The Deredactatron: De-redacts PDF's redacted by drawing rectangles over the text. Needs Python3 and PyMuPDF

deredactatron.py

#!/usr/bin/env python3

# De-redacts PDF's that have been redacted by drawing PDF-native rectangles over the text
# This removes ALL rectangles so might also nuke any shape diagrams a bit

# python deredactatron.py a_redacted_doc.pdf
# Outputs: deredacted-a_redacted_doc.pdf

import sys
import re
import fitz  # PyMuPDF==1.17.5

doc = fitz.open(sys.argv[1])
for page_number, page in enumerate(doc, start=1):
    for xref in page._getContents():
        stream = doc.xrefStream(xref).decode()
        # Format: `36.95 806.58 149.32 -14.6 re\n`
        stream = re.sub("-?\d+(\.\d+)? -?\d+(\.\d+)? -?\d+(\.\d+)? -?\d+(\.\d+)? re\r?\n", "", stream)
        doc.updateStream(xref, stream.encode())

doc.save("deredacted-" + sys.argv[1], clean=True)