The issue was with pdfminer interpreting the Unicode mapping data. If Ghostscript rewrote it, it could have worked around the issue. Even a one byte adjustment could have been a workaround. --redo-ocr ...

Summary: pdfminer3k is a Python 3 port of pdfminer — a tool for extracting information from PDF documents pdfminer3k is a Python 3 port of pdfminer. PDFMiner is a tool for extracting information from ...