bulkrax 9.0.2 → 9.2.0

data/app/parsers/bulkrax/csv_parser.rb

@@ -38,7 +38,7 @@ module Bulkrax
             #       We aren't right now because so many Bulkrax users are in between Fedora and Valkyrie
             if model.casecmp('collection').zero? || model.casecmp('collectionresource').zero?
               @collections << r
-            elsif model.casecmp('fileset').zero?
+            elsif model.casecmp('fileset').zero? || model.casecmp('hyrax::fileset').zero?
               @file_sets << r
             else
               @works << r
@@ -248,14 +248,22 @@ module Bulkrax
       if file_sets.nil? # for valkyrie
         file_sets = record.respond_to?(:file_sets) ? record.file_sets : record.members&.select(&:file_set?)
       end
-      file_sets << record.thumbnail if exporter.include_thumbnails && record.thumbnail.present? && record.work?
+
+      if importerexporter.include_thumbnails?
+        thumbnail = Bulkrax.object_factory.thumbnail_for(resource: record)
+        file_sets << thumbnail if thumbnail.present?
+      end
+
       file_sets.each do |fs|
         path = File.join(exporter_export_path, folder_count, 'files')
         FileUtils.mkdir_p(path) unless File.exist? path
+
+        original_file = Bulkrax.object_factory.original_file(fileset: fs)
+        next if original_file.blank?
         file = filename(fs)
-        next if file.blank? || fs.original_file.blank?
 
-        io = fs.original_file.respond_to?(:uri) ? open(fs.original_file.uri) : fs.original_file.file.io
+        io = original_file.respond_to?(:uri) ? open(original_file.uri) : original_file.file.io
+
         File.open(File.join(path, file), 'wb') do |f|
           f.write(io.read)
           f.close
@@ -263,6 +271,8 @@ module Bulkrax
       end
     rescue Ldp::Gone
       return
+    rescue StandardError => e
+      raise StandardError, "Unable to retrieve files for identifier #{identifier} - #{e.message}"
     end
 
     def export_key_allowed(key)
@@ -341,7 +351,16 @@ module Bulkrax
         file_mapping = Bulkrax.field_mappings.dig(self.class.to_s, 'file', :from)&.first&.to_sym || :file
         next if r[file_mapping].blank?
 
-        r[file_mapping].split(Bulkrax.multi_value_element_split_on).map do |f|
+        split_value = Bulkrax.field_mappings.dig(self.class.to_s, :file, :split)
+        split_pattern = case split_value
+                        when Regexp
+                          split_value
+                        when String
+                          Regexp.new(split_value)
+                        else
+                          Bulkrax.multi_value_element_split_on
+                        end
+        r[file_mapping].split(split_pattern).map do |f|
           file = File.join(path_to_files, f.tr(' ', '_'))
           if File.exist?(file) # rubocop:disable Style/GuardClause
             file
@@ -360,6 +379,11 @@ module Bulkrax
       @path_to_files = File.join(
           zip? ? importer_unzip_path : File.dirname(import_file_path), 'files', filename
         )
+
+      return @path_to_files if File.exist?(@path_to_files)
+
+      # TODO: This method silently returns nil if there is no file & no zip file
+      File.join(importer_unzip_path, 'files', filename) if file? && zip?
     end
 
     private
@@ -379,9 +403,15 @@ module Bulkrax
     # We expect a single CSV at the top level of the zip in the CSVParser
     # but we are willing to go look for it if need be
     def real_import_file_path
-      return Dir["#{importer_unzip_path}/**/*.csv"].first if file? && zip?
+      return Dir["#{importer_unzip_path}/**/*.csv"].reject { |path| in_files_dir?(path) }.first if file? && zip?
 
       parser_fields['import_file_path']
     end
+
+    # If there are CSVs that are meant to be attachments in the files directory,
+    # we don't want to consider them as the import CSV
+    def in_files_dir?(path)
+      File.dirname(path).ends_with?('files')
+    end
   end
 end

data/app/parsers/bulkrax/oai_dc_parser.rb

@@ -61,8 +61,6 @@ module Bulkrax
       ['contributor', 'coverage', 'creator', 'date', 'description', 'format', 'identifier', 'language', 'publisher', 'relation', 'rights', 'source', 'subject', 'title', 'type']
     end
 
-    delegate :list_sets, to: :client
-
     def create_objects(types = [])
       types.each do |object_type|
         send("create_#{object_type.pluralize}")

data/app/parsers/bulkrax/xml_parser.rb

@@ -103,7 +103,7 @@ module Bulkrax
     end
 
     def good_file_type?(path)
-      %w[.xml .xls .xsd].include?(File.extname(path)) || ::Marcel::MimeType.for(path).include?('application/xml')
+      %w[.xml .xls .xsd].include?(File.extname(path)) || ::Marcel::MimeType.for(path: path).include?('application/xml')
     end
 
     def total

data/app/services/bulkrax/factory_class_finder.rb

@@ -66,27 +66,68 @@ module Bulkrax
       entry.default_work_type.constantize
     end
 
+    private
+
     ##
     # @api private
     # @return [String]
     def name
-      fc = if entry.parsed_metadata&.[]('model').present?
-             Array.wrap(entry.parsed_metadata['model']).first
-           elsif entry.importerexporter&.mapping&.[]('work_type').present?
-             # Because of delegation's nil guard, we're reaching rather far into the implementation
-             # details.
-             Array.wrap(entry.parsed_metadata['work_type']).first
-           else
-             entry.default_work_type
-           end
-
-      # Let's coerce this into the right shape; we're not mutating the string because it might well
-      # be frozen.
-      fc = fc.tr(' ', '_')
-      fc = fc.downcase if fc.match?(/[-_]/)
-      fc.camelcase
+      # Try each strategy in order until one returns a value
+      fc = find_factory_class_name || entry.default_work_type
+
+      # Normalize the string format
+      normalize_class_name(fc)
     rescue
       entry.default_work_type
     end
+
+    ##
+    # Try each strategy in sequence to find a factory class name
+    # @return [String, nil] the factory class name or nil if none found
+    def find_factory_class_name
+      prioritized_strategies = [
+        :model_from_parsed_metadata,
+        :work_type_from_parsed_metadata,
+        :model_from_raw_metadata,
+        :model_from_mapped_field
+      ]
+
+      # Return the first non-nil result
+      prioritized_strategies.each do |strategy|
+        result = send(strategy)
+        return result if result.present?
+      end
+
+      nil
+    end
+
+    def model_from_parsed_metadata
+      Array.wrap(entry.parsed_metadata['model']).first if entry.parsed_metadata&.[]('model').present?
+    end
+
+    def work_type_from_parsed_metadata
+      Array.wrap(entry.parsed_metadata['work_type']).first if entry.importerexporter&.mapping&.[]('work_type').present?
+    end
+
+    def model_from_raw_metadata
+      Array.wrap(entry.raw_metadata&.[]('model'))&.first if entry.raw_metadata&.[]('model').present?
+    end
+
+    def model_from_mapped_field
+      return nil unless entry.parser.model_field_mappings.any? { |field| entry.raw_metadata&.[](field).present? }
+      field = entry.parser.model_field_mappings.find { |f| entry.raw_metadata&.[](f).present? }
+      Array.wrap(entry.raw_metadata[field]).first
+    end
+
+    ##
+    # Normalize a class name string to proper format
+    # @param name [String] the class name to normalize
+    # @return [String] the normalized class name
+    def normalize_class_name(name)
+      name = name.to_s
+      name = name.tr(' ', '_')
+      name = name.downcase if name.match?(/[-_]/)
+      name.camelcase
+    end
   end
 end

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

@@ -6,7 +6,7 @@ module Hyrax
     # @see https://github.com/samvera/valkyrie/wiki/Queries#custom-queries
     class FindBySourceIdentifier
       def self.queries
-        [:find_by_model_and_property_value]
+        [:find_by_property_value]
       end
 
       def initialize(query_service:)
@@ -18,30 +18,25 @@ module Hyrax
       delegate :orm_class, to: :resource_factory
 
       ##
-      # @param model [Class, #internal_resource]
       # @param property [#to_s] the name of the property we're attempting to
       #        query.
       # @param value [#to_s] the propety's value that we're trying to match.
       #
       # @return [NilClass] when no record was found
       # @return [Valkyrie::Resource] when a record was found
-      #
-      # @note This is not a real estate transaction nor a Zillow lookup.
-      def find_by_model_and_property_value(model:, property:, value:)
-        sql_query = sql_for_find_by_model_and_property_value
-        # NOTE: Do we need to ask the model for it's internal_resource?
-        # TODO: no => undefined method `internal_resource' for Image:Class
-        query_service.run_query(sql_query, model, property, value).first
+      def find_by_property_value(property:, value:, **)
+        sql_query = sql_for_find_by_property_value
+        query_service.run_query(sql_query, property, value.to_s).first
       end
 
       private
 
-      def sql_for_find_by_model_and_property_value
+      def sql_for_find_by_property_value
         # NOTE: This is querying the first element of the property, but we might
         # want to check all of the elements.
         <<-SQL
           SELECT * FROM orm_resources
-          WHERE internal_resource = ? AND metadata -> ? ->> 0 = ?
+          WHERE metadata -> ? ->> 0 = ?
           LIMIT 1;
         SQL
       end

data/app/services/wings/custom_queries/find_by_source_identifier.rb

@@ -6,7 +6,7 @@ module Wings
       # Custom query override specific to Wings
 
       def self.queries
-        [:find_by_model_and_property_value]
+        [:find_by_property_value]
       end
 
       attr_reader :query_service
@@ -16,11 +16,20 @@ module Wings
         @query_service = query_service
       end
 
-      def find_by_model_and_property_value(model:, property:, value:, use_valkyrie: Hyrax.config.use_valkyrie?)
-        # NOTE: This is using the Bulkrax::ObjectFactory (e.g. the one
-        # envisioned for ActiveFedora).  In doing this, we avoid the situation
-        # where Bulkrax::ValkyrieObjectFactory calls this custom query.
-        af_object = Bulkrax::ObjectFactory.search_by_property(value: value, klass: model, field: property)
+      # rubocop:disable Lint/UnusedMethodArgument
+      def find_by_property_value(property:, value:, search_field:, use_valkyrie: Hyrax.config.use_valkyrie?)
+        # rubocop:enable Lint/UnusedMethodArgument
+        # NOTE: This is using the Bulkrax::ObjectFactory (e.g. the one envisioned for ActiveFedora).
+        # In doing this, we avoid the situation where Bulkrax::ValkyrieObjectFactory calls this custom query.
+
+        # This is doing a solr search first, so we have to use the search_field instead of the property for "field"
+        # If it cannot find the object in Solr, it falls back to an ActiveFedora search, if the object is an ActiveFedora object
+        af_object = Bulkrax::ObjectFactory.search_by_property(
+          value: value,
+          klass: ActiveFedora::Base,
+          field: search_field,
+          name_field: property
+        )
 
         return if af_object.blank?
         return af_object unless use_valkyrie

data/app/views/bulkrax/entries/show.html.erb

@@ -33,17 +33,23 @@
 
     <p class='bulkrax-p-align'>
       <% if @importer.present? %>
-	<%# TODO Consider how to account for Bulkrax.collection_model_class %>
-        <% factory_record = @entry.factory.find %>
-        <% if factory_record.present? && @entry.factory_class %>
-          <strong><%= @entry.factory_class.model_name.human %> Link:</strong>
-          <% if defined?(Hyrax) && @entry.factory_class.model_name.human == 'Collection' %>
-            <%= link_to @entry.factory_class.model_name.human, hyrax.polymorphic_path(factory_record) %>
+	    <%# TODO Consider how to account for Bulkrax.collection_model_class %>
+        <% begin %>
+          <% factory_record = @entry.factory.find rescue nil %>
+          <% if factory_record.present? %>
+            <% factory_record_class = factory_record.class %>
+            <% factory_record_class_human = factory_record_class.model_name.human %>
+            <strong><%= factory_record_class_human %> Link:</strong>
+            <% if defined?(Hyrax) && factory_record_class_human == 'Collection' %>
+              <%= link_to factory_record_class_human, hyrax.polymorphic_path(factory_record) %>
+            <% else %>
+              <%= link_to factory_record_class_human, main_app.polymorphic_path(factory_record) %>
+            <% end %>
           <% else %>
-            <%= link_to @entry.factory_class.model_name.human, main_app.polymorphic_path(factory_record) %>
+            <strong>Item Link:</strong> Item has not yet been imported successfully
           <% end %>
-        <% else %>
-          <strong>Item Link:</strong> Item has not yet been imported successfully
+        <% rescue => e %>
+          <strong>Item Link:</strong> Unable to retrieve item (<%= e.message %>)
         <% end %>
       <% else %>
         <% record = @entry&.hyrax_record %>

data/app/views/bulkrax/importers/_bagit_fields.html.erb

@@ -35,7 +35,7 @@
           input_html: { class: 'form-control' } ,
           required: false
           %>
-    <%= fi.input :override_rights_statement, as: :boolean, hint: 'If checked, always use the selected rights statment. If unchecked, use rights or rights_statement from the record and only use the provided value if dc:rights is blank.', input_html: { checked: (importer.parser_fields['override_rights_statement'] == "1") } %>
+    <%= fi.input :override_rights_statement, as: :boolean, hint: 'If checked, always use the selected rights statement. If unchecked, use rights or rights_statement from the record and only use the provided value if dc:rights is blank.', input_html: { checked: (importer.parser_fields['override_rights_statement'] == "1") } %>
   <% end %>
   <h4>Bag or Bags to Import:</h4>
   <p>File upload and Cloud File upload must be a Zip file containing a single BagIt Bag, or a folder containing multiple BagIt Bags.</p>

data/app/views/bulkrax/importers/_csv_fields.html.erb

@@ -20,7 +20,7 @@
           input_html: { class: 'form-control' },
           required: false
           %>
-    <%= fi.input :override_rights_statement, as: :boolean, hint: 'If checked, always use the selected rights statment. If unchecked, use rights or rights_statement from the record and only use the provided value if dc:rights is blank.', input_html: { checked: (importer.parser_fields['override_rights_statement'] == "1") } %>
+    <%= fi.input :override_rights_statement, as: :boolean, hint: 'If checked, always use the selected rights statement. If unchecked, use rights or rights_statement from the record and only use the provided value if dc:rights is blank.', input_html: { checked: (importer.parser_fields['override_rights_statement'] == "1") } %>
   <% end %>
   <h4>Add CSV or ZIP File to Import:</h4>
   <%# accept a single file upload; data files and bags will need to be added another way %>

data/app/views/bulkrax/importers/_oai_fields.html.erb

@@ -26,7 +26,7 @@
           input_html: { class: 'form-control' },
           required: false
           %>
-    <%= fi.input :override_rights_statement, as: :boolean, hint: 'If checked, always use the selected rights statment. If unchecked, use dc:rights from the record and only use the provided value if dc:rights is blank.', input_html: { checked: (importer.parser_fields['override_rights_statement'] == "1") } %>
+    <%= fi.input :override_rights_statement, as: :boolean, hint: 'If checked, always use the selected rights statement. If unchecked, use dc:rights from the record and only use the provided value if dc:rights is blank.', input_html: { checked: (importer.parser_fields['override_rights_statement'] == "1") } %>
   <% end %>
   <%= fi.input :thumbnail_url, required: false, as: :string, input_html: { value: importer.parser_fields['thumbnail_url'] } %>
   <div class="help-block well well-sm">

data/app/views/bulkrax/importers/_xml_fields.html.erb

@@ -41,7 +41,7 @@
           item_helper: rights_statements.method(:include_current_value),
           input_html: { class: 'form-control' },
           required: false %>
-    <%= fi.input :override_rights_statement, as: :boolean, hint: 'If checked, always use the selected rights statment. If unchecked, use rights or rights_statement from the record and only use the provided value if dc:rights is blank.', input_html: { checked: (importer.parser_fields['override_rights_statement'] == "1") } %>
+    <%= fi.input :override_rights_statement, as: :boolean, hint: 'If checked, always use the selected rights statement. If unchecked, use rights or rights_statement from the record and only use the provided value if dc:rights is blank.', input_html: { checked: (importer.parser_fields['override_rights_statement'] == "1") } %>
   <% end %>
   <h4>XML and files to Import:</h4>
   <p>File upload and Cloud File upload MUST be a either a single XML file (for metadata only import) OR a Zip file containing the XML files and data files, each in a separate folder.</p>

data/app/views/bulkrax/importers/show.html.erb

@@ -1,10 +1,10 @@
-<div class="col-xs-12 main-header">
+<div class="col-xs-12 main-header d-flex justify-content-between align-items-center">
   <h1><span class="fa fa-cloud-upload" aria-hidden="true"></span> Importer: <%= @importer.name %></h1>
   <div class="pull-right">
-    <%= link_to 'Download Original File', importer_original_file_path(@importer.id), class: 'btn btn-primary', data: { turbolinks: false } if @importer.original_file %>
+    <%= link_to 'Download Original File', importer_original_file_path(@importer.id), class: 'btn btn-primary text-nowrap', data: { turbolinks: false } if @importer.original_file %>
     <% if @importer.failed_entries? %>
-      <%= link_to 'Export Errored Entries', importer_export_errors_path(@importer.id), class: 'btn btn-primary', data: { turbolinks: false }%>
-      <%= link_to 'Upload Corrected Entries', importer_upload_corrected_entries_path(@importer.id), class: 'btn btn-primary' if @importer.parser.is_a?(Bulkrax::CsvParser) %>
+      <%= link_to 'Export Errored Entries', importer_export_errors_path(@importer.id), class: 'btn btn-primary text-nowrap', data: { turbolinks: false }%>
+      <%= link_to 'Upload Corrected Entries', importer_upload_corrected_entries_path(@importer.id), class: 'btn btn-primary text-nowrap' if @importer.parser.is_a?(Bulkrax::CsvParser) %>
     <% end %>
   </div>
 </div>

data/app/views/bulkrax/shared/_entries_tab.html.erb

@@ -12,5 +12,5 @@
       </tr>
     </thead>
   </table>
-  <div id='importer-entry-classes' class='hidden'><%= [item.parser.entry_class.to_s, item.parser.collection_entry_class.to_s, item.parser.file_set_entry_class.to_s].compact.join('|') %></div>
+  <div id='importer-entry-classes' class='hidden d-none'><%= [item.parser.entry_class.to_s, item.parser.collection_entry_class.to_s, item.parser.file_set_entry_class.to_s].compact.join('|') %></div>
 </div>

data/config/locales/bulkrax.en.yml

@@ -1,9 +1,9 @@
 en:
-  helpers: 
-    action: 
+  helpers:
+    action:
       importer:
         new: "New"
-      exporter: 
+      exporter:
         new: "New"
   bulkrax:
     admin:
@@ -75,6 +75,8 @@ en:
         identifier: Identifier
         entry_id: Entry ID
         status: Status
+        type: Type
+        updated_at: Updated At
         errors: Errors
         status_set_at: Status Set At
         actions: Actions

data/lib/bulkrax/engine.rb

@@ -35,7 +35,7 @@ module Bulkrax
         ActionController::Base.view_paths = paths.uniq
 
         custom_query_strategies = {
-          find_by_model_and_property_value: :find_single_or_nil
+          find_by_property_value: :find_single_or_nil
         }
 
         if defined?(::Goddess::CustomQueryContainer)

data/lib/bulkrax/version.rb

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
 
 module Bulkrax
-  VERSION = '9.0.2'
+  VERSION = '9.2.0'
 end

data/lib/bulkrax.rb

@@ -107,11 +107,8 @@ module Bulkrax
       # Hyrax::FileSet.try(:internal_resource) || 'hi'
       # => #<Dry::Types::Result::Failure input=:internal_resource error=...
       # ```
-      if collection_model_class.respond_to?(:internal_resource)
-        collection_model_class.internal_resource
-      else
-        collection_model_class.to_s
-      end
+      instance = collection_model_class.new
+      instance.respond_to?(:internal_resource) ? instance.internal_resource : collection_model_class.to_s
     end
 
     def file_model_class
@@ -130,11 +127,8 @@ module Bulkrax
       # Hyrax::FileSet.try(:internal_resource) || 'hi'
       # => #<Dry::Types::Result::Failure input=:internal_resource error=...
       # ```
-      if file_model_class.respond_to?(:internal_resource)
-        file_model_class.internal_resource
-      else
-        file_model_class.to_s
-      end
+      instance = file_model_class.new
+      instance.respond_to?(:internal_resource) ? instance.internal_resource : file_model_class.to_s
     end
 
     def curation_concerns
@@ -154,7 +148,8 @@ module Bulkrax
         # Hyrax::FileSet.try(:internal_resource) || 'hi'
         # => #<Dry::Types::Result::Failure input=:internal_resource error=...
         # ```
-        cc.respond_to?(:internal_resource) ? cc.internal_resource : cc.to_s
+        instance = cc.new
+        instance.respond_to?(:internal_resource) ? instance.internal_resource : cc.to_s
       end.uniq
     end
 

data/lib/generators/bulkrax/templates/bin/importer

@@ -16,7 +16,7 @@ def main(opts = {})
   headers['Authorization'] = "Token: #{opts.delete(:auth_token)}"
   params = build_params(opts)
 
-  logger.info("POST to #{url} - PARAMS #{params}")
+  Rails.logger.info("POST to #{url} - PARAMS #{params}")
 
   conn = Faraday.new(
     url: url,
@@ -92,10 +92,6 @@ def build_url(importer_id, url, port = nil)
   return url
 end
 
-def logger
-  Rails.logger
-end
-
 def version
   puts "Bulkrax #{Bulkrax::VERSION}"
   puts "Slop #{Slop::VERSION}"

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: bulkrax
 version: !ruby/object:Gem::Version
-  version: 9.0.2
+  version: 9.2.0
 platform: ruby
 authors:
 - Rob Kaufman
 autorequire:
 bindir: bin
 cert_chain: []
-date: 2025-01-17 00:00:00.000000000 Z
+date: 2025-09-10 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: rails
@@ -17,6 +17,9 @@ dependencies:
     - - ">="
       - !ruby/object:Gem::Version
         version: 5.1.6
+    - - "<"
+      - !ruby/object:Gem::Version
+        version: 7.0.0
   type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
@@ -24,6 +27,9 @@ dependencies:
     - - ">="
       - !ruby/object:Gem::Version
         version: 5.1.6
+    - - "<"
+      - !ruby/object:Gem::Version
+        version: 7.0.0
 - !ruby/object:Gem::Dependency
   name: bagit
   requirement: !ruby/object:Gem::Requirement
@@ -345,7 +351,6 @@ files:
 - app/factories/bulkrax/object_factory.rb
 - app/factories/bulkrax/object_factory_interface.rb
 - app/factories/bulkrax/valkyrie_object_factory.rb
-- app/factories/bulkrax/valkyrize-hyku.code-workspace
 - app/helpers/bulkrax/application_helper.rb
 - app/helpers/bulkrax/exporters_helper.rb
 - app/helpers/bulkrax/importers_helper.rb

data/app/factories/bulkrax/valkyrize-hyku.code-workspace

@@ -1,19 +0,0 @@
-{
-  "folders": [
-    {
-      "path": "../../../../../hyku"
-    },
-    {
-      "path": "../../../../hyrax"
-    },
-    {
-      "path": "../../.."
-    },
-    {
-      "path": "../../../../iiif_print"
-    }
-  ],
-  "settings": {
-    "github.copilot.inlineSuggest.enable": true
-  }
-}