Generates a custom MD5 hash for a PDF after stripping all the metadata and using a deterministic PDF ID

get_hash_from_pdf.py

import sys
import hashlib
import subprocess
from tempfile import NamedTemporaryFile

def get_hash_from_pdf(input_pdf):
    with NamedTemporaryFile(delete=True) as temp_pdf:
        # Run qpdf to strip metadata and output to a temporary file
        subprocess.run([
            'qpdf', '-empty', '-static-id', '-pages', input_pdf, '1-z', '--', temp_pdf.name
        ], check=True, stderr=subprocess.DEVNULL, stdout=subprocess.DEVNULL)
        # Compute and print MD5 hash of the final PDF (this is pretty much an invariant, tried it on a bunch of pdfs)
        file_data = temp_pdf.read()
        md5_hash = hashlib.md5(file_data).hexdigest()
        return md5_hash


if __name__ == "__main__":
    if len(sys.argv) != 2:
        print("Usage: python get_hash_from_pdf.py input_pdf.pdf")
        sys.exit(1)
    print(get_hash_from_pdf(sys.argv[1]))

The only dependency here is qpdf and that can be installed with sudo apt-get install qpdf or brew install qpdf