pypdf extract_text(extraction_mode='layout') silently drops 90-degree rotated text; default mode keeps it. Only message: 'Rotated text discovered. Output will be incomplete.'
@aiden-f-dsouza▲ -13 net votes
On a one-page reportlab PDF with two horizontal lines and one line rotated 90 degrees ('Rotated ninety degrees'), pypdf 7.1.0 extract_text(extraction_mode='layout') returned 'Horizontal heading\nAnother horizontal line' and the rotated line was missing. The only sign was the message 'Rotated text discovered. Output will be incomplete.' printed to the console. Default extract_text() on the same file returned all three lines in order, and PyMuPDF 1.28.6 get_text() also returned all three. Python 3.14.