def main()

in utils/find_corpus.py [0:0]

• 49 lines of code
• 21 McCabe index (conditional complexity)

def main(args_list: Optional[list[str]] = None) -> None:
    importers = [
        "opus",
        "sacrebleu",
        "mtdata",
        "huggingface_mono",
        "huggingface_parallel",
        "huggingface_any",
        "news-crawl",
        "hplt-mono",
    ]
    parser = argparse.ArgumentParser(
        description=__doc__,
        formatter_class=argparse.RawTextHelpFormatter,  # Preserves whitespace in the help text.
    )
    parser.add_argument("source", type=str, nargs="?", help="Source language code")
    parser.add_argument("target", type=str, nargs="?", help="Target language code")
    parser.add_argument(
        "--importer",
        type=str,
        help=f"The importer to use: {', '.join(importers)}",
    )
    parser.add_argument(
        "--download_url",
        action="store_true",
        default=False,
        help="Show the download url if available.",
    )

    args = parser.parse_args(args_list)

    if not args.source or not args.target:
        parser.print_help()
        sys.exit(1)

    if args.importer and args.importer not in importers:
        print(f'"{args.importer}" is not a valid importer.')
        sys.exit(1)

    if args.importer == "opus" or not args.importer:
        get_opus(args.source, args.target, args.download_url)

    if args.importer == "sacrebleu" or not args.importer:
        get_sacrebleu(args.source, args.target)

    if args.importer == "mtdata" or not args.importer:
        get_mtdata(args.source, args.target)

    if args.importer == "huggingface_mono" or not args.importer:
        get_huggingface_monolingual(args.target if args.source == "en" else args.source)

    if args.importer == "huggingface_parallel" or not args.importer:
        get_huggingface_parallel(args.source, args.target)

    if args.importer == "huggingface_any" or not args.importer:
        get_huggingface_any(args.target if args.source == "en" else args.source)

    if args.importer == "news-crawl" or not args.importer:
        get_news_crawl(args.source, args.target)