fix-encoding

#!/usr/bin/env python3

import sys, os, re
import html

def help():
    string = """
    Usage: %s [FILE]
    Try to revert encoded passwords:
        - \\xFF syntax
        - unescape html entities

    With no FILE, or when FILE is -, read standard input.
    """
    string = string.replace("%s", sys.argv[0])
    string = string.replace(os.linesep + "    ", os.linesep)
    string = os.linesep.join(string.split(os.linesep)[1:-1])
    print(string)
    sys.exit(0)

# handle reading from stdin
stdin = "/dev/stdin"
if len(sys.argv) == 1 and not sys.stdin.isatty():
    sys.argv.append(stdin)
elif len(sys.argv) == 2 and sys.argv[1] == "-":
    sys.argv[1] = stdin

if len(sys.argv) != 2 or not os.path.exists(sys.argv[1]):
    help()

# regex to extract line contents without numeric prefix
regex = re.compile(b"\\\\x[0-9a-fA-F]{2}")

for line in open(sys.argv[1], 'rb'):
    # \xFF syntax
    # 
    for match in regex.findall(line):
        match = match.lower()
        val = bytes.fromhex(match[2:].decode())
        line = line.replace(match, val)

    # html entities
    try:
        line = html.unescape(line.decode()).encode()
    except:
        pass

    sys.stdout.buffer.write(line)