minipuke.py

import io, struct, argparse

class Chunk:
    def __init__(self, id):
        self.id = id
        self.children = []

def process_chunk( input, parent_chunk, remaining_data, depth = 0 ):
    if None != parent_chunk.id and not parent_chunk.id in [ 'MULT', 'WRAP', 'TALK', 'TLKB', 'LECF', 'LFLF', 'SONG', 'NEST', 'SOUN', 'DIGI', 'AKOS', 'AWIZ' ]:
        input.seek( input.tell() + remaining_data )
        print( ' ' * depth + f'{parent_chunk.id} size={remaining_data}' )
        return
    print( ' ' * depth + f'processing {parent_chunk.id}' )
    while 0 < remaining_data:
        id = input.read( 4 ).decode( 'ascii' )
        size = struct.unpack( '>I', input.read( 4 ))[0]
        chunk = Chunk( id )
        parent_chunk.children.append( chunk )
        process_chunk( input, chunk, size - 8, depth + 1 )
        remaining_data -= size

def parse_chunks_for_file( filename, xor_key = None ):
    with open(filename,'rb') as f: raw_data = f.read()
    if xor_key: raw_data = bytes( a ^ xor_key for a in raw_data )
    input = io.BytesIO( raw_data )
    root_chunk = Chunk( None )
    process_chunk( input, root_chunk, len( raw_data ) )
    return root_chunk

def main():
    parser = argparse.ArgumentParser( description='dump humongous chunks' )
    parser.add_argument( 'filename', help='filename' )
    parser.add_argument( '-x', '--xor', type=int, default=None, help='xor decryption key (use 105)')
    args = parser.parse_args()
    file_chunks = parse_chunks_for_file( args.filename, xor_key=args.xor )

if __name__ == "__main__":
    main()