action.rb in documentcloud-cloud-crowd-0.1.0

- old
+ new

@@ -26,11 +26,10 @@
     def initialize(status, input, options, store)
       @input, @options, @store = input, options, store
       @job_id, @work_unit_id = options['job_id'], options['work_unit_id']
       @work_directory = File.expand_path(File.join(@store.temp_storage_path, storage_prefix))
       FileUtils.mkdir_p(@work_directory) unless File.exists?(@work_directory)
-      Dir.chdir @work_directory
       status == MERGING ? parse_input : download_input
     end
     
     # Each Action subclass must implement a +process+ method, overriding this.
     def process
@@ -51,27 +50,25 @@
     # Takes a local filesystem path, saves the file to S3, and returns the 
     # public (or authenticated) url on S3 where the file can be accessed. 
     def save(file_path)
       save_path = File.join(storage_prefix, File.basename(file_path))
       @store.save(file_path, save_path)
-      return @store.url(save_path)
     end
     
     # After the Action has finished, we remove the work directory and return
     # to the root directory (where daemons run by default).
     def cleanup_work_directory
-      Dir.chdir '/'
       FileUtils.rm_r(@work_directory) if File.exists?(@work_directory)
     end
     
     
     private
     
     # Convert an unsafe URL into a filesystem-friendly filename.
     def safe_filename(url)
-      ext = File.extname(url)
-      name = File.basename(url).gsub(/%\d+/, '-').gsub(/[^a-zA-Z0-9_\-.]/, '')
+      ext  = File.extname(url)
+      name = URI.unescape(File.basename(url)).gsub(/[^a-zA-Z0-9_\-.]/, '-').gsub(/-+/, '-')
       File.basename(name, ext).gsub('.', '-') + ext
     end
     
     # The directory prefix to use for both local and S3 storage.
     # [action_name]/job_[job_id]/unit_[work_unit_it]
@@ -88,14 +85,16 @@
       @input = JSON.parse(@input)
     end
     
     # If the input is a URL, download the file before beginning processing.
     def download_input
-      input_is_url = !!URI.parse(@input) rescue false
-      return unless input_is_url
-      @input_path = File.join(@work_directory, safe_filename(@input))
-      @file_name = File.basename(@input_path, File.extname(@input_path))
-      download(@input, @input_path)
+      Dir.chdir(@work_directory) do
+        input_is_url = !!URI.parse(@input) rescue false
+        return unless input_is_url
+        @input_path = File.join(@work_directory, safe_filename(@input))
+        @file_name = File.basename(@input_path, File.extname(@input_path))
+        download(@input, @input_path)
+      end
     end
     
   end
   
 end
\ No newline at end of file