Fix pgfplots matrix plot artifacts (https://tex.stackexchange.com/questions/445977/pgfplots-matrix-plot-perfect-rectangles)

README.md

This script fixes the problem described in: https://tex.stackexchange.com/questions/445977/pgfplots-matrix-plot-perfect-rectangles

The script needs pdftk to be available on the path.

Then you can run it with:

$ python fix_figure.py -i input_file.pdf -o fixed_file.pdf

fix_figure.py

#!/usr/bin/env python
# -*- coding: utf-8 -*-

"""
This must be one of the more nit-picky things I've done.

This script requires pdftk to be available on the path:
https://www.pdflabs.com/tools/pdftk-the-pdf-toolkit/

It fixes artifacts in the pdf images by changing the "fill and stroke" paths to 
simple "fill" paths.

Author: Gertjan van den Burg

"""

import argparse
import os
import re
import subprocess
import sys
import tempfile


def parse_args():
    parser = argparse.ArgumentParser()
    parser.add_argument("-i", "--input-file", help="Input file")
    parser.add_argument("-o", "--output-file", help="Output file")
    return parser.parse_args()


def fix_stack(stack):
    """ Fix the block if we encounter it """
    # We are specifically looking for blocks of the following form:
    #
    # B
    # float float float RG
    # float float float rg
    # float float m
    # float float l
    # float float l
    # float float l
    # h
    #
    # Anything that does not match this format is not affected
    # The fix is to replace the 'B' line with an 'f' line.

    if any(s is None for s in stack):
        return

    block = b"".join(stack)
    regex = b"B \n((\d|0\.\d+)\ ){3}RG\ \n((\d|0\.\d+)\ ){3}rg\ \n(\-?(\d+|\d+\.\d+)\ ){2}m\ \n((\-?(\d+|\d+\.\d+)\ ){2}l\ \n){3}h\ \n"
    if re.fullmatch(regex, block):
        stack[0] = b"f \n"


def uncompress(source, target):
    # decompress the pdf
    status = subprocess.call(["pdftk", source, "output", target, "uncompress"])
    if not status == 0:
        print("Failed to decompress the file with pdftk", file=sys.stderr)
        raise SystemExit(1)


def compress(source, target):
    status = subprocess.call(["pdftk", source, "output", target, "compress"])
    if not status == 0:
        print("Failed to compress the file with pdftk", file=sys.stderr)
        raise SystemExit(2)


def run_fix(uncompressed, fixed):
    # replace the blocks where necessary
    stack = [None] * 8
    with open(fixed, "wb") as ofp, open(uncompressed, "rb") as ifp:
            for line in ifp:
                first = stack.pop(0)
                if not first is None:
                    ofp.write(first)
                stack.append(line)
                fix_stack(stack)
            for line in stack:
                ofp.write(line)


def main():
    args = parse_args()
    with tempfile.TemporaryDirectory(prefix="pdf_fix_") as tmpdir:
        t1 = os.path.join(tmpdir, "uncompressed.pdf")
        t2 = os.path.join(tmpdir, "fixed.pdf")
        
        uncompress(args.input_file, t1)
        run_fix(t1, t2)
        compress(t2, args.output_file)


if __name__ == "__main__":
    main()