Pure python reimplementation of .cpio.xz content extraction from pbzx file payload for OS X packages

parse_pbzx.py

# Pure python reimplementation of .cpio.xz content extraction from pbzx file payload originally here:
# http://www.tonymacx86.com/general-help/135458-pbzx-stream-parser.html
#
# Cleaned up C version (as the basis for my code) here, thanks to Pepijn Bruienne / @bruienne
# https://gist.github.com/bruienne/029494bbcfb358098b41

# Example usage:
# parse_pbzx('PayloadJava', 'PayloadJava.cpio.xz')

import struct

def parse_pbzx(pbzx_path, xar_out_path):
    f = open(pbzx_path, 'rb')
    pbzx = f.read()
    f.close()
    magic, pbzx = pbzx[:4],pbzx[4:]
    if magic != 'pbzx':
        raise "Error: Not a pbzx file"
    # Read 8 bytes for initial flags
    flags, pbzx = pbzx[:8],pbzx[8:]
    # Interpret the flags as a 64-bit big-endian unsigned int
    flags = struct.unpack('>Q', flags)[0]
    xar_f = open(xar_out_path, 'wb')
    while (flags & (1 << 24)):
        # Read in more flags
        flags, pbzx = pbzx[:8],pbzx[8:]
        flags = struct.unpack('>Q', flags)[0]
        # Read in length
        f_length, pbzx = pbzx[:8],pbzx[8:]
        f_length = struct.unpack('>Q', f_length)[0]
        xzmagic, pbzx = pbzx[:6],pbzx[6:]
        if xzmagic != '\xfd7zXZ\x00':
            xar_f.close()
            raise "Error: Header is not xar file header"
        f_length -= 6
        f_content, pbzx = pbzx[:f_length],pbzx[f_length:]
        if f_content[-2:] != 'YZ':
            xar_f.close()
            raise "Error: Footer is not xar file footer"
        xar_f.write(xzmagic)
        xar_f.write(f_content)
    try:
        xar_f.close()
    except:
        pass