Wouldn't it be easier/faster to just parse the desired URL from the link?