Using Incremental Saves - pymupdf/PyMuPDF GitHub Wiki

Since version 1.9.1, incremental saves are possible for PDF documents.

Saving incrementally has the following advantages (also see chapter 3.4.5 of the Adobe PDF Reference)

• it is normally a lot faster, because changes are appended to the file, it is not rewritten as a whole
• it spares handling intermediate files if all you want is actually updating a specific document

The call pattern of incremental saves is as follows:

doc.save(doc.name, incremental=True, ...)

Prerequisites

There are a number of prerequisites that must be met to use this facility:

1. Incremental saves are not possible for encrypted files - even after they have been successfully decrypted.
2. The internal structure of the document must be intact, i.e. no open in repair mode. If errors occur during opening the PDF, a flag will be set that prevents using incremental save later. A normal save is still possible.
3. Option incremental=True excludes options garbage and linear.
4. The file to save to must obviously be the original one. Therefore, documents opened from a memory area cannot be saved incrementally.

Typical Uses

The most typical uses are small changes to the document, like adding or deleting a small number of pages, updating bookmarks or metadata, etc. If changes become significant, there will always be a breakeven point when saving to a new file is better performance-wise.

The following code snippet deletes empty pages from a text oriented PDF (like the Adobe manual ...):

lst = list(range(doc.pageCount))
for i in lst:
    if not doc.getPageText(i):  # no text in page?
        lst.remove(i)           # do not keep it
if len(lst) < doc.pageCount:    # any pages without text?
    doc.select(lst)             # delete those
    try:                        # try incremental
        doc.save(doc.name, incremental=True)
    except:                     # save new if in repair mode
        doc.save("new.pdf", garbage=3)