Process everything in chunks

The old implementation loaded files into memory to provide an IO interface. This has the obvious downside of loading any file included in the import, e.g. a 10GB video file, into memory. ZipKit has no native IO object for reading but it provides all the necessary methods to implement one.

app/models/zip_file/reader.rb

@@ -10,12 +10,11 @@ class ZipFile::Reader
     raise ArgumentError, "Cannot read directory entry: #{file_path}" if entry.filename.end_with?("/")
 
     extractor = entry.extractor_from(@io)
-    content = extractor.extract
 
     if block_given?
-      yield StringIO.new(content)
+      yield ZipFile::Reader::ExtractorIO.new(extractor)
     else
-      content
+      extractor.extract
     end
   end
 

app/models/zip_file/reader/extractor_io.rb

@@ -0,0 +1,23 @@
+class ZipFile::Reader::ExtractorIO
+  def initialize(extractor)
+    @extractor = extractor
+  end
+
+  def read(length = nil, buffer = nil)
+    return nil if @extractor.eof?
+
+    data = @extractor.extract(length)
+    return nil if data.nil?
+
+    if buffer
+      buffer.replace(data)
+      buffer
+    else
+      data
+    end
+  end
+
+  def eof?
+    @extractor.eof?
+  end
+end

app/models/zip_file/writer.rb

@@ -26,16 +26,12 @@ class ZipFile::Writer
 
   def add_file(path, content = nil, compress: true)
     @entries << path
-    compression = compress ? :deflate : :stored
+    write_method = compress ? :write_deflated_file : :write_stored_file
 
     if block_given?
-      streamer.write_deflated_file(path) do |sink|
-        yield sink
-      end
+      streamer.public_send(write_method, path) { |sink| yield sink }
     else
-      streamer.write_deflated_file(path) do |sink|
-        sink.write(content)
-      end
+      streamer.public_send(write_method, path) { |sink| sink.write(content) }
     end
   end