bulkrax 3.0.0.beta8 → 3.1.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 7cfcae9e154506ea5b6699b530da3f2d370d478963de1cb5ce854db1e0ded3ed
-  data.tar.gz: 7b1511cbaad98ef523ac5c1f53442d0f568b7088aba2312df77e104c41c8e192
+  metadata.gz: 23cd36b2c3dc13769f7d20da3c0d53f2d5c2cabef15d5beeeccee75aed6d679a
+  data.tar.gz: 868f0ec5c54862943e126af0ecf63d49771bc9a4ac88e566074205d7077e063c
 SHA512:
-  metadata.gz: af73e975cd84be460ff29e2364a79e4fdae77793ca2a1b50534b8582e602138bd9c700af2068221907eee055f593f2305afab6d304385ba4f53e5ed72c64678f
-  data.tar.gz: 06bc7006d8b701dda2dd68487252e563f8b46484012124d5f81e43d3aa8af4916ba3c378948ad91157dce125768e2ea01e117673288fc8dbe1e9001cef03eb7e
+  metadata.gz: a4f9c359a539054dba54000ca30bd1209bb862254c75463c2598807e8fd91e7e562e3bd6203f5f347ffe17da22f58d7915c1090109e88aabccac774c0281b1ee
+  data.tar.gz: 11c0083e67c4add719f04c01d20fac3216f6e994407dfdc48ae5f6d82696241a2b81ef6156279568b92e4bd56d3598085e83bac3718d04892a4fc96bbefd1028

data/app/jobs/bulkrax/create_relationships_job.rb

@@ -82,18 +82,18 @@ module Bulkrax
     # Work-Collection membership is added to the child as member_of_collection_ids
     # This is adding the reverse relationship, from the child to the parent
     def collection_parent_work_child
-      child_records[:works].each do |child_record|
-        ::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
+      child_work_ids = child_records[:works].map(&:id)
+      parent_record.reindex_extent = Hyrax::Adapters::NestingIndexAdapter::LIMITED_REINDEX
+
+      parent_record.add_member_objects(child_work_ids)
+      ImporterRun.find(importer_run_id).increment!(:processed_relationships, child_work_ids.count) # rubocop:disable Rails/SkipsModelValidations
     end
 
     # Collection-Collection membership is added to the as member_ids
     def collection_parent_collection_child
       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
+        ImporterRun.find(importer_run_id).increment!(:processed_relationships) # rubocop:disable Rails/SkipsModelValidations
       end
     end
 
@@ -103,14 +103,12 @@ module Bulkrax
       child_records[:works].each_with_index do |child_record, i|
         records_hash[i] = { id: child_record.id }
       end
-      attrs = {
-        work_members_attributes: records_hash
-      }
-      parent_record.reindex_extent = Hyrax::Adapters::NestingIndexAdapter::LIMITED_REINDEX if parent_record.respond_to?(:reindex_extent)
+      attrs = { work_members_attributes: records_hash }
+      parent_record.try(:reindex_extent=, Hyrax::Adapters::NestingIndexAdapter::LIMITED_REINDEX)
       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
+      ImporterRun.find(importer_run_id).increment!(:processed_relationships, child_records[:works].count) # rubocop:disable Rails/SkipsModelValidations
     end
 
     def reschedule(parent_identifier:, importer_run_id:)

data/app/jobs/bulkrax/export_work_job.rb

@@ -27,6 +27,7 @@ module Bulkrax
       end
       exporter_run = ExporterRun.find(args[1])
       return entry if exporter_run.enqueued_records.positive?
+
       if exporter_run.failed_records.positive?
         exporter_run.exporter.status_info('Complete (with failures)')
       else

data/app/models/bulkrax/csv_entry.rb

@@ -46,6 +46,7 @@ module Bulkrax
       add_visibility
       add_metadata_for_model
       add_rights_statement
+      sanitize_controlled_uri_values!
       add_local
 
       self.parsed_metadata

data/app/models/bulkrax/exporter.rb

@@ -96,7 +96,8 @@ module Bulkrax
     end
 
     def current_run
-      @current_run ||= self.exporter_runs.create!(total_work_entries: self.limit || parser.total)
+      total = self.limit || parser.total
+      @current_run ||= self.exporter_runs.create!(total_work_entries: total, enqueued_records: total)
     end
 
     def last_run

data/app/models/bulkrax/importer.rb

@@ -142,7 +142,7 @@ module Bulkrax
 
     def import_objects(types_array = nil)
       self.only_updates ||= false
-      types = types_array || %w[work collection file_set relationship]
+      types = types_array || %w[collection work file_set relationship]
       if parser.class == Bulkrax::CsvParser
         parser.create_objects(types)
       else

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

@@ -2,7 +2,7 @@
 
 module Bulkrax
   # Import Behavior for Entry classes
-  module ImportBehavior
+  module ImportBehavior # rubocop:disable Metrics/ModuleLength
     extend ActiveSupport::Concern
 
     def build_for_importer
@@ -21,6 +21,8 @@ module Bulkrax
         status_info(e)
       else
         status_info
+      ensure
+        self.save!
       end
       return @item
     end
@@ -103,6 +105,65 @@ module Bulkrax
       end
     end
 
+    # Attempt to sanitize Questioning Authority URI values for configured controlled fields of common
+    # data entry mistakes. Controlled URI values are only valid if they are an exact match.
+    # Example:
+    #   Valid value:     http://rightsstatements.org/vocab/InC/1.0/
+    #   Provided value:  https://rightsstatements.org/vocab/InC/1.0
+    #   Sanitized value: http://rightsstatements.org/vocab/InC/1.0/ ("s" from "https" removed, trailing "/" added)
+    #
+    # @return [Boolean] true if all controlled URI values are sanitized successfully
+    def sanitize_controlled_uri_values!
+      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
+          end
+        end
+      end
+
+      true
+    end
+
+    # @param value [String] value to validate
+    # @param field [String] name of the controlled property
+    # @return [String, nil] validated URI value or nil
+    def validate_value(value, field)
+      if value.match?(::URI::DEFAULT_PARSER.make_regexp)
+        value = value.strip.chomp
+        # add trailing forward slash unless one is already present
+        value << '/' unless value.match?(%r{/$})
+      end
+
+      valid = if active_id_for_authority?(value, field)
+                true
+              else
+                value.include?('https') ? value.sub!('https', 'http') : value.sub!('http', 'https')
+                active_id_for_authority?(value, field)
+              end
+
+      valid ? value : nil
+    end
+
+    # @param value [String] value to check
+    # @param field [String] name of the controlled property
+    # @return [Boolean] provided value is a present, active authority ID for the provided field
+    def active_id_for_authority?(value, field)
+      field_service = ('Hyrax::' + "#{field}_service".camelcase).constantize
+      active_authority_ids = field_service.new.active_elements.map { |ae| ae['id'] }
+
+      active_authority_ids.include?(value)
+    end
+
     def factory
       @factory ||= Bulkrax::ObjectFactory.new(attributes: self.parsed_metadata,
                                               source_identifier_value: identifier,

data/app/parsers/bulkrax/csv_parser.rb

@@ -270,7 +270,7 @@ module Bulkrax
     #   Changed to grep as wc -l counts blank lines, and ignores the final unescaped line (which may or may not contain data)
     def total
       @total = importer.parser_fields['total'] || 0 if importer?
-      @total = importerexporter.entries.count if exporter?
+      @total = limit || current_record_ids.count if exporter?
 
       return @total || 0
     rescue StandardError

data/lib/bulkrax/version.rb

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
 
 module Bulkrax
-  VERSION = '3.0.0.beta8'
+  VERSION = '3.1.0'
 end

data/lib/bulkrax.rb

@@ -12,13 +12,13 @@ module Bulkrax
                    :related_children_field_mapping,
                    :related_parents_field_mapping,
                    :reserved_properties,
+                   :qa_controlled_properties,
                    :field_mappings,
                    :import_path,
                    :export_path,
                    :removed_image_path,
                    :server_name,
-                   :api_definition,
-                   :removed_image_path
+                   :api_definition
 
     self.parsers = [
       { name: "OAI - Dublin Core", class_name: "Bulkrax::OaiDcParser", partial: "oai_fields" },
@@ -119,6 +119,11 @@ module Bulkrax
       original_url
       relative_path
     ]
+
+    # List of Questioning Authority properties that are controlled via YAML files in
+    # the config/authorities/ directory. For example, the :rights_statement property
+    # is controlled by the active terms in config/authorities/rights_statements.yml
+    self.qa_controlled_properties = %w[rights_statement license]
   end
 
   def self.api_definition
@@ -131,8 +136,6 @@ module Bulkrax
     )
   end
 
-  self.removed_image_path = 'app/assets/images/bulkrax/removed.png'
-
   # this function maps the vars from your app into your engine
   def self.setup
     yield self

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

@@ -61,6 +61,12 @@ Bulkrax.setup do |config|
 
   # Properties that should not be used in imports/exports. They are reserved for use by Hyrax.
   # config.reserved_properties += ['my_field']
+
+  # List of Questioning Authority properties that are controlled via YAML files in
+  # the config/authorities/ directory. For example, the :rights_statement property
+  # is controlled by the active terms in config/authorities/rights_statements.yml
+  # Defaults: 'rights_statement' and 'license'
+  # config.qa_controlled_properties += ['my_field']
 end
 
 # Sidebar for hyrax 3+ support

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: bulkrax
 version: !ruby/object:Gem::Version
-  version: 3.0.0.beta8
+  version: 3.1.0
 platform: ruby
 authors:
 - Rob Kaufman
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2022-05-11 00:00:00.000000000 Z
+date: 2022-05-19 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: rails
@@ -393,9 +393,9 @@ required_ruby_version: !ruby/object:Gem::Requirement
       version: '0'
 required_rubygems_version: !ruby/object:Gem::Requirement
   requirements:
-  - - ">"
+  - - ">="
     - !ruby/object:Gem::Version
-      version: 1.3.1
+      version: '0'
 requirements: []
 rubygems_version: 3.1.4
 signing_key: