hyrax 4.0.0.rc3 → 5.0.0.rc1

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

@@ -1,20 +1,24 @@
 # frozen_string_literal: true
+
 require 'oauth2'
 require 'signet/oauth_2/client'
 
-# rubocop:disable Metrics/ModuleLength
 module Hyrax
   module Analytics
+    # rubocop:disable Metrics/ModuleLength
     module Google
       extend ActiveSupport::Concern
       # rubocop:disable Metrics/BlockLength
       class_methods do
-        # Loads configuration options from config/analytics.yml. Expected structure:
+        # Loads configuration options from config/analytics.yml. You only need PRIVATE_KEY_PATH or
+        # PRIVATE_KEY_VALUE. VALUE takes precedence.
+        # Expected structure:
         # `analytics:`
         # `  google:`
         # `    app_name: <%= ENV['GOOGLE_OAUTH_APP_NAME']`
         # `    app_version: <%= ENV['GOOGLE_OAUTH_APP_VERSION']`
         # `    privkey_path: <%= ENV['GOOGLE_OAUTH_PRIVATE_KEY_PATH']`
+        # `    privkey_value: <%= ENV['GOOGLE_OAUTH_PRIVATE_KEY_VALUE']`
         # `    privkey_secret: <%= ENV['GOOGLE_OAUTH_PRIVATE_KEY_SECRET']`
         # `    client_email: <%= ENV['GOOGLE_OAUTH_CLIENT_EMAIL']`
         # @return [Config]
@@ -41,7 +45,8 @@ module Hyrax
             new config
           end
 
-          REQUIRED_KEYS = %w[analytics_id app_name app_version privkey_path privkey_secret client_email].freeze
+          KEYS = %w[analytics_id app_name app_version privkey_path privkey_value privkey_secret client_email].freeze
+          REQUIRED_KEYS = %w[analytics_id app_name app_version privkey_secret client_email].freeze
 
           def initialize(config)
             @config = config
@@ -49,18 +54,16 @@ module Hyrax
 
           # @return [Boolean] are all the required values present?
           def valid?
-            config_keys = @config.keys
-            REQUIRED_KEYS.all? { |required| config_keys.include?(required) }
-          end
+            return false unless @config['privkey_value'].present? || @config['privkey_path'].present?
 
-          REQUIRED_KEYS.each do |key|
-            class_eval %{ def #{key};  @config.fetch('#{key}'); end }
+            REQUIRED_KEYS.all? { |required| @config[required].present? }
           end
 
-          # This method allows setting the analytics id in the initializer
-          # @deprecated set the analytics id in either ENV['GOOGLE_ANALYTICS_ID'] or config/analytics.yaml
-          def analytics_id=(value)
-            @config['analytics_id'] = value
+          KEYS.each do |key|
+            # rubocop:disable Style/EvalWithLocation
+            class_eval %{ def #{key}; @config.fetch('#{key}'); end }
+            class_eval %{ def #{key}=(value); @config['#{key}'] = value; end }
+            # rubocop:enable Style/EvalWithLocation
           end
         end
 
@@ -77,8 +80,12 @@ module Hyrax
         end
 
         def auth_client(scope)
-          raise "Private key file for Google analytics was expected at '#{config.privkey_path}', but no file was found." unless File.exist?(config.privkey_path)
-          private_key = File.read(config.privkey_path)
+          private_key = Base64.decode64(config.privkey_value) if config.privkey_value.present?
+          if private_key.blank?
+            raise "Private key file for Google analytics was expected at '#{config.privkey_path}', but no file was found." unless File.exist?(config.privkey_path)
+
+            private_key = File.read(config.privkey_path)
+          end
           Signet::OAuth2::Client.new token_credential_uri: 'https://accounts.google.com/o/oauth2/token',
                                      audience: 'https://accounts.google.com/o/oauth2/token',
                                      scope: scope,
@@ -123,7 +130,7 @@ module Hyrax
 
           [start_date, end_date]
         end
-        # rubocop:enabl e Metrics/MethodLength
+        # rubocop:enable Metrics/MethodLength
 
         def keyword_conversion(date)
           case date
@@ -199,6 +206,7 @@ module Hyrax
       end
       # rubocop:enable Metrics/BlockLength
     end
+    # rubocop:enable Metrics/ModuleLength
   end
 end
 # rubocop:enable Metrics/ModuleLength

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

@@ -11,15 +11,16 @@ module Hyrax
         # Loads configuration options from config/analytics.yml. Expected structure:
         # `analytics:`
         # `  matomo:`
-        # `    base_url: <%= ENV['MATOMOT_BASE_URL']`
-        # `    site_id: <%= ENV['MATOMOT_SITE_ID']`
-        # `    auth_token: <%= ENV['MATOMOT_AUTH_TOKEN']`
+        # `    base_url: <%= ENV['MATOMO_BASE_URL']`
+        # `    site_id: <%= ENV['MATOMO_SITE_ID']`
+        # `    auth_token: <%= ENV['MATOMO_AUTH_TOKEN']`
         # @return [Config]
         def config
           @config ||= Config.load_from_yaml
         end
 
         class Config
+          # TODO: test matomo and see if it needs any of the updates from https://github.com/samvera/hyrax/pull/6063
           def self.load_from_yaml
             filename = Rails.root.join('config', 'analytics.yml')
             yaml = YAML.safe_load(ERB.new(File.read(filename)).result)

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

@@ -27,25 +27,27 @@ 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/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

@@ -17,6 +17,7 @@ module Hyrax
   #    - "Enforced" means the object's visibility matches the pre-release
   #      visibility of the embargo; i.e. the embargo has been applied,
   #      but not released.
+  #    - "Deactivate" means that the existing embargo will be removed
   #
   # 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 +75,7 @@ module Hyrax
   #   resource.visibility # => 'open'
   #   manager.enforced? => false
   #
-  class EmbargoManager
+  class EmbargoManager # rubocop:disable Metrics/ClassLength
     ##
     # @!attribute [rw] resource
     #   @return [Hyrax::Resource]
@@ -117,6 +118,51 @@ module Hyrax
         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 and logs a message to the embargo_history property
+    def deactivate!
+      embargo_state = embargo.active? ? 'active' : 'expired'
+      embargo_record = embargo_history_message(
+        embargo_state,
+        Time.zone.today,
+        embargo.embargo_release_date,
+        embargo.visibility_during_embargo,
+        embargo.visibility_after_embargo
+      )
+
+      release(force: true)
+      nullify(force: true)
+      embargo.embargo_history += [embargo_record]
     end
 
     ##
@@ -128,7 +174,7 @@ 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 +198,8 @@ 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 +209,28 @@ 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`.
+    #
+    # @param force [boolean] force the nullify even when the embargo period is current
     #
     # @return [void]
-    def nullify
-      return unless under_embargo?
+    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
     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?
       return true if embargo.visibility_after_embargo.nil?
 
       resource.visibility = embargo.visibility_after_embargo
@@ -209,5 +263,17 @@ module Hyrax
     def core_attribute_keys
       [:visibility_after_embargo, :visibility_during_embargo, :embargo_release_date]
     end
+
+    protected
+
+    # Create the log message used when deactivating 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/file_set_derivatives_service.rb

@@ -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)
@@ -86,7 +86,7 @@ module Hyrax
                                                        url: derivative_url('thumbnail'),
                                                        layer: 0
                                                      }])
-      extract_full_text(filename, uri)
+      extract_full_text(filename, derivative_url('extracted_text'))
     end
 
     def create_audio_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