bulkrax 4.4.0 → 5.0.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: d53012e0252f8033f5da334d5336c134deb3a6221040cfe4c1c6c01bb473d617
-  data.tar.gz: e986a3506c073aa533c4ea303a123e8866d2bf7ac6d8ab6df3949682f61b6c05
+  metadata.gz: 49a1f0ccc806cf73a7872a634c8c2819dac62f98f0a9dc163e7ada8931d9b1fe
+  data.tar.gz: ecf5c8ad3e4864110665cfb5c6c5f6329e3b868414815bba163646db839400e7
 SHA512:
-  metadata.gz: 8ea20593d1164f62fdfda2dbe84bd8cc4e199a65d7e85eea0b68b27e71aeff7579f1b0059042f55eec1c7ab36f8dec9db406ad58cc3a315b29b8e4900f7cb450
-  data.tar.gz: 89a03cd842d855d48f8c4952b6e9a6daea757e80cb8c0f9ee370b59f90e434b39dfb548deb1757f36463914bd5cccbcf1df6c5ffd9102bb5f0d54b5b7ea6476f
+  metadata.gz: df394e4fbbc6ca0a71eb595c075c5181d24098b1ab889ff185da10991ec01c2eaf60239658564838b67df4c0a421306de923d2eab4bbeffafb31321acb9ba1ff
+  data.tar.gz: 35b691e96d0e59f83efadc35364283239f7f741ab4bf9d859756dacc8483e380090512cd10bf3b6ec4ce81936d00c88eb4d1ed96742c384e0f3082028ffdcf45

data/app/controllers/bulkrax/entries_controller.rb

@@ -7,6 +7,7 @@ module Bulkrax
   class EntriesController < ApplicationController
     include Hyrax::ThemedLayoutController
     before_action :authenticate_user!
+    before_action :check_permissions
     with_themed_layout 'dashboard'
 
     def show
@@ -40,5 +41,9 @@ module Bulkrax
       add_breadcrumb @exporter.name, bulkrax.exporter_path(@exporter.id)
       add_breadcrumb @entry.id
     end
+
+    def check_permissions
+      raise CanCan::AccessDenied unless current_ability.can_import_works? || current_ability.can_export_works?
+    end
   end
 end

data/app/controllers/bulkrax/exporters_controller.rb

@@ -7,6 +7,7 @@ module Bulkrax
     include Hyrax::ThemedLayoutController
     include Bulkrax::DownloadBehavior
     before_action :authenticate_user!
+    before_action :check_permissions
     before_action :set_exporter, only: [:show, :edit, :update, :destroy]
     with_themed_layout 'dashboard'
 
@@ -131,5 +132,9 @@ module Bulkrax
     def file_path
       "#{@exporter.exporter_export_zip_path}/#{params['exporter']['exporter_export_zip_files']}"
     end
+
+    def check_permissions
+      raise CanCan::AccessDenied unless current_ability.can_export_works?
+    end
   end
 end

data/app/controllers/bulkrax/importers_controller.rb

@@ -14,6 +14,7 @@ module Bulkrax
     protect_from_forgery unless: -> { api_request? }
     before_action :token_authenticate!, if: -> { api_request? }, only: [:create, :update, :delete]
     before_action :authenticate_user!, unless: -> { api_request? }
+    before_action :check_permissions
     before_action :set_importer, only: [:show, :edit, :update, :destroy]
     with_themed_layout 'dashboard'
 
@@ -76,6 +77,9 @@ module Bulkrax
       @importer = Importer.new(importer_params)
       field_mapping_params
       @importer.validate_only = true if params[:commit] == 'Create and Validate'
+      # the following line is needed to handle updating remote files of a FileSet
+      # on a new import otherwise it only gets updated during the update path
+      @importer.parser_fields['update_files'] = true if params[:commit] == 'Create and Import'
       if @importer.save
         files_for_import(file, cloud_files)
         if params[:commit] == 'Create and Import'
@@ -316,6 +320,10 @@ module Bulkrax
       end
       @importer.save
     end
+
+    def check_permissions
+      raise CanCan::AccessDenied unless current_ability.can_import_works?
+    end
   end
   # rubocop:enable Metrics/ClassLength
 end

data/app/factories/bulkrax/object_factory.rb

@@ -6,6 +6,27 @@ module Bulkrax
     include Bulkrax::FileFactory
     include DynamicRecordLookup
 
+    # @api private
+    #
+    # These are the attributes that we assume all "work type" classes (e.g. the given :klass) will
+    # have in addition to their specific attributes.
+    #
+    # @return [Array<Symbol>]
+    # @see #permitted_attributes
+    class_attribute :base_permitted_attributes,
+      default: %i[id edit_users edit_groups read_groups visibility work_members_attributes admin_set_id]
+
+    # @return [Boolean]
+    #
+    # @example
+    #   Bulkrax::ObjectFactory.transformation_removes_blank_hash_values = true
+    #
+    # @see #transform_attributes
+    # @see https://github.com/samvera-labs/bulkrax/pull/708 For discussion concerning this feature
+    # @see https://github.com/samvera-labs/bulkrax/wiki/Interacting-with-Metadata For documentation
+    #      concerning default behavior.
+    class_attribute :transformation_removes_blank_hash_values, default: false
+
     define_model_callbacks :save, :create
     attr_reader :attributes, :object, :source_identifier_value, :klass, :replace_files, :update_files, :work_identifier, :related_parents_parsed_mapping, :importer_run_id
 
@@ -58,7 +79,7 @@ module Bulkrax
         elsif klass == FileSet
           update_file_set(attrs)
         else
-          work_actor.update(environment(attrs))
+          update_work(attrs)
         end
       end
       object.apply_depositor_metadata(@user) && object.save! if object.depositor.nil?
@@ -104,7 +125,7 @@ module Bulkrax
           elsif klass == FileSet
             create_file_set(attrs)
           else
-            work_actor.create(environment(attrs))
+            create_work(attrs)
           end
         end
       end
@@ -139,6 +160,14 @@ module Bulkrax
       Hyrax::CurationConcern.actor
     end
 
+    def create_work(attrs)
+      work_actor.create(environment(attrs))
+    end
+
+    def update_work(attrs)
+      work_actor.update(environment(attrs))
+    end
+
     def create_collection(attrs)
       attrs = clean_attrs(attrs)
       attrs = collection_type(attrs)
@@ -227,12 +256,32 @@ module Bulkrax
     def transform_attributes(update: false)
       @transform_attributes = attributes.slice(*permitted_attributes)
       @transform_attributes.merge!(file_attributes(update_files)) if with_files
+      @transform_attributes = remove_blank_hash_values(@transform_attributes) if transformation_removes_blank_hash_values?
       update ? @transform_attributes.except(:id) : @transform_attributes
     end
 
     # Regardless of what the Parser gives us, these are the properties we are prepared to accept.
     def permitted_attributes
-      klass.properties.keys.map(&:to_sym) + %i[id edit_users edit_groups read_groups visibility work_members_attributes admin_set_id]
+      klass.properties.keys.map(&:to_sym) + base_permitted_attributes
+    end
+
+    # Return a copy of the given attributes, such that all values that are empty or an array of all
+    # empty values are fully emptied.  (See implementation details)
+    #
+    # @param attributes [Hash]
+    # @return [Hash]
+    #
+    # @see https://github.com/emory-libraries/dlp-curate/issues/1973
+    def remove_blank_hash_values(attributes)
+      dupe = attributes.dup
+      dupe.each do |key, values|
+        if values.is_a?(Array) && values.all? { |value| value.is_a?(String) && value.empty? }
+          dupe[key] = []
+        elsif values.is_a?(String) && values.empty?
+          dupe[key] = nil
+        end
+      end
+      dupe
     end
   end
 end

data/app/jobs/bulkrax/import_file_set_job.rb

@@ -2,6 +2,7 @@
 
 module Bulkrax
   class MissingParentError < ::StandardError; end
+
   class ImportFileSetJob < ApplicationJob
     include DynamicRecordLookup
 

data/app/jobs/bulkrax/import_work_job.rb

@@ -5,7 +5,7 @@ module Bulkrax
     queue_as :import
 
     # rubocop:disable Rails/SkipsModelValidations
-    def perform(entry_id, run_id, *)
+    def perform(entry_id, run_id, time_to_live = 3, *)
       entry = Entry.find(entry_id)
       importer_run = ImporterRun.find(run_id)
       entry.build
@@ -24,13 +24,21 @@ module Bulkrax
       entry.save!
       entry.importer.current_run = importer_run
       entry.importer.record_status
-    rescue Bulkrax::CollectionsCreatedError
-      reschedule(entry_id, run_id)
+    rescue Bulkrax::CollectionsCreatedError => e
+      Rails.logger.warn("#{self.class} entry_id: #{entry_id}, run_id: #{run_id} encountered #{e.class}: #{e.message}")
+      # You get 3 attempts at the above perform before we have the import exception cascade into
+      # the Sidekiq retry ecosystem.
+      # rubocop:disable Style/IfUnlessModifier
+      if time_to_live <= 1
+        raise "Exhauted reschedule limit for #{self.class} entry_id: #{entry_id}, run_id: #{run_id}.  Attemping retries"
+      end
+      # rubocop:enable Style/IfUnlessModifier
+      reschedule(entry_id, run_id, time_to_live)
     end
     # rubocop:enable Rails/SkipsModelValidations
 
-    def reschedule(entry_id, run_id)
-      ImportWorkJob.set(wait: 1.minute).perform_later(entry_id, run_id)
+    def reschedule(entry_id, run_id, time_to_live)
+      ImportWorkJob.set(wait: 1.minute).perform_later(entry_id, run_id, time_to_live - 1)
     end
   end
 end

data/app/matchers/bulkrax/application_matcher.rb

@@ -6,6 +6,10 @@ module Bulkrax
   class ApplicationMatcher
     attr_accessor :to, :from, :parsed, :if, :split, :excluded, :nested_type
 
+    # New parse methods will need to be added here; you'll also want to define a corresponding
+    # "parse_#{field}" method.
+    class_attribute :parsed_fields, instance_writer: false, default: ['remote_files', 'language', 'subject', 'types', 'model', 'resource_type', 'format_original']
+
     def initialize(args)
       args.each do |k, v|
         send("#{k}=", v)
@@ -38,8 +42,6 @@ module Bulkrax
     end
 
     def process_parse
-      # New parse methods will need to be added here
-      parsed_fields = ['remote_files', 'language', 'subject', 'types', 'model', 'resource_type', 'format_original']
       # This accounts for prefixed matchers
       parser = parsed_fields.find { |field| to&.include? field }
 

data/app/models/bulkrax/csv_entry.rb

@@ -7,7 +7,7 @@ module Bulkrax
   # We do too much in these entry classes. We need to extract the common logic from the various
   # entry models into a module that can be shared between them.
   class CsvEntry < Entry # rubocop:disable Metrics/ClassLength
-    serialize :raw_metadata, JSON
+    serialize :raw_metadata, Bulkrax::NormalizedJson
 
     def self.fields_from_data(data)
       data.headers.flatten.compact.uniq
@@ -36,10 +36,14 @@ module Bulkrax
 
     def build_metadata
       raise StandardError, 'Record not found' if record.nil?
-      raise StandardError, "Missing required elements, missing element(s) are: #{importerexporter.parser.missing_elements(keys_without_numbers(record.keys)).join(', ')}" unless importerexporter.parser.required_elements?(keys_without_numbers(record.keys))
+      unless importerexporter.parser.required_elements?(keys_without_numbers(record.keys))
+        raise StandardError,
+"Missing required elements, missing element(s) are: #{importerexporter.parser.missing_elements(keys_without_numbers(record.keys)).join(', ')}"
+      end
 
       self.parsed_metadata = {}
       add_identifier
+      establish_factory_class
       add_ingested_metadata
       # TODO(alishaevn): remove the collections stuff entirely and only reference collections via the new parents code
       add_collections
@@ -56,6 +60,12 @@ module Bulkrax
       self.parsed_metadata[work_identifier] = [record[source_identifier]]
     end
 
+    def establish_factory_class
+      parser.model_field_mappings.each do |key|
+        add_metadata('model', record[key]) if record.key?(key)
+      end
+    end
+
     def add_metadata_for_model
       if factory_class == Collection
         add_collection_type_gid
@@ -107,7 +117,9 @@ module Bulkrax
     # Metadata required by Bulkrax for round-tripping
     def build_system_metadata
       self.parsed_metadata['id'] = hyrax_record.id
-      self.parsed_metadata[source_identifier] = hyrax_record.send(work_identifier)
+      source_id = hyrax_record.send(work_identifier)
+      source_id = source_id.to_a.first if source_id.is_a?(ActiveTriples::Relation)
+      self.parsed_metadata[source_identifier] = source_id
       self.parsed_metadata[key_for_export('model')] = hyrax_record.has_model.first
     end
 

data/app/models/bulkrax/entry.rb

@@ -4,6 +4,7 @@ module Bulkrax
   # Custom error class for collections_created?
   class CollectionsCreatedError < RuntimeError; end
   class OAIError < RuntimeError; end
+
   class Entry < ApplicationRecord
     include Bulkrax::HasMatchers
     include Bulkrax::ImportBehavior
@@ -15,7 +16,7 @@ module Bulkrax
     alias importer importerexporter
     alias exporter importerexporter
 
-    serialize :parsed_metadata, JSON
+    serialize :parsed_metadata, Bulkrax::NormalizedJson
     # Do not serialize raw_metadata as so we can support xml or other formats
     serialize :collection_ids, Array
 

data/app/models/bulkrax/oai_entry.rb

@@ -5,7 +5,7 @@ require 'ostruct'
 
 module Bulkrax
   class OaiEntry < Entry
-    serialize :raw_metadata, JSON
+    serialize :raw_metadata, Bulkrax::NormalizedJson
 
     delegate :record, to: :raw_record
 
@@ -28,13 +28,16 @@ module Bulkrax
     def build_metadata
       self.parsed_metadata = {}
       self.parsed_metadata[work_identifier] = [record.header.identifier]
+      self.raw_metadata = { xml: record.metadata.to_s }
 
-      record.metadata.children.each do |child|
-        child.children.each do |node|
-          add_metadata(node.name, node.content)
-        end
-      end
-      add_metadata('thumbnail_url', thumbnail_url)
+      # We need to establish the #factory_class before we proceed with the metadata.  See
+      # https://github.com/samvera-labs/bulkrax/issues/702 for further details.
+      #
+      # tl;dr - if we don't have the right factory_class we might skip properties that are
+      # specifically assigned to the factory class
+      establish_factory_class
+      add_metadata_from_record
+      add_thumbnail_url
 
       add_visibility
       add_rights_statement
@@ -53,6 +56,37 @@ module Bulkrax
       end
     end
 
+    # To ensure we capture the correct parse data, we first need to establish the factory_class.
+    # @see https://github.com/samvera-labs/bulkrax/issues/702
+    def establish_factory_class
+      model_field_names = parser.model_field_mappings
+
+      each_candidate_metadata_node do |node|
+        next unless model_field_names.include?(node.name)
+        add_metadata(node.name, node.content)
+      end
+    end
+
+    def add_metadata_from_record
+      each_candidate_metadata_node do |node|
+        add_metadata(node.name, node.content)
+      end
+    end
+
+    # A method that you could override to better handle the shape of the record's metadata.
+    # @yieldparam node [Object<#name, #content>]
+    def each_candidate_metadata_node
+      record.metadata.children.each do |child|
+        child.children.each do |node|
+          yield(node)
+        end
+      end
+    end
+
+    def add_thumbnail_url
+      add_metadata('thumbnail_url', thumbnail_url)
+    end
+
     # Retrieve list of collections for the entry; add to collection_ids
     # If OAI-PMH doesn't return setSpec in the headers for GetRecord, use parser.collection_name
     #   in this case, if 'All' is selected, records will not be added to a collection.

data/app/models/bulkrax/rdf_entry.rb

@@ -3,7 +3,7 @@
 require 'rdf'
 module Bulkrax
   class RdfEntry < Entry
-    serialize :raw_metadata, JSON
+    serialize :raw_metadata, Bulkrax::NormalizedJson
 
     def self.read_data(path)
       RDF::Reader.open(path)

data/app/models/bulkrax/xml_entry.rb

@@ -4,7 +4,7 @@ require 'nokogiri'
 module Bulkrax
   # Generic XML Entry
   class XmlEntry < Entry
-    serialize :raw_metadata, JSON
+    serialize :raw_metadata, Bulkrax::NormalizedJson
 
     def self.fields_from_data(data); end
 
@@ -43,15 +43,14 @@ module Bulkrax
       raise StandardError, "Missing source identifier (#{source_identifier})" if self.raw_metadata[source_identifier].blank?
       self.parsed_metadata = {}
       self.parsed_metadata[work_identifier] = [self.raw_metadata[source_identifier]]
-      xml_elements.each do |element_name|
-        elements = record.xpath("//*[name()='#{element_name}']")
-        next if elements.blank?
-        elements.each do |el|
-          el.children.map(&:content).each do |content|
-            add_metadata(element_name, content) if content.present?
-          end
-        end
-      end
+
+      # We need to establish the #factory_class before we proceed with the metadata.  See
+      # https://github.com/samvera-labs/bulkrax/issues/702 for further details.
+      #
+      # tl;dr - if we don't have the right factory_class we might skip properties that are
+      # specifically assigned to the factory class
+      establish_factory_class
+      add_metadata_from_record
       add_visibility
       add_rights_statement
       add_admin_set_id
@@ -63,11 +62,54 @@ module Bulkrax
       self.parsed_metadata
     end
 
-    # Grab the class from the real parser
-    def xml_elements
+    def establish_factory_class
+      model_field_names = parser.model_field_mappings
+
+      each_candidate_metadata_node_name_and_content(elements: parser.model_field_mappings) do |name, content|
+        next unless model_field_names.include?(name)
+        add_metadata(name, content)
+      end
+    end
+
+    def add_metadata_from_record
+      each_candidate_metadata_node_name_and_content do |name, content|
+        add_metadata(name, content)
+      end
+    end
+
+    def each_candidate_metadata_node_name_and_content(elements: field_mapping_from_values_for_xml_element_names)
+      elements.each do |name|
+        # NOTE: the XML element name's case matters
+        nodes = record.xpath("//*[name()='#{name}']")
+        next if nodes.empty?
+
+        nodes.each do |node|
+          node.children.each do |content|
+            next if content.to_s.blank?
+
+            yield(name, content.to_s)
+          end
+        end
+      end
+    end
+
+    # Returns the explicitly declared "from" key's value of each parser's element's value.  (Yes, I
+    # would like a proper class for the thing I just tried to describe.)
+    #
+    # @return [Array<String>]
+    #
+    # @todo Additionally, we may want to revisit the XML parser fundamental logic; namely we only
+    #       parse nodes that are explicitly declared with in the `from`.  This is a bit different
+    #       than other parsers, in that they will make assumptions about each encountered column (in
+    #       the case of CSV) or node (in the case of OAI).  tl;dr - Here there be dragons.
+    def field_mapping_from_values_for_xml_element_names
       Bulkrax.field_mappings[self.importerexporter.parser_klass].map do |_k, v|
         v[:from]
       end.flatten.compact.uniq
     end
+
+    # Included for potential downstream adopters
+    alias xml_elements field_mapping_from_values_for_xml_element_names
+    deprecation_deprecate xml_elements: "Use '#{self}#field_mapping_from_values_for_xml_element_names' instead"
   end
 end

data/app/models/concerns/bulkrax/file_factory.rb

@@ -45,9 +45,15 @@ module Bulkrax
     end
 
     def new_remote_files
-      return if object.is_a? FileSet
-
-      @new_remote_files ||= if object.present? && object.file_sets.present?
+      @new_remote_files ||= if object.is_a? FileSet
+                              parsed_remote_files.select do |file|
+                                # is the url valid?
+                                is_valid = file[:url]&.match(URI::ABS_URI)
+                                # does the file already exist
+                                is_existing = object.import_url && object.import_url == file[:url]
+                                is_valid && !is_existing
+                              end
+                            elsif object.present? && object.file_sets.present?
                               parsed_remote_files.select do |file|
                                 # is the url valid?
                                 is_valid = file[:url]&.match(URI::ABS_URI)

data/app/models/concerns/bulkrax/import_behavior.rb

@@ -117,23 +117,30 @@ module Bulkrax
       Bulkrax.qa_controlled_properties.each do |field|
         next if parsed_metadata[field].blank?
 
-        parsed_metadata[field].each_with_index do |value, i|
-          next if value.blank?
-
-          if (validated_uri_value = validate_value(value, field))
-            parsed_metadata[field][i] = validated_uri_value
-          else
-            debug_msg = %(Unable to locate active authority ID "#{value}" in config/authorities/#{field.pluralize}.yml)
-            Rails.logger.debug(debug_msg)
-            error_msg = %("#{value}" is not a valid and/or active authority ID for the :#{field} field)
-            raise ::StandardError, error_msg
+        if multiple?(field)
+          parsed_metadata[field].each_with_index do |value, i|
+            next if value.blank?
+            parsed_metadata[field][i] = sanitize_controlled_uri_value(field, value)
           end
+        else
+          parsed_metadata[field] = sanitize_controlled_uri_value(field, parsed_metadata[field])
         end
       end
 
       true
     end
 
+    def sanitize_controlled_uri_value(field, value)
+      if (validated_uri_value = validate_value(value, field))
+        validated_uri_value
+      else
+        debug_msg = %(Unable to locate active authority ID "#{value}" in config/authorities/#{field.pluralize}.yml)
+        Rails.logger.debug(debug_msg)
+        error_msg = %("#{value}" is not a valid and/or active authority ID for the :#{field} field)
+        raise ::StandardError, error_msg
+      end
+    end
+
     # @param value [String] value to validate
     # @param field [String] name of the controlled property
     # @return [String, nil] validated URI value or nil

data/app/parsers/bulkrax/application_parser.rb

@@ -51,6 +51,12 @@ module Bulkrax
       raise NotImplementedError, 'must be defined'
     end
 
+    # @api public
+    # @abstract Subclass and override {#file_set_entry_class} to implement behavior for the parser.
+    def file_set_entry_class
+      raise NotImplementedError, 'must be defined'
+    end
+
     # @api public
     # @abstract Subclass and override {#records} to implement behavior for the parser.
     def records(_opts = {})
@@ -153,7 +159,7 @@ module Bulkrax
     # @see #create_relationships
     def create_objects(types = [])
       types.each do |object_type|
-        parser.send("create_#{object_type.pluralize}")
+        send("create_#{object_type.pluralize}")
       end
     end
 

data/app/parsers/bulkrax/csv_parser.rb

@@ -196,10 +196,13 @@ module Bulkrax
         @collection_ids = ActiveFedora::SolrService.query("has_model_ssim:Collection #{extra_filters}", method: :post, rows: 2_147_483_647).map(&:id)
         @file_set_ids = ActiveFedora::SolrService.query("has_model_ssim:FileSet #{extra_filters}", method: :post, rows: 2_147_483_647).map(&:id)
       when 'collection'
-        @work_ids = ActiveFedora::SolrService.query("member_of_collection_ids_ssim:#{importerexporter.export_source + extra_filters} AND has_model_ssim:(#{Hyrax.config.curation_concerns.join(' OR ')})", method: :post, rows: 2_000_000_000).map(&:id)
+        @work_ids = ActiveFedora::SolrService.query(
+"member_of_collection_ids_ssim:#{importerexporter.export_source + extra_filters} AND has_model_ssim:(#{Hyrax.config.curation_concerns.join(' OR ')})", method: :post, rows: 2_000_000_000
+).map(&:id)
         # get the parent collection and child collections
         @collection_ids = ActiveFedora::SolrService.query("id:#{importerexporter.export_source} #{extra_filters}", method: :post, rows: 2_147_483_647).map(&:id)
-        @collection_ids += ActiveFedora::SolrService.query("has_model_ssim:Collection AND member_of_collection_ids_ssim:#{importerexporter.export_source}", method: :post, rows: 2_147_483_647).map(&:id)
+        @collection_ids += ActiveFedora::SolrService.query("has_model_ssim:Collection AND member_of_collection_ids_ssim:#{importerexporter.export_source}", method: :post,
+                                                                                                                                                            rows: 2_147_483_647).map(&:id)
         find_child_file_sets(@work_ids)
       when 'worktype'
         @work_ids = ActiveFedora::SolrService.query("has_model_ssim:#{importerexporter.export_source + extra_filters}", method: :post, rows: 2_000_000_000).map(&:id)
@@ -234,7 +237,7 @@ module Bulkrax
         instance_variable_set(instance_var, ActiveFedora::SolrService.post(
           extra_filters.to_s,
           fq: [
-            %(#{::Solrizer.solr_name(work_identifier)}:("#{complete_entry_identifiers.join('" OR "')}")),
+            %(#{solr_name(work_identifier)}:("#{complete_entry_identifiers.join('" OR "')}")),
             "has_model_ssim:(#{models_to_search.join(' OR ')})"
           ],
           fl: 'id',
@@ -243,6 +246,10 @@ module Bulkrax
       end
     end
 
+    def solr_name(base_name)
+      Module.const_defined?(:Solrizer) ? ::Solrizer.solr_name(base_name) : ::ActiveFedora.index_field_mapper.solr_name(base_name)
+    end
+
     def create_new_entries
       current_record_ids.each_with_index do |id, index|
         break if limit_reached?(limit, index)

data/app/parsers/bulkrax/xml_parser.rb

@@ -12,6 +12,12 @@ module Bulkrax
     # @todo not yet supported
     def create_collections; end
 
+    # @todo not yet supported
+    def file_set_entry_class; end
+
+    # @todo not yet supported
+    def create_file_sets; end
+
     # TODO: change to differentiate between collection and work records when adding ability to import collection metadata
     def works_total
       total

data/app/views/hyrax/dashboard/sidebar/_bulkrax_sidebar_additions.html.erb

@@ -1,7 +1,9 @@
-<% if ENV.fetch('HYKU_BULKRAX_ENABLED', 'true') == 'true' %>
+<% if current_ability.can_import_works? %>
   <%= menu.nav_link(bulkrax.importers_path) do %>
     <span class="fa fa-cloud-upload" aria-hidden="true"></span> <span class="sidebar-action-text"><%= t('bulkrax.admin.sidebar.importers') %></span>
   <% end %>
+<% end %>
+<% if current_ability.can_export_works? %>
   <%= menu.nav_link(bulkrax.exporters_path) do %>
     <span class="fa fa-cloud-download" aria-hidden="true"></span> <span class="sidebar-action-text"><%= t('bulkrax.admin.sidebar.exporters') %></span>
   <% end %>

data/app/views/hyrax/dashboard/sidebar/_repository_content.html.erb

@@ -1,27 +1,30 @@
-  <li class="h5 nav-item"><%= t('hyrax.admin.sidebar.repository_objects') %></li>
+<li class="h5 nav-item"><%= t('hyrax.admin.sidebar.repository_objects') %></li>
 
-  <%= menu.nav_link(hyrax.my_collections_path,
-                    class: "nav-link",
-                    onclick: "dontChangeAccordion(event);",
-                    also_active_for: hyrax.dashboard_collections_path) do %>
-    <span class="fa fa-folder-open" aria-hidden="true"></span> <span class="sidebar-action-text"><%= t('hyrax.admin.sidebar.collections') %></span>
-  <% end %>
+<%= menu.nav_link(hyrax.my_collections_path,
+                  class: "nav-link",
+                  onclick: "dontChangeAccordion(event);",
+                  also_active_for: hyrax.dashboard_collections_path) do %>
+  <span class="fa fa-folder-open" aria-hidden="true"></span> <span class="sidebar-action-text"><%= t('hyrax.admin.sidebar.collections') %></span>
+<% end %>
 
-  <%= menu.nav_link(hyrax.my_works_path,
-                    class: "nav-link",
-                    onclick: "dontChangeAccordion(event);",
-                    also_active_for: hyrax.dashboard_works_path) do %>
-    <span class="fa fa-file" aria-hidden="true"></span> <span class="sidebar-action-text"><%= t('hyrax.admin.sidebar.works') %></span>
-  <% end %>
+<%= menu.nav_link(hyrax.my_works_path,
+                  class: "nav-link",
+                  onclick: "dontChangeAccordion(event);",
+                  also_active_for: hyrax.dashboard_works_path) do %>
+  <span class="fa fa-file" aria-hidden="true"></span> <span class="sidebar-action-text"><%= t('hyrax.admin.sidebar.works') %></span>
+<% end %>
 
-  <% if ::Hyrax::DashboardController&.respond_to?(:sidebar_partials) %>
+<% if ::Hyrax::DashboardController&.respond_to?(:sidebar_partials) %>
   <%= render 'hyrax/dashboard/sidebar/menu_partials', menu: menu, section: :repository_content %>
-  <% else %>
-  <%= menu.nav_link(bulkrax.importers_path) do %>
-    <span class="fa fa-cloud-upload" aria-hidden="true"></span> <span class="sidebar-action-text"><%= t('bulkrax.admin.sidebar.importers') %></span>
-  <% end %>
-
-  <%= menu.nav_link(bulkrax.exporters_path) do %>
-    <span class="fa fa-cloud-download" aria-hidden="true"></span> <span class="sidebar-action-text"><%= t('bulkrax.admin.sidebar.exporters') %></span>
+<% else %>
+  <% if current_ability.can_import_works? %>
+    <%= menu.nav_link(bulkrax.importers_path) do %>
+      <span class="fa fa-cloud-upload" aria-hidden="true"></span> <span class="sidebar-action-text"><%= t('bulkrax.admin.sidebar.importers') %></span>
+    <% end %>
   <% end %>
+  <% if current_ability.can_export_works? %>
+    <%= menu.nav_link(bulkrax.exporters_path) do %>
+      <span class="fa fa-cloud-download" aria-hidden="true"></span> <span class="sidebar-action-text"><%= t('bulkrax.admin.sidebar.exporters') %></span>
+    <% end %>
   <% end %>
+<% end %>

data/lib/bulkrax/version.rb

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
 
 module Bulkrax
-  VERSION = '4.4.0'
+  VERSION = '5.0.0'
 end

data/lib/bulkrax.rb

@@ -4,8 +4,13 @@ require "bulkrax/version"
 require "bulkrax/engine"
 require 'active_support/all'
 
+# rubocop:disable Metrics/ModuleLength
 module Bulkrax
   class << self
+    # @todo Move from module attribute methods to a configuration class.  With module attributes,
+    #       when we make a change we are polluting the global space.  This means that our tests that
+    #       modify these config values are modifying global state.  Which is not desirous, as it can
+    #       introduce unexpected flakey tests.
     mattr_accessor :api_definition,
                    :default_field_mapping,
                    :default_work_type,
@@ -168,4 +173,46 @@ module Bulkrax
   def self.setup
     yield self
   end
+
+  # Responsible for stripping hidden characters from the given string.
+  #
+  # @param value [#to_s]
+  # @return [String] with hidden characters removed
+  #
+  # @see https://github.com/samvera-labs/bulkrax/issues/688
+  def self.normalize_string(value)
+    # Removing [Byte Order Mark (BOM)](https://en.wikipedia.org/wiki/Byte_order_mark)
+    value.to_s.delete("\xEF\xBB\xBF")
+  end
+
+  # This class confirms to the Active::Support.serialze interface.  It's job is to ensure that we
+  # don't have keys with the tricksy Byte Order Mark character.
+  #
+  # @see https://api.rubyonrails.org/classes/ActiveRecord/AttributeMethods/Serialization/ClassMethods.html#method-i-serialize
+  class NormalizedJson
+    def self.normalize_keys(hash)
+      return hash unless hash.respond_to?(:each_pair)
+      returning_value = {}
+      hash.each_pair do |key, value|
+        returning_value[Bulkrax.normalize_string(key)] = value
+      end
+      returning_value
+    end
+
+    # When we write the serialized data to the database, we "dump" the value into that database
+    # column.
+    def self.dump(value)
+      JSON.dump(normalize_keys(value))
+    end
+
+    # When we load the serialized data from the database, we pass the database's value into "load"
+    # function.
+    #
+    # rubocop:disable Security/JSONLoad
+    def self.load(string)
+      normalize_keys(JSON.load(string))
+    end
+    # rubocop:enable Security/JSONLoad
+  end
 end
+# rubocop:disable Metrics/ModuleLength

data/lib/generators/bulkrax/install_generator.rb

@@ -55,6 +55,26 @@ class Bulkrax::InstallGenerator < Rails::Generators::Base
     end
   end
 
+  def add_ability
+    file = 'app/models/ability.rb'
+    file_text = File.read(file)
+    import_line = 'def can_import_works?'
+    export_line = 'def can_export_works?'
+    unless file_text.include?(import_line)
+      insert_into_file file, before: /^end/ do
+        "  def can_import_works?\n    can_create_any_work?\n  end"
+      end
+    end
+
+    # rubocop:disable Style/GuardClause
+    unless file_text.include?(export_line)
+      insert_into_file file, before: /^end/ do
+        "  def can_export_works?\n    can_create_any_work?\n  end"
+      end
+    end
+    # rubocop:enable Style/GuardClause
+  end
+
   def add_css
     ['css', 'scss', 'sass'].map do |ext|
       file = "app/assets/stylesheets/application.#{ext}"

data/lib/generators/bulkrax/templates/config/initializers/bulkrax.rb

@@ -80,4 +80,6 @@ Bulkrax.setup do |config|
 end
 
 # Sidebar for hyrax 3+ support
-Hyrax::DashboardController.sidebar_partials[:repository_content] << "hyrax/dashboard/sidebar/bulkrax_sidebar_additions" if Object.const_defined?(:Hyrax) && ::Hyrax::DashboardController&.respond_to?(:sidebar_partials)
+if Object.const_defined?(:Hyrax) && ::Hyrax::DashboardController&.respond_to?(:sidebar_partials)
+  Hyrax::DashboardController.sidebar_partials[:repository_content] << "hyrax/dashboard/sidebar/bulkrax_sidebar_additions"
+end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: bulkrax
 version: !ruby/object:Gem::Version
-  version: 4.4.0
+  version: 5.0.0
 platform: ruby
 authors:
 - Rob Kaufman
 autorequire:
 bindir: bin
 cert_chain: []
-date: 2022-12-01 00:00:00.000000000 Z
+date: 2023-01-24 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: rails
@@ -106,14 +106,14 @@ dependencies:
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: 3.1.0
+        version: 3.2.4
   type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: 3.1.0
+        version: 3.2.4
 - !ruby/object:Gem::Dependency
   name: loofah
   requirement: !ruby/object:Gem::Requirement