fitz.get_text ignores 'pages' kwarg

[ananyablonko]

Description of the bug

The function fitz.get_text ignores the optional keyword argument 'pages'. Oddly enough, it does not ignore it if the method used to read pages is 'mp' (multiprocessing).
The culprit is fitz.apply_pages.

How to reproduce the bug

import fitz
from pathlib import Path

pdf_name = "C-standard"
path = str(Path(f"~/Downloads/{pdf_name}.pdf").expanduser())
pages = list(range(2, 3))
text1 = fitz.get_text(path, pages=pages)
text2 = fitz.get_text(path, pages=pages, method='mp')
assert text1 == text2

I would expect text1 and text2 to be the same, but text1 includes text from the entire file whilst text2 is as expected (only required page(s)).

The culprit:
lines 20633-20640:

    if method == 'single':
        if initfn:
            initfn(*initfn_args, **initfn_kwargs)
        ret = list()
        document = Document(path)
        for page in document:
            r = pagefn(page, *pagefn_args, **initfn_kwargs)
            ret.append(r)

Instead of this:

    ret = list()
    ...
    for page in document:
        r = pagefn(page, *pagefn_args, **initfn_kwargs)
        ret.append(r)

I would expect something like:

    ret = [pagefn(page, *pagefn_args, **initfn_kwargs) for page in doc if page.number in pages]

I would create a pull request but I am unsure if there's a faster way to get only specific pages.

Thank you for your time.

PyMuPDF version

1.26.0

Operating system

Windows

Python version

3.13

[julian-smith-artifex-com]

Many thanks for the report. I have a fix in my tree, which will hopefully be merged soon.

[julian-smith-artifex-com]

Fixed in PyMuPDF-1.26.1.