extractText() extracts broken text from pdf

[spagliarini]

Description of the bug

Hi,

I noticed a bug in PyMuPDF version > 1.23.9 (included) when using get_text to extract text from PDF documents.

To reproduce the bug

• Consider the attached PDF file: test_file.pdf

• Extract text using the code below (see "How to reproduce the bug")

• To reproduce the correct behavior install a PyMuPDF version < 1.23.9 (e.g., 1.23.8). We obtain the following complete text: doc_text_1238.txt

• To reproduce the bug behavior install a PyMuPDF version >= 1.23.9 (e.g., 1.23.24). We obtain the following broken text: doc_text_12324.txt

ADDITIONAL NOTES

• The bug behavior can only be observed on certain documents (e.g., the one attached above)
• extractBLOCKS, extractWORDS and extractDICT work fine, the bug seems to show only for extractTEXT
• We tried both on windows and linux and neither works

Thank you for your help

How to reproduce the bug

Extract text using the following code

fitz_doc = fitz.open(pdf_path)

doc_text = list()
for page in fitz_doc:
    doc_text.append(page.get_text())

doc_text = ' '.join(doc_text)

To reproduce the bug behavior install a PyMuPDF version >= 1.23.9 (e.g., 1.23.24).

PyMuPDF version

1.23.24

Operating system

Windows

Python version

3.10

[julian-smith-artifex-com]

Fixed in 1.23.25.