Is there an automated process to take the OCR'd text from a PDF and make an EPUB like this or does it take a lot of manual adjustment?