Merge pull request #944 from OregonDigital/migration_fix_task

Migration fix task

lib/tasks/migration/fixes.rake

@@ -0,0 +1,56 @@
+# frozen_string_literal: true
+
+namespace :migration do
+  namespace :fixes do
+    desc 'Fix batch by retrying file attach'
+    task retry_file_attach: :environment do
+      batch = ENV['batch']
+
+      batch_path = File.join(Hyrax::Migrator.config.ingest_local_path, batch)
+
+      pids = Dir.entries(batch_path)
+                .select { |e| File.file?(File.join(batch_path, e)) && File.extname(e) == '.zip' }
+                .map { |zip_file| File.basename(zip_file, File.extname(zip_file)) }
+
+      pids.each do |pid|
+        assess(pid)
+      end
+    end
+
+    desc 'Fix one asset by retrying file attach'
+    task retry_single_file_attach: :environment do
+      pid = ENV['pid']
+      assess(pid)
+    end
+  end
+end
+
+def assess(pid)
+  af_object = ActiveFedora::Base.exists?(pid) ? ActiveFedora::Base.find(pid) : nil
+  attach_file(af_object) if !af_object.nil? && af_object.representative_id.nil?
+end
+
+# Looks for the content file in the file_system_path
+# TODO Enable retrieving content file from the zip?
+def attach_file(af_object)
+  filename = files.find { |f| f.downcase.start_with? af_object.id }
+  puts "Unable to find content file for #{af_object.id}" && return if filename.nil?
+
+  puts "Enqueueing job for #{af_object.id}"
+  AttachFilesToWorkJob.perform_later(af_object, [upload_file(migration_user, filename)])
+end
+
+def migration_user
+  @migration_user ||= User.where(email: Hyrax::Migrator.config.migration_user).first
+end
+
+def files
+  @files ||= Dir.entries(Hyrax::Migrator.config.file_system_path)
+end
+
+def upload_file(migration_user, filename)
+  file = File.open(File.join(Hyrax::Migrator.config.file_system_path, filename))
+  uploaded_file = Hyrax::UploadedFile.new(user: migration_user, file: file)
+  uploaded_file.save
+  uploaded_file
+end

spec/tasks/migration/rake_spec.rb

@@ -5,6 +5,7 @@
 require 'pry'
 
 RSpec.describe 'Rake tasks' do
+  include ActiveJob::TestHelper
   describe 'migration:bulk_approve' do
     let(:work1) { build(:work, user: other_user, state: state) }
     let(:migration_user) { User.new(email: '[email protected]') }
@@ -71,4 +72,41 @@
       end
     end
   end
+
+  describe 'migration:fixes:retry_file_attach' do
+    context 'when there is a batch of persisted assets with no files' do
+      let(:migration_user) { User.new(email: '[email protected]') }
+      let(:work1) { build(:work, user: migration_user, id: 'abcde1234') }
+      let(:work2) { build(:work, user: migration_user, id: 'abcde5678') }
+      let(:path) { Rails.root.join 'spec/fixtures' }
+      let(:upload_file) { Hyrax::UploadedFile.new(user: migration_user, file: file) }
+      let(:file) { File.open(File.join(path, 'abcde1234_content.txt')) }
+      let(:run_rake_task) do
+        ENV['batch'] = 'batch1'
+        Rake.application.invoke_task 'migration:fixes:retry_file_attach'
+      end
+
+      before do
+        work1.save
+        work2.save
+        migration_user.save
+        upload_file.save
+        Hyrax::Migrator.config.ingest_local_path = path
+        Hyrax::Migrator.config.file_system_path = path
+        Rake.application.rake_require 'tasks/migration/fixes'
+        Rake::Task.define_task(:environment)
+        allow(Hyrax::UploadedFile).to receive(:new).and_return(upload_file)
+        run_rake_task
+      end
+
+      after do
+        clear_enqueued_jobs
+        clear_performed_jobs
+      end
+
+      it 'enqueues jobs' do
+        expect(ActiveJob::Base.queue_adapter.enqueued_jobs.size).to eq 2
+      end
+    end
+  end
 end