Unpacking LZMA-compressed files from BSPs

bsp_lzma.py

#!/usr/bin/python3

import struct, lzma, sys

def unpack(fmt, stream):
	'''
	unpacks values from a stream
	source: https://stackoverflow.com/a/17537253
	'''
	size = struct.calcsize(fmt)
	buf = stream.read(size)
	return struct.unpack(fmt, buf)

if __name__ == "__main__":
	import argparse, os
	parser = argparse.ArgumentParser(
			description="Decompressese an LZMA file extracted from a Source Engine BSP.",
			usage='%(prog)s FILE')
	
	parser.add_argument('filepath', metavar='FILE', help="file to decompress")
	
	args = parser.parse_args()
	
	if not os.path.isfile(args.filepath):
		raise ValueError("file not defined")
	
	with open(f'{args.filepath}', mode='rb') as f:
		# id, actual_size, lzma_size, *properties = unpack('<III5B', f)
		
		# properties can be unpacked directly, [0] = d, [1:4] = dict_size
		# see LzmaProps_Decode in mp/src/utils/lzma/C/LzmaDec.c
		header_id, actual_size, lzma_size, d, dict_size = unpack('<IIIBI', f)
		
		if not header_id == 0x414D5A4C:
			raise AssertionError('Not a valid Source Engine (PC) LZMA file')
		
		if d >= (9 * 5 * 5):
			raise AssertionError('throw SZ_ERROR_UNSUPPORTED')
		
		lc, pb, lp = d % 9, (d // 9) // 5, (d // 9) % 5
		
		# TODO what filter do the other files use?
		lzma_filter = [
			{ 'id': lzma.FILTER_LZMA1, 'dict_size': dict_size, 'lc': lc, 'lp': lp, 'pb': pb, 'preset': 6}
		]
		
		decompressor = lzma.LZMADecompressor(format = lzma.FORMAT_RAW, filters = lzma_filter)
		stream_decompressed = decompressor.decompress(f.read(), max_length = actual_size)
		
		sys.stdout.write(stream_decompressed.decode('ascii'))