Export files attached to exported works (#207)

Updates the importer to export links to the files attached to each
exported work.

Files are exported in the same ordering they are attached to the work.

In Hyrax, the file we want is the `:original_file` attached
to each FileSet.  The Hyrax downloads controller abstracts this so we
can just provide the `:id` of the desired FileSet.

app/lib/tenejo/csv_exporter.rb

@@ -5,8 +5,8 @@
 
 module Tenejo
   class CsvExporter
-    EXCLUDE_FROM_EXPORT = [:date_modified, :identifier, :label, :arkivo_checksum, :state].freeze
-    HEADER_ROW = (([:primary_identifier, :error, :class, :parent, :title] \
+    EXCLUDE_FROM_EXPORT = [:date_modified, :label, :arkivo_checksum, :state].freeze
+    HEADER_ROW = (([:identifier, :error, :object_type, :visibility, :parent, :title, :file_url] \
                   + Tenejo::CsvImporter.collection_attributes_to_copy.keys \
                   + Tenejo::CsvImporter.work_attributes_to_copy.keys).uniq \
                   - EXCLUDE_FROM_EXPORT).freeze
@@ -28,14 +28,11 @@ def run
     def generate_csv
       csv_string = CSV.generate(encoding: 'UTF-8', write_headers: true) do |csv|
         csv << HEADER_ROW
-        csv << CSV::Row.new([:primary_identifier, :error], ["missing", "No identifiers provided"]) if @export.identifiers.empty?
+        csv << CSV::Row.new([:identifier, :error], ["missing", "No identifiers provided"]) if @export.identifiers.empty?
         @export.identifiers.each do |id|
           find_and_export(id, csv)
         end
       end
-
-      # TODO: remove this after refactoring Tenejo metadata to rename primary_identifier and identifer
-      csv_string.gsub!('primary_identifier,error,class,', 'identifier,error,object type,')
       csv_string
     end
 
@@ -53,7 +50,7 @@ def export_name
     # @param csv[CSV] a CSV IO object to append the metadata to
     def find_and_export(id, csv)
       obj = ActiveFedora::Base.where(primary_identifier_ssi: id).last
-      csv << CSV::Row.new([:primary_identifier, :error], [id, "No match for identifier"]) unless obj
+      csv << CSV::Row.new([:identifier, :error], [id, "No match for identifier"]) unless obj
       serialize_with_descendants(obj, nil, csv)
     end
 
@@ -72,13 +69,23 @@ def serialize_children(csv, obj)
       parent_id = obj.primary_identifier
       obj.try(:child_collections)&.map { |child| serialize_with_descendants(child, parent_id, csv) }
       obj.try(:child_works)&.map { |child| serialize_with_descendants(child, parent_id, csv) }
+      obj.try(:ordered_file_sets)&.map { |child| serialize_with_descendants(child, parent_id, csv) }
+    end
+
+    def download_url(obj)
+      return unless obj.is_a? FileSet
+      Hyrax::Engine.routes.url_helpers.download_url(obj.id)
     end
 
     def serialize(obj, parent_id = nil)
       return unless obj
-      obj.define_singleton_method(:parent) { parent_id }
       values = HEADER_ROW.map { |attr| pack_field(obj.try(attr)) }
-      CSV::Row.new(HEADER_ROW, values)
+      row = CSV::Row.new(HEADER_ROW, values)
+      row[:parent] = parent_id
+      row[:identifier] = obj.primary_identifier || obj.id
+      row[:file_url] = download_url(obj)
+      row[:object_type] = obj.class.to_s.gsub('FileSet', 'File')
+      row
     end
 
     # Handle multi-value fields regardless of underlying class

config/initializers/hyrax.rb

@@ -309,6 +309,8 @@
 require './app/lib/hydra/file_characterization/characterizers/fits_servlet'
 Date::DATE_FORMATS[:standard] = "%m/%d/%Y"
 
+Hyrax::Engine.routes.default_url_options = Rails.application.routes.default_url_options
+
 Qa::Authorities::Local.register_subauthority('subjects', 'Qa::Authorities::Local::TableBasedAuthority')
 Qa::Authorities::Local.register_subauthority('languages', 'Qa::Authorities::Local::TableBasedAuthority')
 Qa::Authorities::Local.register_subauthority('genres', 'Qa::Authorities::Local::TableBasedAuthority')

spec/lib/tenejo/csv_exporter_spec.rb

@@ -64,13 +64,13 @@
       expect(rows[0]['identifier']).to eq 'COL001'
       expect(rows[0]['error']).to be_nil
       expect(rows[0]['title']).to include 'Test collection'
-      expect(rows[0]['object type']).to eq "Collection"
+      expect(rows[0]['object_type']).to eq "Collection"
 
       # Work WRK001
       expect(rows[1]['identifier']).to eq 'WRK001'
       expect(rows[1]['error']).to be_nil
       expect(rows[1]['title']).to include 'Test work'
-      expect(rows[1]['object type']).to eq "Work"
+      expect(rows[1]['object_type']).to eq "Work"
     end
 
     it 'includes children', :aggregate_failures do
@@ -84,10 +84,34 @@
 
       expect(rows[0]['parent']).to be_blank
       expect(rows[0]['identifier']).to eq 'COL001'
+      expect(rows[0]['object_type']).to eq 'Collection'
       expect(rows[1]['parent']).to eq 'COL001'
       expect(rows[1]['identifier']).to eq 'WRK001'
+      expect(rows[1]['object_type']).to eq 'Work'
       expect(rows[2]['parent']).to eq 'WRK001'
       expect(rows[2]['identifier']).to eq 'WRK002'
+      expect(rows[2]['object_type']).to eq 'Work'
+    end
+
+    it 'includes files', :aggregate_failures do
+      file001 = FileSet.new(primary_identifier: 'FIL001', id: 'placeholder')
+      file002 = FileSet.new(id: 'auto-generated')
+      allow(ActiveFedora::Base).to receive(:where).and_return([work001])
+      allow(work001).to receive(:ordered_file_sets).and_return([file001, file002])
+
+      export.identifiers = ['WRK001']
+      csv_string = described_class.new(export).generate_csv
+      rows = CSV.parse(csv_string, headers: true)
+
+      expect(rows[0]['object_type']).to eq 'Work'
+      expect(rows[0]['identifier']).to eq 'WRK001'
+      expect(rows[0]['file_url']).to be_blank
+      expect(rows[1]['object_type']).to eq 'File'
+      expect(rows[1]['identifier']).to eq 'FIL001'
+      expect(rows[1]['file_url']).to eq 'http://localhost:3000/downloads/placeholder'
+      expect(rows[2]['object_type']).to eq 'File'
+      expect(rows[2]['identifier']).to eq 'auto-generated'
+      expect(rows[2]['file_url']).to eq 'http://localhost:3000/downloads/auto-generated'
     end
   end
 
@@ -131,6 +155,10 @@
       expect(serialized[:parent]).to eq 'PARENT_ID'
     end
 
+    it "returns the row type" do
+      expect(serialized[:object_type]).to eq 'Work'
+    end
+
     it "handles all the fields" do
       # Just check for one value deep in the array
       # Otherwise we need to check the whole exact string and the test becomes fragile
@@ -148,5 +176,9 @@
     it "handles embedded commas" do
       expect(serialized[:creator]).to eq 'Anon., 16th Century'
     end
+
+    it "defaults to private(restricted) visibility" do
+      expect(serialized[:visibility]).to eq 'restricted'
+    end
   end
 end

spec/rails_helper.rb

@@ -11,6 +11,7 @@
 require_relative 'support/controller_macros'
 # Add additional requires below this line. Rails is not loaded until this point!
 require 'capybara/rails'
+require 'noid/rails/rspec'
 
 # use this in specs to avoid actually using a working virus scanner during tests (very slow)
 Hyrax.config.virus_scanner = Hyrax::VirusScanner
@@ -46,6 +47,16 @@
   config.include Devise::Test::IntegrationHelpers, type: :system
   config.include Devise::Test::IntegrationHelpers, type: :request
 
+  # Avoid rollback of id minter-state to avoid errors like
+  # `Ldp::Conflict, "Can't call create on an existing resource"`
+  # see testing notes at https://github.com/samvera/noid-rails#overriding-default-behavior
+  include Noid::Rails::RSpec
+  config.before(:suite) { disable_production_minter! }
+  config.after(:suite)  { enable_production_minter! }
+
+  # Clean out ActiveFedora
+  config.before(:suite) { ActiveFedora::Cleaner.clean! }
+
   config.before do |_example|
     class_double(Tenejo::VirusScanner)
     allow(Tenejo::VirusScanner).to receive(:infected?).and_return(false)