgument
+    def generate(resource:, file:, original_filename:)
+      raise ArgumentError, "original_filename must be provided" unless original_filename
+      Pathname.new(base_path).join(*bucketed_path(resource.id)).join(original_filename)
+    end
+    # rubocop:enable Lint/UnusedMethodArgument
+
+    def bucketed_path(id)
+      # We want to use the same code the derivative process uses so that items end up
+      # stored in the place we expect them.
+      Hyrax::DerivativePath.new(id.to_s).pair_directory
+    end
+  end
+end

data/app/services/hyrax/derivative_path.rb

@@ -35,6 +35,18 @@ module Hyrax
       end
     end
 
+    def pairs
+      @pairs ||= id.split('').each_slice(2).map(&:join)
+    end
+
+    def pair_directory
+      pairs[0..-2]
+    end
+
+    def pair_path
+      (pair_directory + pairs[-1..-1]).join('/')
+    end
+
     private
 
     # @return [String] Returns the root path where derivatives will be generated into.
@@ -47,10 +59,6 @@ module Hyrax
       Pathname.new(Hyrax.config.derivatives_path).join(pair_path)
     end
 
-    def pair_path
-      id.split('').each_slice(2).map(&:join).join('/')
-    end
-
     def file_name
       return unless destination_name
       destination_name + extension
@@ -60,6 +68,8 @@ module Hyrax
       case destination_name
       when 'thumbnail'
         ".#{MIME::Types.type_for('jpg').first.extensions.first}"
+      when 'extracted_text'
+        ".#{MIME::Types.type_for('txt').first.extensions.first}"
       else
         ".#{destination_name}"
       end

data/app/services/hyrax/embargo_manager.rb

@@ -12,11 +12,13 @@ module Hyrax
   #       date is today or later.
   #    - "Applied" means the embargo's pre-release visibility has been set on
   #      the resource.
-  #    - "Released" means the embargo's post-release visibility has been set on
-  #      the resource.
   #    - "Enforced" means the object's visibility matches the pre-release
   #      visibility of the embargo; i.e. the embargo has been applied,
   #      but not released.
+  #    - "Released" means the embargo's post-release visibility has been set on
+  #      the resource.
+  #    - "Deactivate" means that the existing embargo will be removed, even
+  #      if it active.
   #
   # Note that an resource may be `#under_embargo?` even if the embargo is not
   # be `#enforced?` (in this case, the application should seek to apply the
@@ -74,7 +76,7 @@ module Hyrax
   #   resource.visibility # => 'open'
   #   manager.enforced? => false
   #
-  class EmbargoManager
+  class EmbargoManager # rubocop:disable Metrics/ClassLength
     ##
     # @!attribute [rw] resource
     #   @return [Hyrax::Resource]
@@ -108,15 +110,74 @@ module Hyrax
           .embargo
       end
 
+      # @return [Boolean]
+      def deactivate_embargo_for(resource:, query_service: Hyrax.query_service)
+        new(resource: resource, query_service: query_service)
+          .deactivate
+      end
+
+      # @return [Boolean]
+      def deactivate_embargo_for!(resource:, query_service: Hyrax.query_service)
+        new(resource: resource, query_service: query_service)
+          .deactivate!
+      end
+
+      # @return [Boolean]
       def release_embargo_for(resource:, query_service: Hyrax.query_service)
         new(resource: resource, query_service: query_service)
           .release
       end
 
+      # @return [Boolean]
       def release_embargo_for!(resource:, query_service: Hyrax.query_service)
         new(resource: resource, query_service: query_service)
           .release!
       end
+
+      # Creates or updates an existing embargo on a member to match the embargo on the parent work
+      # @param [Array<Valkyrie::Resource>] members
+      # @param [Hyrax::Work] work
+      # rubocop:disable Metrics/AbcSize, Metrics/MethodLength
+      def create_or_update_embargo_on_members(members, work)
+        # TODO: account for all members and levels, not just file sets. ref: #6131
+
+        members.each do |member|
+          member_embargo_needs_updating = work.embargo.updated_at > member.embargo&.updated_at if member.embargo
+
+          if member.embargo && member_embargo_needs_updating
+            member.embargo.embargo_release_date = work.embargo['embargo_release_date']
+            member.embargo.visibility_during_embargo = work.embargo['visibility_during_embargo']
+            member.embargo.visibility_after_embargo = work.embargo['visibility_after_embargo']
+            member.embargo = Hyrax.persister.save(resource: member.embargo)
+          else
+            work_embargo_manager = Hyrax::EmbargoManager.new(resource: work)
+            work_embargo_manager.copy_embargo_to(target: member)
+            member = Hyrax.persister.save(resource: member)
+          end
+
+          user ||= ::User.find_by_user_key(member.depositor)
+          # the line below works in that it indexes the file set with the necessary lease properties
+          # I do not know however if this is the best event_id to pass
+          Hyrax.publisher.publish('object.metadata.updated', object: member, user: user)
+        end
+      end
+      # rubocop:enable Metrics/AbcSize, Metrics/MethodLength
+    end
+
+    ##
+    # Deactivates the embargo
+    # @return [Boolean]
+    def deactivate
+      release(force: true) &&
+        nullify(force: true)
+    end
+
+    ##
+    # Deactivates the embargo
+    # @return [Boolean]
+    def deactivate!
+      release!(force: true)
+      nullify(force: true)
     end
 
     ##
@@ -128,7 +189,7 @@ module Hyrax
     def copy_embargo_to(target:)
       return false unless under_embargo?
 
-      target.embargo = Embargo.new(clone_attributes)
+      target.embargo = Hyrax.persister.save(resource: Embargo.new(clone_attributes))
       self.class.apply_embargo_for(resource: target)
     end
 
@@ -152,7 +213,8 @@ module Hyrax
     ##
     # @return [Boolean]
     def enforced?
-      embargo.visibility_during_embargo.to_s == resource.visibility
+      embargo.embargo_release_date.present? &&
+        (embargo.visibility_during_embargo.to_s == resource.visibility)
     end
 
     ##
@@ -162,21 +224,40 @@ module Hyrax
     end
 
     ##
-    # Drop the embargo by setting its release date to `nil`.
+    # Drop the embargo by setting its release date and visibility settings to `nil`.
     #
-    # @return [void]
-    def nullify
-      return unless under_embargo?
+    # @param force [Boolean] force the nullify even when the embargo period is current
+    #
+    # @return [Boolean]
+    def nullify(force: false)
+      return false if !force && under_embargo?
+
       embargo.embargo_release_date = nil
+      embargo.visibility_during_embargo = nil
+      embargo.visibility_after_embargo = nil
+      true
     end
 
     ##
-    # Sets the visibility of the resource to the embargo's visibility condition.
-    # no-op if the embargo period is current.
+    # Sets the visibility of the resource to the embargo's after embargo visibility.
+    # no-op if the embargo period is current and the force flag is false.
+    #
+    # @param force [boolean] force the release even when the embargo period is current
     #
     # @return [Boolean] truthy if the embargo has been applied
-    def release
-      return false if under_embargo?
+    def release(force: false)
+      return false if !force && under_embargo?
+
+      embargo_state = embargo.active? ? 'active' : 'expired'
+      history_record = embargo_history_message(
+        embargo_state,
+        Hyrax::TimeService.time_in_utc,
+        embargo.embargo_release_date,
+        embargo.visibility_during_embargo,
+        embargo.visibility_after_embargo
+      )
+      embargo.embargo_history += [history_record]
+
       return true if embargo.visibility_after_embargo.nil?
 
       resource.visibility = embargo.visibility_after_embargo
@@ -209,5 +290,17 @@ module Hyrax
     def core_attribute_keys
       [:visibility_after_embargo, :visibility_during_embargo, :embargo_release_date]
     end
+
+    protected
+
+    # Create the log message used when releasing an embargo
+    def embargo_history_message(state, deactivate_date, release_date, visibility_during, visibility_after)
+      I18n.t 'hydra.embargo.history_message',
+              state: state,
+              deactivate_date: deactivate_date,
+              release_date: release_date,
+              visibility_during: visibility_during,
+              visibility_after: visibility_after
+    end
   end
 end

data/app/services/hyrax/embargo_service.rb

@@ -1,24 +1,26 @@
 # frozen_string_literal: true
 module Hyrax
+  ##
+  # Methods for Querying Repository to find Embargoed Objects
   class EmbargoService < RestrictionService
     class << self
-      #
-      # Methods for Querying Repository to find Embargoed Objects
-      #
-
       # Returns all assets with embargo release date set to a date in the past
-      def assets_with_expired_embargoes
+      def assets_with_expired_enforced_embargoes
         builder = Hyrax::ExpiredEmbargoSearchBuilder.new(self)
         presenters(builder)
       end
+      alias assets_with_expired_embargoes assets_with_expired_enforced_embargoes
 
-      # Returns all assets with embargo release date set
-      #   (assumes that when lease visibility is applied to assets
-      #    whose leases have expired, the lease expiration date will be removed from its metadata)
-      def assets_under_embargo
+      ##
+      # Returns all assets with embargoes that are currently enforced,
+      # regardless of whether the embargoes are active or expired.
+      #
+      # @see Hyrax::EmbargoManager
+      def assets_with_enforced_embargoes
         builder = Hyrax::EmbargoSearchBuilder.new(self)
-        presenters(builder)
+        presenters(builder).select(&:enforced?)
       end
+      alias assets_under_embargo assets_with_enforced_embargoes
 
       # Returns all assets that have had embargoes deactivated in the past.
       def assets_with_deactivated_embargoes

data/app/services/hyrax/file_set_derivatives_service.rb

@@ -5,7 +5,7 @@ module Hyrax
     attr_reader :file_set
     delegate :mime_type, to: :file_set
 
-    # @param file_set [Hyrax::FileSet] At least for this class, it must have #uri and #mime_type
+    # @param file_set [Hyrax::FileSet, Hyrax::FileMetadata] At least for this class, it must have #uri and #mime_type
     def initialize(file_set)
       @file_set = file_set
     end
@@ -31,11 +31,11 @@ module Hyrax
 
     def create_derivatives(filename)
       case mime_type
-      when *file_set.class.pdf_mime_types             then create_pdf_derivatives(filename)
-      when *file_set.class.office_document_mime_types then create_office_document_derivatives(filename)
-      when *file_set.class.audio_mime_types           then create_audio_derivatives(filename)
-      when *file_set.class.video_mime_types           then create_video_derivatives(filename)
-      when *file_set.class.image_mime_types           then create_image_derivatives(filename)
+      when *Hyrax.config.derivative_mime_type_mappings[:pdf]    then create_pdf_derivatives(filename)
+      when *Hyrax.config.derivative_mime_type_mappings[:office] then create_office_document_derivatives(filename)
+      when *Hyrax.config.derivative_mime_type_mappings[:audio]  then create_audio_derivatives(filename)
+      when *Hyrax.config.derivative_mime_type_mappings[:video]  then create_video_derivatives(filename)
+      when *Hyrax.config.derivative_mime_type_mappings[:image] then create_image_derivatives(filename)
       end
     end
 
@@ -75,7 +75,7 @@ module Hyrax
                                                   url: derivative_url('thumbnail'),
                                                   layer: 0
                                                 }])
-      extract_full_text(filename, uri)
+      extract_full_text(filename, derivative_url('extracted_text'))
     end
 
     def create_office_document_derivatives(filename)
@@ -86,20 +86,20 @@ module Hyrax
                                                        url: derivative_url('thumbnail'),
                                                        layer: 0
                                                      }])
-      extract_full_text(filename, uri)
+      extract_full_text(filename, derivative_url('extracted_text'))
     end
 
     def create_audio_derivatives(filename)
       Hydra::Derivatives::AudioDerivatives.create(filename,
-                                                  outputs: [{ label: 'mp3', format: 'mp3', url: derivative_url('mp3') },
-                                                            { label: 'ogg', format: 'ogg', url: derivative_url('ogg') }])
+                                                  outputs: [{ label: 'mp3', format: 'mp3', url: derivative_url('mp3'), mime_type: 'audio/mpeg', container: 'service_file' },
+                                                            { label: 'ogg', format: 'ogg', url: derivative_url('ogg'), mime_type: 'audio/ogg', container: 'service_file' }])
     end
 
     def create_video_derivatives(filename)
       Hydra::Derivatives::VideoDerivatives.create(filename,
-                                                  outputs: [{ label: :thumbnail, format: 'jpg', url: derivative_url('thumbnail') },
-                                                            { label: 'webm', format: 'webm', url: derivative_url('webm') },
-                                                            { label: 'mp4', format: 'mp4', url: derivative_url('mp4') }])
+                                                  outputs: [{ label: :thumbnail, format: 'jpg', url: derivative_url('thumbnail'), mime_type: 'image/jpeg' },
+                                                            { label: 'webm', format: 'webm', url: derivative_url('webm'), mime_type: 'video/webm', container: 'service_file' },
+                                                            { label: 'mp4', format: 'mp4', url: derivative_url('mp4'), mime_type: 'video/mp4', container: 'service_file' }])
     end
 
     def create_image_derivatives(filename)
@@ -122,6 +122,7 @@ module Hyrax
     # @param [String] uri to the file set (deligated to file_set)
     def extract_full_text(filename, uri)
       return unless Hyrax.config.extract_full_text?
+
       Hydra::Derivatives::FullTextExtract.create(filename,
                                                  outputs: [{ url: uri, container: "extracted_text" }])
     end

data/app/services/hyrax/fixity_check_failure_service.rb

@@ -14,7 +14,7 @@ module Hyrax
     end
 
     def message
-      uri = file_set.original_file.uri.to_s
+      uri = checksum_audit_log.checked_uri
       file_title = file_set.title.first
       I18n.t('hyrax.notifications.fixity_check_failure.message', log_date: log_date, file_title: file_title, uri: uri)
     end

data/app/services/hyrax/identifier/dispatcher.rb

@@ -53,8 +53,15 @@ module Hyrax
       #
       # @see #assign_for
       def assign_for!(object:, attribute: :identifier)
-        assign_for(object: object, attribute: attribute).save!
-        object
+        result = assign_for(object: object, attribute: attribute)
+
+        case result
+        when Valkyrie::Resource
+          Hyrax.persister.save(resource: result)
+        else
+          result.save
+          result
+        end
       end
     end
   end