Provide a "lightweight" cleaning mode for PDF

The way to go would by PyPDF2, but it's not actively maintained anymore, the documentation for fancy things is close to non-existent, has issues dealing with images and doesn't play well with Python3.