Getting a list of header dependencies using GCC -MM

listdeps.py

#!/usr/bin/env python3

import argparse
import json
import re
import sys
from collections import OrderedDict
from subprocess import Popen, PIPE

RE_INC = re.compile(r'^(?P<target>.+?):\s+(?P<deps>.+?)$', re.MULTILINE)
SPACE_TOK = '\x1B'

def parse_deps(input_text):
    text = input_text.replace('\\\n', ' ').replace('\\ ', SPACE_TOK)
    deps = OrderedDict()
    for match in RE_INC.finditer(text):
        target = match.group('target').replace(SPACE_TOK, ' ')
        files = [f.replace(SPACE_TOK, ' ') for f in match.group('deps').split()]
        deps[target] = files
    return deps

def list_stdin_deps(encoding='utf-8'):
    return parse_deps(sys.stdin.read().decode(encoding))

def list_cmd_deps(cmd, encoding='utf-8'):
    cmd = [cmd[0], '-MM'] + cmd[1:] if len(cmd) > 1 else []
    proc = Popen(cmd, stdout=PIPE)
    text = proc.stdout.read().decode(encoding)
    return parse_deps(text)

if __name__ == '__main__':

    parser = argparse.ArgumentParser()
    parser.add_argument('-e', '--encoding', metavar='ENC', dest='encoding',
                        default='utf-8', help="input encoding (default 'utf-8')")
    parser.add_argument('-o', '--output', metavar='FILE', dest='output',
                        default='-', help="output file (default '-' for stdout)")
    parser.add_argument('-c', '--command', metavar='CMD', dest='command',
                        nargs=argparse.REMAINDER, default='-',
                        help="read from CMD command (default '-' for stdin)")
    args = parser.parse_args()

    if args.command == '-':
        deps = list_stdin_deps(encoding=args.encoding)
    else:
        deps = list_cmd_deps(args.command, encoding=args.encoding)

    if args.output == '-':
        print(json.dumps(deps, indent=2))
    else:
        with open(args.output, 'w') as out_file:
            out_file.write(json.dumps(deps, indent=2))