Maybe you can try using https://github.com/pymupdf/PyMuPDF to iterate through all annotations, obtain the deleted text based on the deletion marks, and find the associated pop-up comments.