unpack Unreal Engine .pck files into .wav

unpck.py

"""unpack Unreal Engine .pck audio packs and convert them into .wav

quickbms and vgmstream must be installed and available on your PATH as 'quickbms' and 'vgmstream-cli'.
If not, specify the paths using the '-q' and '-v' options:

unpck -q=<path to quickbms binary> -v=<path to vgmstream-cli> ...

known issues:
- vgmstream does not support Unicode filenames on Windows, so the WEM file will be renamed temporarily to use an ASCII filename if it contains Unicode characters.
"""

import argparse
import sys
import subprocess
import shutil
from pathlib import Path

PCK_EXT = "*.pck"
WEM_EXT = "*.wem"

QUICKBMS = shutil.which("quickbms")
VGMSTREAM = shutil.which("vgmstream-cli")

# wavescan.bms
WAVESCAN = """
# scan data for wave files
# RIFF and RIFX header supported
# note: There are wave files with a wrong file size after RIFF/RIFX
#       This script takes the stream size, adds the header size and writes the correct size after RIFF/RIFX
# (c) 2012-06-26 by AlphaTwentyThree
#
# future update plans:
# - option to also write data between found wave files to disk
# - option to automatically transform the file to a playable or at least decodable format

for i = 1                        # run through loop with count variable i
   FindLoc OFFSET string "WAVE" 0 ""   # search for "WAVE", save position as variable OFFSET
   if OFFSET == ""                  # when nothing is found
      cleanexit                  # the script exits (e.g. at end of file)
   endif
   math OFFSET -= 8               # jump to possible
   goto OFFSET                     # RIFF/RIFX file start
   getDstring IDENT 4               # read string of 4 bytes, save variable as IDENT
   if IDENT == "RIFX"               # differentiate between header possibilities
      endian big                  # set endianness to big, if file has RIFX identifier
      callfunction write 1         # see function section below
   elif IDENT == "RIFF"            # endianness stays little
      callfunction write 1         # also run function
   else                        # string "WAVE" found, but doesn't belong to wave file
      set SIZE 0xc               # do as if something with 0xc bytes was found to continue search from the right position
   endif
   set SEARCH OFFSET               # set marker to position from where to search next
   math SEARCH += SIZE               # (that would be after the file that was found)
   if SEARCH == FSIZE               # in case the last found file ends with the main file, we exit
      cleanexit
   endif
   goto SEARCH                     # if we haven't exited the script above, we set out cursor to after the last found file
next i

startfunction write                  # function "write" starts here, is called when a wave file is found above
   get NAME basename               # save name without extension under variable NAME
   string NAME += "_"               # add underscore to the name
   string NAME += i               # add the loop variable to the name
   goto OFFSET                     # set cursor to the beginning of the found file
   get DUMMY long                  # RIFF/RIFX identifier, not needed
   get DUMMY long                  # riff size, not needed
   get DUMMY long                  # "WAVE", not needed, we arrive at the "fmt " section
   for                           # loop search for the "data" section at the start of the stream (get the stream size from there)
      getDstring AREA_NAME 4         # name of area in header
      get AREA_SIZE long            # size of area in header
      savepos MYOFF               # save position we are at
      if AREA_NAME == "data"         # when we arrive at the needed "data" area:
         break                  # we exit the loop
      else                     # otherwise:
         if AREA_NAME == "LIST"      # that's the area of the marker names
            callfunction retrievename 1 # see below
         endif
         math MYOFF += AREA_SIZE      # not reached "data" area -> adjust cursor position...
         goto MYOFF               # ... and go there
      endif
   next                        # remember: the cursor is now directly at the stream start
   set STREAMSIZE AREA_SIZE         # the last AREA_SIZE is the size we need (size of the audio stream)
   set HEADERSIZE MYOFF            # 
   math HEADERSIZE -= OFFSET         # calculate the size of the stream header (offset - offset = size)
   set SIZE HEADERSIZE               # 
   math SIZE += STREAMSIZE            # calculate complete file size (header + stream = file)
   log MEMORY_FILE OFFSET SIZE         # write file to memory
   math SIZE -= 8                  # subtract 8 bytes to get the riff size
   putVarChr MEMORY_FILE 4 SIZE long    # write the correct riff size to the header inside the memory
   string NAME += ".wem"            # add extension to the name (the name could contain the name of the first marker if the file has markers)
   math SIZE += 8                  # add the subtracted 8 bytes again
   log NAME 0 SIZE MEMORY_FILE         # write file in memory to disk
endfunction                        # remember that we continue with our next i now!

startfunction retrievename            # get possible file name from first marker name, rmember: our cursor is after the size of the LIST area
   get DUMMY long                   # always "adtl", not needed
   get DUMMY long                  # always "labl", not needed
   get MSIZE long                  # size of the label area for this marker
   math MSIZE -= 4                  # subtracting 4 bytes leaves us with the length of the marker label
   get DUMMY long                  # marker type, not needed
   getDstring MNAME MSIZE            # cursor is at the beginning of the label name, now get the marker name with the desired length MSIZE
   string NAME += "~"
   string NAME += MNAME            # add the marker name to the file name
endfunction
"""


def cmd(args, ignore):
    result = subprocess.run(args, capture_output=True, encoding="utf-8")
    if result.returncode != 0:
        print(
            f"error: command returned exit code {result.returncode}.",
            f"args: {result.args}",
            f"output: {result.stderr}",
            sep="\n",
        )
        if not ignore:
            sys.exit(1)


def main():
    parser = argparse.ArgumentParser(
        description=__doc__, formatter_class=argparse.RawDescriptionHelpFormatter
    )
    parser.add_argument("-q", "--quickbms", help="path to quickbms binary")
    parser.add_argument("-v", "--vgmstream", help="path to vgmstream-cli binary")
    parser.add_argument(
        "-i",
        "--ignore-errors",
        action="store_true",
        help="ignore errors when running commands",
    )
    parser.add_argument(
        "-k",
        "--keep-wem",
        action="store_true",
        help="keep the wem directory after conversion (it is in the dest folder)",
    )
    parser.add_argument("dest", help="where to output the .wav files")
    parser.add_argument(
        "files",
        nargs="+",
        help=".pck files or a folder with .pck files",
    )

    args = parser.parse_args()

    quickbms = args.quickbms or QUICKBMS
    vgmstream = args.vgmstream or VGMSTREAM

    dest = Path(args.dest)
    dest.mkdir(exist_ok=True)

    # check if path is a folder
    path = Path(args.files[0])
    if path.is_dir():
        files = path.glob(PCK_EXT)
    else:
        files = [Path(f) for f in args.files]

    script = dest / "wavescan.bms"
    script.write_text(WAVESCAN)

    wem_dir = dest / "wem"
    wem_dir.mkdir(exist_ok=True)

    for file in files:
        if file.is_dir():
            print("error: you can only specify 1 folder or any number of files")
            sys.exit(1)

        print(f"[*] unpacking {file}")
        cmd(
            [
                quickbms,
                str(script),
                str(file),
                wem_dir,
            ],
            args.ignore_errors,
        )

    print("[+] unpacking done!")

    for file in Path(wem_dir).glob(WEM_EXT):
        print(f"[*] converting {file}")

        name = file.stem
        unicode = not name.isascii()

        if unicode:
            print(
                f"[!] filename {name} contains Unicode characters, temporarily renaming with ASCII..."
            )
            new_file = file.rename(
                file.with_stem(name.encode("ascii", "ignore").decode("utf-8"))
            )
        else:
            # no need to rename
            new_file = file

        cmd(
            [
                vgmstream,
                "-o",
                str(dest / f"{new_file.stem}.wav"),
                str(new_file),
            ],
            args.ignore_errors,
        )

        if unicode:
            # rename back to original unicode filename
            new_file.rename(file)

    print("[+] converting done!")

    if not args.keep_wem:
        script.unlink()
        shutil.rmtree(wem_dir)


if __name__ == "__main__":
    main()