In pymupdf4llm, if a page has multiple images, only 1 image per-page is extracted

[dhavalsingh]

Description of the bug

test_with_images (5).pdf

For example, this pdf with multiple images per page, using

pymupdf4llm.to_markdown(
                updated_save_path, write_images=True, image_path=image_path
            )

we get only 3 images, also, the last image per page.

How to reproduce the bug

1. use the above pdf or any pdf with multiple images per page
2. Run the code

pymupdf4llm.to_markdown(
                updated_save_path, write_images=True, image_path=image_path
            )

Check the output folder.

PyMuPDF version

1.24.13

Operating system

MacOS

Python version

3.12

[JorjMcKie]

This has nothing to do with PyMuPDF itself, but with this package.
For this package, you did not mention a version number.
Your finding looks like a know bug in v0.0.17. Please try the version 0.0.16 instead.

[Amraneze]

@JorjMcKie You are correct, the issue is with PyMuPDFLLM and exactly in this line.

#

RAG/pymupdf4llm/pymupdf4llm/helpers/pymupdf_rag.py

Line 710 in b257182

del parms.img_rects[i] # do not touch this image twice

The problem is that if you remove an element of the list, which means the index has been changed, if we have a list of [1, 2, 3, 4, 5], in the first iteration i = 0 we remove the first element of the list, that will results us in the following list [2, 3, 4, 5], but with enumerate it will goes to the element 3 and skip 2 because the next index is i = 1. Instead of processing the 5 elements it will process only 3 elements