mc_hex2bin.py

import re
import binascii
import sys

def convert_hex(input_file, output_file):
    """ input: list containing raw lines of hex
        output: continuous binary data
    """
    last_addr = None
    for line in input_file:
        addr = parse_prefix(line)
        if addr == 0:
            last_addr = addr
            output_file.write(line_h2b(line))
        elif (addr - last_addr) == 0x10:
            last_addr = addr
            output_file.write(line_h2b(line))
        elif (addr - last_addr) > 0x10:
            for null_addr in range((last_addr+0x10), (addr), 0x10):
                output_file.write("\x00" * 0x10)
            last_addr = addr
            output_file.write(line_h2b(line))

        else:
            sys.exit("bad hex data? parsing line: %x" % addr)
        
def parse_prefix(line):
    """ input: raw hex line, example:
            0000:   0000 4400 0000 0000 0000 0000 0000 0000   ..D............. 
        output: integer value of the line location
    """
    addr = re.search("^(\w\w\w\w):", line)
    if addr:
        return int(addr.group(1), 16)
    else:
        print "[DEBUG] no prefix addr returned: %s" % line
        return None

def line_h2b(line):
    """ input: raw hex line, example:
            0000:   0000 4400 0000 0000 0000 0000 0000 0000   ..D............. 
        output: binary equivelant of the line of hex
    """
    stripped_line = re.search("^\w{4}:\s*(\w{4}\s\w{4}\s\w{4}\s\w{4}\s\w{4}\s\w{4}\s\w{4}\s\w{4})\s*", line)
    if stripped_line:
        hex_data = re.sub(r'\s', "", stripped_line.group(1))
        return binascii.a2b_hex(hex_data)
    elif re.search("^\w{4}:\s*\*", line):
        return "\x00" * 0x10
    else:
        return None

def main():
    if len(sys.argv) < 3:
        sys.exit("usage: %s <hex file> <binary output>" % sys.argv[0])
    hex_f = open(sys.argv[1], 'rb')
    bin_f = open(sys.argv[2], 'wb')
    bindata = convert_hex(hex_f, bin_f)
    hex_f.close()
    bin_f.close()

if __name__ == '__main__':
    main()