Split downloads by size (#582)

* can break up bagit zips by a hardcoded record amount

* breaking bags into folders appropriately

* splitting bagit zips by size

* missed an "end"

* zip using only the new method

* add rerun_all_exporters task

* move a puts

* multiple zips show up on the exporter index page

* wip: new download button

* able to download zips from index page again!
co-author: shana@scientist.com

* able to download zips from the exporter show page
co-author: shana@scientist.com

* splitting download sizes for csv exports too

* style the download buttons
co-author: shana@scientist.com

* lint fix

* more linting

* specs pass

* fix extra count issue

* sort download options in ui

* handle works with more than one file_set, move 1000 to a method

* always find or create the file set entry on export

* sort the exported zip files properly

* fixing specs

* add bagit parser spec for #find_child_file_sets

* 100% coverage for #find_child_file_sets

* remove byebug

* add spec for #records_split_count

* create exporter specs

* create spec for #setup_export_file

* add bagit parser specs for new folder system

* consistent use of "subject" instead of "parser" in the bagit parser specs

* all specs passing

* add a spec for #write_files

* linting

* refactor #find_child_file_sets

Co-authored-by: kirkkwang <k3wang@gmail.com>

app/controllers/bulkrax/exporters_controller.rb

@@ -127,7 +127,7 @@ def add_exporter_breadcrumbs
     # Download methods
 
     def file_path
-      @exporter.exporter_export_zip_path
+      "#{@exporter.exporter_export_zip_path}/#{params['exporter']['exporter_export_zip_files']}"
     end
   end
 end

app/models/bulkrax/exporter.rb

@@ -124,9 +124,13 @@ def exporter_export_path
     end
 
     def exporter_export_zip_path
-      @exporter_export_zip_path ||= File.join(parser.base_path('export'), "export_#{self.id}_#{self.exporter_runs.last.id}.zip")
+      @exporter_export_zip_path ||= File.join(parser.base_path('export'), "export_#{self.id}_#{self.exporter_runs.last.id}")
     rescue
-      @exporter_export_zip_path ||= File.join(parser.base_path('export'), "export_#{self.id}_0.zip")
+      @exporter_export_zip_path ||= File.join(parser.base_path('export'), "export_#{self.id}_0")
+    end
+
+    def exporter_export_zip_files
+      @exporter_export_zip_files ||= Dir["#{exporter_export_zip_path}/**"].map { |zip| Array(zip.split('/').last) }
     end
 
     def export_properties
@@ -137,5 +141,14 @@ def export_properties
     def metadata_only?
       export_type == 'metadata'
     end
+
+    def sort_zip_files(zip_files)
+      zip_files.sort_by do |item|
+        number = item.split('_').last.match(/\d+/)&.[](0) || 0.to_s
+        sort_number = number.rjust(4, "0")
+
+        sort_number
+      end
+    end
   end
 end

app/models/concerns/bulkrax/export_behavior.rb

@@ -7,9 +7,6 @@ module ExportBehavior
 
     def build_for_exporter
       build_export_metadata
-      # TODO(alishaevn): determine if the line below is still necessary
-      # the csv and bagit parsers also have write_files methods
-      write_files if export_type == 'full' && !importerexporter.parser_klass.include?('Bagit')
     rescue RSolr::Error::Http, CollectionsCreatedError => e
       raise e
     rescue StandardError => e
@@ -26,25 +23,6 @@ def hyrax_record
       @hyrax_record ||= ActiveFedora::Base.find(self.identifier)
     end
 
-    def write_files
-      return if hyrax_record.is_a?(Collection)
-
-      file_sets = hyrax_record.file_set? ? Array.wrap(hyrax_record) : hyrax_record.file_sets
-      file_sets << hyrax_record.thumbnail if hyrax_record.thumbnail.present? && hyrax_record.work? && exporter.include_thumbnails
-      file_sets.each do |fs|
-        path = File.join(exporter_export_path, 'files')
-        FileUtils.mkdir_p(path)
-        file = filename(fs)
-        require 'open-uri'
-        io = open(fs.original_file.uri)
-        next if file.blank?
-        File.open(File.join(path, file), 'wb') do |f|
-          f.write(io.read)
-          f.close
-        end
-      end
-    end
-
     # Prepend the file_set id to ensure a unique filename and also one that is not longer than 255 characters
     def filename(file_set)
       return if file_set.original_file.blank?

app/parsers/bulkrax/application_parser.rb

@@ -260,10 +260,14 @@ def unzip(file_to_unzip)
     end
 
     def zip
-      FileUtils.rm_rf(exporter_export_zip_path)
-      Zip::File.open(exporter_export_zip_path, create: true) do |zip_file|
-        Dir["#{exporter_export_path}/**/**"].each do |file|
-          zip_file.add(file.sub("#{exporter_export_path}/", ''), file)
+      FileUtils.mkdir_p(exporter_export_zip_path)
+
+      Dir["#{exporter_export_path}/**"].each do |folder|
+        zip_path = "#{exporter_export_zip_path.split('/').last}_#{folder.split('/').last}.zip"
+        Zip::File.open(File.join("#{exporter_export_zip_path}/#{zip_path}"), create: true) do |zip_file|
+          Dir["#{folder}/**/**"].each do |file|
+            zip_file.add(file.sub("#{folder}/", ''), file)
+          end
         end
       end
     end

app/parsers/bulkrax/bagit_parser.rb

@@ -113,6 +113,9 @@ def current_record_ids
       when 'importer'
         set_ids_for_exporting_from_importer
       end
+
+      find_child_file_sets(@work_ids) if importerexporter.export_from == 'collection' || importerexporter.export_from == 'worktype'
+
       @work_ids + @collection_ids + @file_set_ids
     end
 
@@ -122,18 +125,27 @@ def current_record_ids
     def write_files
       require 'open-uri'
       require 'socket'
+
+      folder_count = 1
+      records_in_folder = 0
+
       importerexporter.entries.where(identifier: current_record_ids)[0..limit || total].each do |entry|
         record = ActiveFedora::Base.find(entry.identifier)
         next unless Hyrax.config.curation_concerns.include?(record.class)
-        bag = BagIt::Bag.new setup_bagit_folder(entry.identifier)
+
         bag_entries = [entry]
+        file_set_entries = Bulkrax::CsvFileSetEntry.where(importerexporter_id: importerexporter.id).where("parsed_metadata LIKE '%#{record.id}%'")
+        file_set_entries.each { |fse| bag_entries << fse }
 
-        record.file_sets.each do |fs|
-          if @file_set_ids.present?
-            file_set_entry = Bulkrax::CsvFileSetEntry.where("parsed_metadata LIKE '%#{fs.id}%'").first
-            bag_entries << file_set_entry unless file_set_entry.nil?
-          end
+        records_in_folder += bag_entries.count
+        if records_in_folder > records_split_count
+          folder_count += 1
+          records_in_folder = bag_entries.count
+        end
+
+        bag ||= BagIt::Bag.new setup_bagit_folder(folder_count, entry.identifier)
 
+        record.file_sets.each do |fs|
           file_name = filename(fs)
           next if file_name.blank?
           io = open(fs.original_file.uri)
@@ -148,17 +160,21 @@ def write_files
           end
         end
 
-        CSV.open(setup_csv_metadata_export_file(entry.identifier), "w", headers: export_headers, write_headers: true) do |csv|
+        CSV.open(setup_csv_metadata_export_file(folder_count, entry.identifier), "w", headers: export_headers, write_headers: true) do |csv|
           bag_entries.each { |csv_entry| csv << csv_entry.parsed_metadata }
         end
-        write_triples(entry)
+
+        write_triples(folder_count, entry)
         bag.manifest!(algo: 'sha256')
       end
     end
     # rubocop:enable Metrics/MethodLength, Metrics/AbcSize
 
-    def setup_csv_metadata_export_file(id)
-      File.join(importerexporter.exporter_export_path, id, 'metadata.csv')
+    def setup_csv_metadata_export_file(folder_count, id)
+      path = File.join(importerexporter.exporter_export_path, folder_count.to_s)
+      FileUtils.mkdir_p(path) unless File.exist?(path)
+
+      File.join(path, id, 'metadata.csv')
     end
 
     def key_allowed(key)
@@ -167,21 +183,27 @@ def key_allowed(key)
         key != source_identifier.to_s
     end
 
-    def setup_triple_metadata_export_file(id)
-      File.join(importerexporter.exporter_export_path, id, 'metadata.nt')
+    def setup_triple_metadata_export_file(folder_count, id)
+      path = File.join(importerexporter.exporter_export_path, folder_count.to_s)
+      FileUtils.mkdir_p(path) unless File.exist?(path)
+
+      File.join(path, id, 'metadata.nt')
     end
 
-    def setup_bagit_folder(id)
-      File.join(importerexporter.exporter_export_path, id)
+    def setup_bagit_folder(folder_count, id)
+      path = File.join(importerexporter.exporter_export_path, folder_count.to_s)
+      FileUtils.mkdir_p(path) unless File.exist?(path)
+
+      File.join(path, id)
     end
 
-    def write_triples(e)
+    def write_triples(folder_count, e)
       sd = SolrDocument.find(e.identifier)
       return if sd.nil?
 
       req = ActionDispatch::Request.new({ 'HTTP_HOST' => Socket.gethostname })
       rdf = Hyrax::GraphExporter.new(sd, req).fetch.dump(:ntriples)
-      File.open(setup_triple_metadata_export_file(e.identifier), "w") do |triples|
+      File.open(setup_triple_metadata_export_file(folder_count, e.identifier), "w") do |triples|
         triples.write(rdf)
       end
     end

app/parsers/bulkrax/csv_parser.rb

@@ -4,6 +4,7 @@
 module Bulkrax
   class CsvParser < ApplicationParser # rubocop:disable Metrics/ClassLength
     include ErroredEntries
+    include ExportBehavior
     attr_writer :collections, :file_sets, :works
 
     def self.export_supported?
@@ -207,6 +208,13 @@ def current_record_ids
       @work_ids + @collection_ids + @file_set_ids
     end
 
+    # find the related file set ids so entries can be made for export
+    def find_child_file_sets(work_ids)
+      work_ids.each do |id|
+        ActiveFedora::Base.find(id).file_set_ids.each { |fs_id| @file_set_ids << fs_id }
+      end
+    end
+
     # Set the following instance variables: @work_ids, @collection_ids, @file_set_ids
     # @see #current_record_ids
     def set_ids_for_exporting_from_importer
@@ -283,6 +291,10 @@ def total
       @total = 0
     end
 
+    def records_split_count
+      1000
+    end
+
     # @todo - investigate getting directory structure
     # @todo - investigate using perform_later, and having the importer check for
     #   DownloadCloudFileJob before it starts
@@ -307,9 +319,37 @@ def retrieve_cloud_files(files)
     # export methods
 
     def write_files
-      CSV.open(setup_export_file, "w", headers: export_headers, write_headers: true) do |csv|
-        importerexporter.entries.where(identifier: current_record_ids)[0..limit || total].each do |e|
-          csv << e.parsed_metadata
+      require 'open-uri'
+      folder_count = 0
+
+      importerexporter.entries.where(identifier: current_record_ids)[0..limit || total].in_groups_of(records_split_count, false) do |group|
+        folder_count += 1
+
+        CSV.open(setup_export_file(folder_count), "w", headers: export_headers, write_headers: true) do |csv|
+          group.each do |entry|
+            csv << entry.parsed_metadata
+            next if importerexporter.metadata_only? || entry.type == 'Bulkrax::CsvCollectionEntry'
+
+            store_files(entry.identifier, folder_count.to_s)
+          end
+        end
+      end
+    end
+
+    def store_files(identifier, folder_count)
+      record = ActiveFedora::Base.find(identifier)
+      file_sets = record.file_set? ? Array.wrap(record) : record.file_sets
+      file_sets << record.thumbnail if exporter.include_thumbnails && record.thumbnail.present? && record.work?
+      file_sets.each do |fs|
+        path = File.join(exporter_export_path, folder_count, 'files')
+        FileUtils.mkdir_p(path) unless File.exist? path
+        file = filename(fs)
+        io = open(fs.original_file.uri)
+        next if file.blank?
+
+        File.open(File.join(path, file), 'wb') do |f|
+          f.write(io.read)
+          f.close
         end
       end
     end
@@ -356,8 +396,11 @@ def sort_headers(headers)
     end
 
     # in the parser as it is specific to the format
-    def setup_export_file
-      File.join(importerexporter.exporter_export_path, "export_#{importerexporter.export_source}_from_#{importerexporter.export_from}.csv")
+    def setup_export_file(folder_count)
+      path = File.join(importerexporter.exporter_export_path, folder_count.to_s)
+      FileUtils.mkdir_p(path) unless File.exist?(path)
+
+      File.join(path, "export_#{importerexporter.export_source}_from_#{importerexporter.export_from}_#{folder_count}.csv")
     end
 
     # Retrieve file paths for [:file] mapping in records

app/views/bulkrax/exporters/_downloads.html.erb

@@ -0,0 +1,8 @@
+<%= form.select :exporter_export_zip_files,
+  exporter.sort_zip_files(form.object.exporter_export_zip_files.flatten),
+  {},
+  {
+    class: 'btn btn-default form-control',
+    style: 'width: 200px'
+  }
+%>

app/views/bulkrax/exporters/index.html.erb

@@ -21,7 +21,7 @@
               <th scope="col">Name</th>
               <th scope="col">Status</th>
               <th scope="col">Date Exported</th>
-              <th scope="col"></th>
+              <th scope="col">Downloadable Files</th>
               <th scope="col"></th>
               <th scope="col"></th>
               <th scope="col"></th>
@@ -35,7 +35,10 @@
                 <td><%= exporter.created_at %></td>
                 <td>
                   <% if File.exist?(exporter.exporter_export_zip_path) %>
-                    <%= link_to raw('<span class="glyphicon glyphicon-download"></span>'), exporter_download_path(exporter) %>
+                    <%= simple_form_for(exporter, method: :get, url: exporter_download_path(exporter)) do |form| %>
+                      <%= render 'downloads', exporter: exporter, form: form %>
+                      <%= form.button :submit, value: 'Download', data: { disable_with: false } %>
+                    <% end %>
                   <% end%>
                 </td>
                 <td><%= link_to raw('<span class="glyphicon glyphicon-info-sign"></span>'), exporter_path(exporter) %></td>

app/views/bulkrax/exporters/show.html.erb

@@ -8,10 +8,11 @@
   <div class='panel-body'>
 
     <% if File.exist?(@exporter.exporter_export_zip_path) %>
-      <p class='bulkrax-p-align'>
+      <%= simple_form_for @exporter, method: :get, url: exporter_download_path(@exporter), html: { class: 'form-inline bulkrax-p-align' } do |form| %>
         <strong>Download:</strong>
-        <%= link_to raw('<span class="glyphicon glyphicon-download"></span>'), exporter_download_path(@exporter) %>
-      </p>
+        <%= render 'downloads', exporter: @exporter, form: form %>
+        <%= form.button :submit, value: 'Download', data: { disable_with: false } %>
+      <% end %>
     <% end %>
 
     <p class='bulkrax-p-align'>
@@ -135,10 +136,6 @@
       <%= page_entries_info(@work_entries) %><br>
       <%= paginate(@work_entries, param_name: :work_entries_page) %>
       <br>
-      <% if File.exist?(@exporter.exporter_export_zip_path) %>
-        <%= link_to 'Download', exporter_download_path(@exporter) %>
-        |
-      <% end %>
       <%= link_to 'Edit', edit_exporter_path(@exporter) %>
       |
       <%= link_to 'Back', exporters_path %>

lib/tasks/bulkrax_tasks.rake

@@ -1,6 +1,30 @@
 # frozen_string_literal: true
 
-# desc "Explaining what the task does"
-# task :bulkrax do
-#   # Task goes here
-# end
+namespace :bulkrax do
+  desc "Remove old exported zips and create new ones with the new file structure"
+  task rerun_all_exporters: :environment do
+    if defined?(::Hyku)
+      Account.find_each do |account|
+        puts "=============== updating #{account.name} ============"
+        next if account.name == "search"
+        switch!(account)
+
+        rerun_exporters_and_delete_zips
+
+        puts "=============== finished updating #{account.name} ============"
+      end
+    else
+      rerun_exporters_and_delete_zips
+    end
+  end
+
+  def rerun_exporters_and_delete_zips
+    begin
+      Bulkrax::Exporter.all.each { |e| Bulkrax::ExporterJob.perform_later(e.id) }
+    rescue => e
+      puts "(#{e.message})"
+    end
+
+    Dir["tmp/exports/**.zip"].each { |zip_path| FileUtils.rm_rf(zip_path) }
+  end
+end

spec/jobs/bulkrax/importer_job_spec.rb

@@ -30,7 +30,7 @@ module Bulkrax
         importer_job.perform(importer.id)
 
         expect(importer.current_run.total_work_entries).to eq(10)
-        expect(importer.current_run.total_collection_entries).to eq(427)
+        expect(importer.current_run.total_collection_entries).to eq(428)
       end
     end