lib/massive_sitemap/writer/file.rb in massive_sitemap-2.0.0.rc5 vs lib/massive_sitemap/writer/file.rb in massive_sitemap-2.0.0.rc6

@@ -15,11 +15,11 @@
         :filename        => "sitemap.xml",
       )
 
       def initialize(options = {})
         super
-        @files = Dir[::File.join(@options[:root], "*#{::File.extname(filename)}")]
+        init_stream_ids
       end
 
       protected
       def open_stream
         #create dir if not exists
@@ -33,29 +33,75 @@
       def close_stream(file)
         file.close
         # Move from tmp_file into acutal file
         ::File.delete(filename) if ::File.exists?(filename)
         ::File.rename(tmp_filename, filename)
-        @files << filename
+        add_stream_id(filename)
         rotate
       end
 
       def init?
-        if !@options[:force_overwrite] && @files.include?(filename)
+        if !@options[:force_overwrite] && find_stream_id(filename)
           error_message = "Can not create file: #{filename} exits"
           rotate #push next possible filename
           raise FileExistsException, error_message
         end
         true
       end
 
-      def stream_ids
-        @files.map do |path|
-          [::File.basename(path), ::File.stat(path).mtime]
+      # Keep state of Files
+      def init_stream_ids
+        @stream_ids = {}
+        load_stream_ids
+
+        if @options[:partial_update]
+          delete_stream_ids upper_stream_ids(@stream_ids.keys)
+        end
+
+        if (keep_file_for = @options[:keep_file_for].to_i) > 0
+          delete_stream_ids chaos_monkey_stream_ids(@stream_ids.keys.sort, keep_file_for)
+        end
+      end
+
+      def load_stream_ids
+        Dir[::File.join(@options[:root], "*#{::File.extname(filename)}")].each do |path|
+          add_stream_id(path, ::File.stat(path).mtime)
+        end
+      end
+
+      def upper_stream_ids(stream_id_keys)
+        {}.tap do |cluster|
+          stream_id_keys.each do |path|
+            filename, rotation,  ext = split_filename(path)
+            _,        rotation2, _   = split_filename(cluster[filename])
+            if rotation.to_i >= rotation2.to_i
+              cluster[filename] = path
+            end
+          end
+        end.values
+      end
+
+      def chaos_monkey_stream_ids(stream_id_keys, days)
+        return [] if days < 1
+        offset = Time.now.to_i / (24 * 60 * 60)
+        (0...stream_id_keys.size).step(days).map do |index|
+          stream_id_keys.at((offset % days) + index)
         end.compact
       end
 
+      def delete_stream_ids(to_delete)
+        @stream_ids.delete_if { |key, value| to_delete.include?(key) }
+      end
+
+      def find_stream_id(path)
+        @stream_ids.keys.include?(::File.basename(path))
+      end
+
+      def add_stream_id(path, last_modified = Time.now)
+        @stream_ids[::File.basename(path)] = last_modified
+      end
+
       def stream_id
         @stream_id && ::File.basename(@stream_id)
       end
 
       private
@@ -77,9 +123,10 @@
       end
 
       def split_filename(filename)
         filename.to_s.scan(/^([^.]*?)(?:-([0-9]+))?(\..+)?$/).flatten
       end
+
     end
 
   end
 end