🐛 Fix CSV file path handling in zip imports

When re running a single entry, a nil file path was getting sent to #read_data and thus resulted in the StandardError.

The issue is in real_import_file_path where we're returning too early when file? && zip? is true but no CSV files are found in the unzipped directory.

app/parsers/bulkrax/csv_parser.rb

@@ -379,7 +379,10 @@ def unique_collection_identifier(collection_hash)
     # 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"].reject { |path| in_files_dir?(path) }.first if file? && zip?
+      if file? && zip?
+        csv_files = Dir["#{importer_unzip_path}/**/*.csv"].reject { |path| in_files_dir?(path) }
+        return csv_files.first if csv_files.any?
+      end
 
       parser_fields['import_file_path']
     end