pymupdf4llm.to_markdown raises error when using importing pymupdf.layout

[Tomaz12345]

Document: at link, its path saved into temp_file variable.

>>> import pymupdf.layout as pymupdf_layout
>>> md = pymupdf4llm.to_markdown(
...     temp_file,
...     write_images=True,
...     image_path='/tmp/imagessdkjfjns/'
... )
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
    md = pymupdf4llm.to_markdown(
        temp_file,
        write_images=True,
        image_path='/tmp/imagessdkjfjns/'
    )
  File "/home/admin/.venv/lib/python3.13/site-packages/pymupdf4llm/helpers/pymupdf_rag.py", line 1207, in to_markdown
    parms = get_page_output(
        doc,
    ...<5 lines>...
        IGNORE_GRAPHICS,
    )
  File "/home/admin/.venv/lib/python3.13/site-packages/pymupdf4llm/helpers/pymupdf_rag.py", line 1035, in get_page_output
    omitted_table_rects.append(pymupdf.Rect(t.bbox))
                                            ^^^^^^
  File "/home/admin/.venv/lib/python3.13/site-packages/pymupdf/table.py", line 1506, in bbox
    min(map(itemgetter(0), c)),
    ~~~^^^^^^^^^^^^^^^^^^^^^^^
ValueError: min() iterable argument is empty```

[JorjMcKie]

There are two problems / questions here:

1. You did not mention the pymupdf4llm version. The problem should not occur with version 0.2.2.
2. Why do you rename pymupdf to pymupdf_layout? The import of pymupdf.layout provides the full PyMuPDF namespace, so you can say pymupdf.open() etc. as usual.

[Tomaz12345]

@JorjMcKie

pip show pymupdf4llm
Name: pymupdf4llm
Version: 0.2.2
Summary: PyMuPDF Utilities for LLM/RAG
Home-page: https://github.com/pymupdf/RAG
Author: Artifex
Author-email: support@artifex.com
License: Dual Licensed - GNU AFFERO GPL 3.0 or Artifex Commercial License

pdf4llm.to_markdown(
    temp_file,
    write_images=True,
    image_path='/tmp/imagessdkjfjns/'
)>>> import pymupdf.layout
>>> md = pymupdf4llm.to_markdown(
...     temp_file,
...     write_images=True,
...     image_path='/tmp/imagessdkjfjns/'
... )
Traceback (most recent call last):
  File "/home/admin/.venv/lib/python3.13/site-packages/pymupdf4llm/helpers/pymupdf_rag.py", line 1041, in get_page_output
    omitted_table_rects.append(pymupdf.Rect(t.bbox))
                                            ^^^^^^
  File "/home/admin/.venv/lib/python3.13/site-packages/pymupdf/table.py", line 1506, in bbox
    min(map(itemgetter(0), c)),
    ~~~^^^^^^^^^^^^^^^^^^^^^^^
ValueError: min() iterable argument is empty

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
    md = pymupdf4llm.to_markdown(
        temp_file,
        write_images=True,
        image_path='/tmp/imagessdkjfjns/'
    )
  File "/home/admin/.venv/lib/python3.13/site-packages/pymupdf4llm/helpers/pymupdf_rag.py", line 1206, in to_markdown
    parms = get_page_output(
        doc, pno, margins, textflags, FILENAME, IGNORE_IMAGES, IGNORE_GRAPHICS
    )
  File "/home/admin/.venv/lib/python3.13/site-packages/pymupdf4llm/helpers/pymupdf_rag.py", line 1043, in get_page_output
    omitted_table_rects.append(pymupdf.Rect(t.bbox))
                                            ^^^^^^
  File "/home/admin/.venv/lib/python3.13/site-packages/pymupdf/table.py", line 1506, in bbox
    min(map(itemgetter(0), c)),
    ~~~^^^^^^^^^^^^^^^^^^^^^^^
ValueError: min() iterable argument is empty```

[JorjMcKie]

Anyway, I just confirmed that pymupdf4llm v0.2.2 successfully processes this file and produces this log output:

>py test-4llm.py
Parsing 422 pages of 'petrol-lp-2024-ang-web.pdf'...
  1%|▌                                         | 6/422 [00:02<02:48,  2.46it/s]Performing OCR on page.number=6[7]...
  6%|██▌                                      | 26/422 [00:19<03:41,  1.79it/s]Performing OCR on page.number=26[27]...
 27%|██████████▌                             | 112/422 [01:15<03:11,  1.62it/s]Performing OCR on page.number=112[113]...
 69%|███████████████████████████▋            | 292/422 [03:56<02:06,  1.03it/s]Performing OCR on page.number=292[293]...
100%|████████████████████████████████████████| 422/422 [07:19<00:00,  1.04s/it]
Generating markdown text...
100%|██████████████████████████████████████| 422/422 [00:00<00:00, 1136.39it/s]

[Tomaz12345]

How did you run your example.

[JorjMcKie]

This is the script:

import pymupdf.layout
import pymupdf4llm
import pathlib

md = pymupdf4llm.to_markdown(
    "petrol-lp-2024-ang-web.pdf", write_images=True, show_progress=True
)

[JorjMcKie]

I am going to close this issue.
I did not look at details of the produced markdown. Please report any problems there as separate issues.