archive.py

import os, subprocess, mimetypes, zlib, sys, getopt

def help():
    help = [
            'Options:',
            '    -i (--input) - input stream file',
            '    -o (--output) - output archive file (.mkv)',
            '    -d (--dir) - archive directory (file)',
    ]
    print("\n".join(help))
    sys.exit()

def get_crc32(file_path):
    with open(file_path, 'rb') as file:
        return zlib.crc32(file.read())

def get_mime(file_path):
    mime_type = mimetypes.guess_type(file_path)[0]
    with open(file_path, 'rb') as file:
        if not mime_type:
            try:
                file.read().decode('utf-8')
            except UnicodeDecodeError:
                mime_type = 'application/octet-stream'
            else:
                mime_type = 'text/plain'
    return mime_type

def create_archive(archive_dir, input_file, output_file):
    file_list = []
    if os.path.isdir(archive_dir):
        for root, dirs, files in os.walk(archive_dir):
            for file in files:
                file_list.append(os.path.join(root, file))
    elif os.path.isfile(archive_dir):
        file_list.append(archive_dir)
    ffargs = ['ffmpeg', '-i', input_file, '-c', 'copy', '-map_metadata', '0']
    file_number = 0
    for file in file_list:
        mime_type = get_mime(file)
        crc32 = get_crc32(file)
        file_size = os.path.getsize(file)
        if mime_type and crc32 and file_size:
            ffargs += [
                        '-metadata:s:t:' + str(file_number), f'mimetype={mime_type}',
                        '-metadata:s:t:' + str(file_number), f'crc32={format(crc32, "08x")}',
                        '-metadata:s:t:' + str(file_number), f'size={file_size}',

                        '-attach', file
            ]
            file_number += 1

    ffargs += [output_file]
    subprocess.run(ffargs)

def main(argv):
    archive_dir = ''
    input_file = 'dummy.mkv'
    output_file = 'archive.mkv'

    if argv:
        opts, args = getopt.getopt(argv,'i:o:d:h',['input=','output=','dir='])
        for opt, arg in opts:
            if opt in ('-i', '--input'):
                input_file = arg
            elif opt in ('-o', '--output'):
                output_file = arg
            elif opt in ('-d', '--dir'):
                archive_dir = arg
            elif opt == '-h':
                help()
    else:
        help()

    create_archive(archive_dir, input_file, output_file)

if __name__ == '__main__':
    main(sys.argv[1:])