When the dependencies has been installed, zip the folder into a zip-file where the folder itself is included. When Amazon unzips it the handler should be metadataparser/index.fetchBatch. One way to do this is to right-click the folder on OS X and chose to compress it.
Go to the AWS Console for your Lambda job and upload the zip-file.
Done. It can take a short while for the new code to reach all the Lambda servers, but that's to be expected and something that happens on our own infrastructure as well.