bulkrax 3.0.0.beta5 → 3.0.0.beta8

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 994ee06cdd4ed63d0dfa79e0e77b1fb6b017528926a7b891d8ff702a15cf7347
-  data.tar.gz: 55298f93ce2078f1157a1426ce63823b78ae8a04cd4b076947a2ba1ad93ad05c
+  metadata.gz: 7cfcae9e154506ea5b6699b530da3f2d370d478963de1cb5ce854db1e0ded3ed
+  data.tar.gz: 7b1511cbaad98ef523ac5c1f53442d0f568b7088aba2312df77e104c41c8e192
 SHA512:
-  metadata.gz: 512b254802875cca2dae6377a1317439b25d35038aacd56ce2db18ee2aaaf9d0f1d22bb2f521ce16366a7a5630d5768468f3949579216fc02502b36e06e6159a
-  data.tar.gz: b25f1d98da59db8bc91f02d90fe44003350d7bf7eb7b67d5610d2006f7e0f5e1b72e9d37601b3a41f01faf34c8c07d8c388e173084edc6ebaf985e5f124f7d4a
+  metadata.gz: af73e975cd84be460ff29e2364a79e4fdae77793ca2a1b50534b8582e602138bd9c700af2068221907eee055f593f2305afab6d304385ba4f53e5ed72c64678f
+  data.tar.gz: 06bc7006d8b701dda2dd68487252e563f8b46484012124d5f81e43d3aa8af4916ba3c378948ad91157dce125768e2ea01e117673288fc8dbe1e9001cef03eb7e

data/app/factories/bulkrax/object_factory.rb

@@ -143,13 +143,11 @@ module Bulkrax
       object.attributes = attrs
       object.apply_depositor_metadata(@user)
       object.save!
-      pass_collection_to_be_persisted(parent: find_collection(attributes[related_parents_parsed_mapping]), child: object) if attributes[related_parents_parsed_mapping].present?
     end
 
     def update_collection(attrs)
       object.attributes = attrs
       object.save!
-      pass_collection_to_be_persisted(parent: find_collection(attributes[related_parents_parsed_mapping]), child: object) if attributes[related_parents_parsed_mapping].present?
     end
 
     # This method is heavily inspired by Hyrax's AttachFilesToWorkJob
@@ -182,37 +180,6 @@ module Bulkrax
       actor.update_metadata(file_set_attrs)
     end
 
-    def pass_collection_to_be_persisted(parent:, child:)
-      if parent.is_a? Array
-        parent.each { |par| persist_collection_memberships(par, child) }
-      else
-        persist_collection_memberships(parent, child)
-      end
-    end
-
-    # Add child to parent's #member_collections
-    # Add parent to child's #member_of_collections
-    def persist_collection_memberships(parent, child)
-      parent.reject!(&:blank?) if parent.respond_to?(:reject!)
-      child.reject!(&:blank?) if child.respond_to?(:reject!)
-      return if parent.blank? || child.blank?
-
-      ::Hyrax::Collections::NestedCollectionPersistenceService.persist_nested_collection_for(parent: parent, child: child)
-    end
-
-    def find_collection(id)
-      case id
-      when Hash
-        Collection.find(id[:id])
-      when String
-        Collection.find(id) if id.present?
-      when Array
-        id.map { |i| find_collection(i) }
-      else
-        []
-      end
-    end
-
     def clean_attrs(attrs)
       # avoid the "ArgumentError: Identifier must be a string of size > 0 in order to be treeified" error
       # when setting object.attributes
@@ -237,7 +204,7 @@ module Bulkrax
 
     # 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 member_of_collections_attributes]
+      klass.properties.keys.map(&:to_sym) + %i[id edit_users edit_groups read_groups visibility work_members_attributes admin_set_id]
     end
   end
 end

data/app/jobs/bulkrax/create_relationships_job.rb

@@ -83,17 +83,7 @@ module Bulkrax
     # This is adding the reverse relationship, from the child to the parent
     def collection_parent_work_child
       child_records[:works].each do |child_record|
-        attrs = { id: child_record.id, member_of_collections_attributes: { 0 => { id: parent_record.id } } }
-        ObjectFactory.new(
-          attributes: attrs,
-          source_identifier_value: nil, # sending the :id in the attrs means the factory doesn't need a :source_identifier_value
-          work_identifier: parent_entry&.parser&.work_identifier,
-          related_parents_parsed_mapping: parent_entry&.parser&.related_parents_parsed_mapping,
-          replace_files: false,
-          user: user,
-          klass: child_record.class,
-          importer_run_id: importer_run_id
-        ).run
+        ::Hyrax::Collections::NestedCollectionPersistenceService.persist_nested_collection_for(parent: parent_record, child: child_record)
         # TODO: add counters for :processed_parents and :failed_parents
         Bulkrax::ImporterRun.find(importer_run_id).increment!(:processed_relationships) # rubocop:disable Rails/SkipsModelValidations
       end
@@ -101,20 +91,10 @@ module Bulkrax
 
     # Collection-Collection membership is added to the as member_ids
     def collection_parent_collection_child
-      child_record = child_records[:collections].first
-      attrs = { id: parent_record.id, child_collection_id: child_record.id }
-      ObjectFactory.new(
-        attributes: attrs,
-        source_identifier_value: nil, # sending the :id in the attrs means the factory doesn't need a :source_identifier_value
-        work_identifier: parent_entry&.parser&.work_identifier,
-        related_parents_parsed_mapping: parent_entry&.parser&.related_parents_parsed_mapping,
-        replace_files: false,
-        user: user,
-        klass: parent_record.class,
-        importer_run_id: importer_run_id
-      ).run
-      # TODO: add counters for :processed_parents and :failed_parents
-      Bulkrax::ImporterRun.find(importer_run_id).increment!(:processed_relationships) # rubocop:disable Rails/SkipsModelValidations
+      child_records[:collections].each do |child_record|
+        ::Hyrax::Collections::NestedCollectionPersistenceService.persist_nested_collection_for(parent: parent_record, child: child_record)
+        Bulkrax::ImporterRun.find(importer_run_id).increment!(:processed_relationships) # rubocop:disable Rails/SkipsModelValidations
+      end
     end
 
     # Work-Work membership is added to the parent as member_ids
@@ -124,19 +104,11 @@ module Bulkrax
         records_hash[i] = { id: child_record.id }
       end
       attrs = {
-        id: parent_record.id,
         work_members_attributes: records_hash
       }
-      ObjectFactory.new(
-        attributes: attrs,
-        source_identifier_value: nil, # sending the :id in the attrs means the factory doesn't need a :source_identifier_value
-        work_identifier: parent_entry&.parser&.work_identifier,
-        related_parents_parsed_mapping: parent_entry&.parser&.related_parents_parsed_mapping,
-        replace_files: false,
-        user: user,
-        klass: parent_record.class,
-        importer_run_id: importer_run_id
-      ).run
+      parent_record.reindex_extent = Hyrax::Adapters::NestingIndexAdapter::LIMITED_REINDEX if parent_record.respond_to?(:reindex_extent)
+      env = Hyrax::Actors::Environment.new(parent_record, Ability.new(user), attrs)
+      Hyrax::CurationConcern.actor.update(env)
       # TODO: add counters for :processed_parents and :failed_parents
       Bulkrax::ImporterRun.find(importer_run_id).increment!(:processed_relationships) # rubocop:disable Rails/SkipsModelValidations
     end

data/app/jobs/bulkrax/import_collection_job.rb

@@ -10,7 +10,6 @@ module Bulkrax
       begin
         entry.build
         entry.save
-        add_user_to_permission_template!(entry)
         ImporterRun.find(args[1]).increment!(:processed_records)
         ImporterRun.find(args[1]).increment!(:processed_collections)
         ImporterRun.find(args[1]).decrement!(:enqueued_records)
@@ -22,22 +21,5 @@ module Bulkrax
       end
     end
     # rubocop:enable Rails/SkipsModelValidations
-
-    private
-
-    def add_user_to_permission_template!(entry)
-      user                = ::User.find(entry.importerexporter.user_id)
-      collection          = entry.factory.find
-      permission_template = Hyrax::PermissionTemplate.find_or_create_by!(source_id: collection.id)
-
-      Hyrax::PermissionTemplateAccess.find_or_create_by!(
-        permission_template_id: permission_template.id,
-        agent_id: user.user_key,
-        agent_type: 'user',
-        access: 'manage'
-      )
-
-      collection.reset_access_controls!
-    end
   end
 end

data/app/models/bulkrax/csv_entry.rb

@@ -59,6 +59,7 @@ module Bulkrax
       if factory_class == Collection
         add_collection_type_gid
       elsif factory_class == FileSet
+        validate_presence_of_filename!
         add_path_to_file
         validate_presence_of_parent!
       else
@@ -250,7 +251,7 @@ module Bulkrax
       split_references = record[parent_field_mapping].split(/\s*[;|]\s*/)
       split_references.each do |c_reference|
         matching_collection_entries = importerexporter.entries.select do |e|
-          (e.raw_metadata[source_identifier] == c_reference) &&
+          (e.raw_metadata&.[](source_identifier) == c_reference) &&
             e.is_a?(CsvCollectionEntry)
         end
         raise ::StandardError, 'Only expected to find one matching entry' if matching_collection_entries.count > 1

data/app/models/bulkrax/csv_file_set_entry.rb

@@ -17,6 +17,12 @@ module Bulkrax
       parsed_metadata['file']
     end
 
+    def validate_presence_of_filename!
+      return if parsed_metadata&.[]('file')&.map(&:present?)&.any?
+
+      raise StandardError, 'File set must have a filename'
+    end
+
     def validate_presence_of_parent!
       return if parsed_metadata[related_parents_parsed_mapping]&.map(&:present?)&.any?
 

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

@@ -11,6 +11,7 @@ module Bulkrax
         unless self.importerexporter.validate_only
           raise CollectionsCreatedError unless collections_created?
           @item = factory.run!
+          add_user_to_permission_templates! if self.class.to_s.include?("Collection")
           parent_jobs if self.parsed_metadata[related_parents_parsed_mapping].present?
           child_jobs if self.parsed_metadata[related_children_parsed_mapping].present?
         end
@@ -24,6 +25,25 @@ module Bulkrax
       return @item
     end
 
+    def add_user_to_permission_templates!
+      permission_template = Hyrax::PermissionTemplate.find_or_create_by!(source_id: @item.id)
+
+      Hyrax::PermissionTemplateAccess.find_or_create_by!(
+        permission_template_id: permission_template.id,
+        agent_id: user.user_key,
+        agent_type: 'user',
+        access: 'manage'
+      )
+      Hyrax::PermissionTemplateAccess.find_or_create_by!(
+        permission_template_id: permission_template.id,
+        agent_id: 'admin',
+        agent_type: 'group',
+        access: 'manage'
+      )
+
+      @item.reset_access_controls!
+    end
+
     def parent_jobs
       self.parsed_metadata[related_parents_parsed_mapping].each do |parent_identifier|
         next if parent_identifier.blank?

data/app/parsers/bulkrax/csv_parser.rb

@@ -4,6 +4,8 @@ require 'csv'
 module Bulkrax
   class CsvParser < ApplicationParser # rubocop:disable Metrics/ClassLength
     include ErroredEntries
+    attr_writer :collections, :file_sets, :works
+
     def self.export_supported?
       true
     end
@@ -17,37 +19,56 @@ module Bulkrax
       @records ||= csv_data.map { |record_data| entry_class.data_for_entry(record_data, nil, self) }
     end
 
-    def collections
-      # retrieve a list of unique collections
-      records.map do |r|
-        collections = []
-        model_field_mappings.each do |model_mapping|
-          collections << r if r[model_mapping.to_sym]&.downcase == 'collection'
+    def build_records
+      @collections = []
+      @works = []
+      @file_sets = []
+
+      if model_field_mappings.map { |mfm| mfm.to_sym.in?(records.first.keys) }.any?
+        records.map do |r|
+          model_field_mappings.map(&:to_sym).each do |model_mapping|
+            next unless r.key?(model_mapping)
+
+            if r[model_mapping].casecmp('collection').zero?
+              @collections << r
+            elsif r[model_mapping].casecmp('fileset').zero?
+              @file_sets << r
+            else
+              @works << r
+            end
+          end
         end
-        collections
-      end.flatten.compact.uniq
+        @collections = @collections.flatten.compact.uniq
+        @file_sets = @file_sets.flatten.compact.uniq
+        @works = @works.flatten.compact.uniq
+      else # if no model is specified, assume all records are works
+        @works = records.flatten.compact.uniq
+      end
+
+      true
     end
 
-    def collections_total
-      collections.size
+    def collections
+      build_records if @collections.nil?
+      @collections
     end
 
     def works
-      records - collections - file_sets
+      build_records if @works.nil?
+      @works
     end
 
-    def works_total
-      works.size
+    def file_sets
+      build_records if @file_sets.nil?
+      @file_sets
     end
 
-    def file_sets
-      records.map do |r|
-        file_sets = []
-        model_field_mappings.each do |model_mapping|
-          file_sets << r if r[model_mapping.to_sym]&.downcase == 'fileset'
-        end
-        file_sets
-      end.flatten.compact.uniq
+    def collections_total
+      collections.size
+    end
+
+    def works_total
+      works.size
     end
 
     def file_sets_total
@@ -96,21 +117,24 @@ module Bulkrax
     end
 
     def create_objects(types_array = nil)
-      (types_array || %w[work collection file_set relationship]).each do |type|
+      index = 0
+      (types_array || %w[collection work file_set relationship]).each do |type|
         if type.eql?('relationship')
           ScheduleRelationshipsJob.set(wait: 5.minutes).perform_later(importer_id: importerexporter.id)
           next
         end
-        send(type.pluralize).each_with_index do |current_record, index|
-          next unless record_has_source_identifier(current_record, records.find_index(current_record))
-          break if limit_reached?(limit, records.find_index(current_record))
+        send(type.pluralize).each do |current_record|
+          next unless record_has_source_identifier(current_record, index)
+          break if limit_reached?(limit, index)
 
           seen[current_record[source_identifier]] = true
           create_entry_and_job(current_record, type)
           increment_counters(index, "#{type}": true)
+          index += 1
         end
         importer.record_status
       end
+      true
     rescue StandardError => e
       status_info(e)
     end
@@ -193,7 +217,7 @@ module Bulkrax
         instance_variable_set(instance_var, ActiveFedora::SolrService.post(
           extra_filters.to_s,
           fq: [
-            "#{work_identifier}_sim:(#{complete_entry_identifiers.join(' OR ')})",
+            "#{::Solrizer.solr_name(work_identifier)}:(#{complete_entry_identifiers.join(' OR ')})",
             "has_model_ssim:(#{models_to_search.join(' OR ')})"
           ],
           fl: 'id',

data/lib/bulkrax/version.rb

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
 
 module Bulkrax
-  VERSION = '3.0.0.beta5'
+  VERSION = '3.0.0.beta8'
 end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: bulkrax
 version: !ruby/object:Gem::Version
-  version: 3.0.0.beta5
+  version: 3.0.0.beta8
 platform: ruby
 authors:
 - Rob Kaufman
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2022-04-27 00:00:00.000000000 Z
+date: 2022-05-11 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: rails