☑️ Verifying pdf splitter finds pre-existing files

Updating a bit of documentation and reworking the filename to account
for a work having multiple PDFs.

- https://github.com/scientist-softserv/adventist-dl/issues/330
- scientist-softserv/iiif_print#220

lib/derivative_rodeo/generators/pdf_split_generator.rb

@@ -30,9 +30,7 @@ class PdfSplitGenerator < BaseGenerator
       #
       # @see #existing_page_locations
       def image_file_basename_template(basename:)
-        # We can do this because the temp files are always local; and we'll need to modify how we
-        # write these files.
-        "pages/#{basename}-%d.#{output_extension}"
+        "#{basename}/pages/#{basename}-%d.#{output_extension}"
       end
 
       ##
@@ -45,10 +43,12 @@ def image_file_basename_template(basename:)
       #         with :tail_glob.
       #
       # @note There is relation to {Generators::BaseGenerator#destination} and this method.
+      #
+      # @note The tail_glob is in relation to the {#image_file_basename_template}
       def existing_page_locations(input_location:)
-        # TODO: Are we adequately accounting for the directory structure necessary to have a work have
-        # more than one PDF and then split each PDF's pages into the correct sub directory?
-        tail_glob = "pages/*.#{output_extension}"
+        # See image_file_basename_template
+        tail_glob = "#{input_location.file_basename}/pages/*.#{output_extension}"
+
         output_locations = input_location.derived_file_from(template: output_location_template).globbed_tail_locations(tail_glob: tail_glob)
         return output_locations if output_locations.count.positive?
 
@@ -69,7 +69,12 @@ def existing_page_locations(input_location:)
       # @yieldparam image_location [StorageLocations::FileLocation] the file and adapter logic.
       # @yieldparam image_path [String] where to find this file in the tmp space
       #
+      # @note This function makes a concession; namely that if it encounters any
+      # {#existing_page_locations} it will use all of that result as the entire number of pages.
+      # We could make this smarter but at the moment we're deferring on that.
+      #
       # @see BaseGenerator#with_each_requisite_location_and_tmp_file_path for further discussion
+      #
       # rubocop:disable Metrics/MethodLength
       def with_each_requisite_location_and_tmp_file_path
         input_files.each do |input_location|

spec/derivative_rodeo/generators/pdf_split_generator_spec.rb

@@ -17,6 +17,33 @@
   end
 
   describe '#generated_files' do
+    context 'when given an already split PDF' do
+      it 'uses the already split components' do
+        Fixtures.with_file_uris_for("minimal-2-page.pdf") do |input_uris|
+          Fixtures.with_temporary_directory do |output_temporary_path|
+            output_location_template = "file://#{output_temporary_path}/{{ dir_parts[-1..-1] }}/{{ filename }}"
+            instance = described_class.new(input_uris: input_uris, output_location_template: output_location_template)
+            output_location = DerivativeRodeo::StorageLocations::FileLocation.build(from_uri: input_uris.first, template: output_location_template)
+
+            # Let's fake a nice TIFF being in a pre-processed location.
+            pre_existing_tiff_path = File.join(output_location.file_dir, output_location.file_basename, "pages/1.tiff")
+            FileUtils.mkdir_p(File.dirname(pre_existing_tiff_path))
+            File.open(pre_existing_tiff_path, "w+") do |f|
+              f.puts "🤠🐮🐴 A muppet man parading as a TIFF."
+            end
+
+            generated_files = instance.generated_files
+            # TODO: The PDF is two pages yet we only check for the presence of one
+            # or more derived files; hence our faked pre-processed derivative is all that we find.
+            expect(generated_files.size).to eq(1)
+
+            # Ensuring that we do in fact have the pre-made file.
+            expect(File.read(generated_files.first.file_path)).to start_with("🤠🐮🐴")
+          end
+        end
+      end
+    end
+
     context 'when given a PDF to split' do
       it 'will create one image per page, writing that to the storage adapter, and then enqueue each page for processing' do
         generated_files = nil

spec/derivative_rodeo/storage_locations/s3_location_spec.rb

@@ -66,11 +66,9 @@
 
   describe '#globbed_tail_locations' do
     it 'searched the bucket' do
-      # The subject's bucket is not the same as the above bucket
-      subject.bucket = bucket
       basename_ish = short_path.split(".").first
       key = File.join(basename_ish, File.basename(__FILE__))
-      bucket.object(key).upload_file(__FILE__)
+      subject.bucket.object(key).upload_file(__FILE__)
 
       subject.globbed_tail_locations(tail_glob: "*.rb")
     end