def get_args()

in build_obelics/01_download_warc.py [0:0]

• 27 lines of code
• 1 McCabe index (conditional complexity)

def get_args():
    parser = argparse.ArgumentParser(description="Download warc files from Common Crawl pointers.")
    parser.add_argument(
        "idx_job",
        type=int,
        help="Index of the job (between 0 and 199).",
    )
    parser.add_argument(
        "--path_metadata_dataset",
        type=str,
        default="s3://m4-datasets/webdocs/pointers_cc_dataset/",
        help="Path of the dataset containing the metadata to retrieve the warc files.",
    )
    parser.add_argument(
        "--path_save_dir_warc_dataset",
        type=str,
        default="s3://m4-datasets/webdocs/warc_dataset/",
        help="The directory to save the warc dataset.",
    )
    parser.add_argument(
        "--num_proc",
        type=int,
        default=4 * cpu_count(),
        help="Number of processes to use for the multiprocessing.",
    )
    args = parser.parse_args()
    return args