def get_args()

in utils/hf_dataset_subsampling.py [0:0]

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

def get_args():
    parser = argparse.ArgumentParser()
    parser.add_argument('--name', type=str, required=True,
                        help="Path to the dataset you're using on the HF hub. Pass e.g. `csv` or `json` and `data_files=path_on_disk` to load something locally")
    parser.add_argument('--subset', type=str, default=None, help="Subset of the dataset you're using, if needed")
    parser.add_argument('--data_files', type=str, default=None, help="Path to the dataset on disk if using local files")
    parser.add_argument('--ratios', nargs='+', type=float, help="Subsampling ratios", required=True)
    parser.add_argument('--names', nargs='+', type=str, help="Names for the produced subsets", required=False)
    parser.add_argument('--pre_shuffle', action="store_true", help="Whether to shuffle the dataset in advance")
    parser.add_argument('--shuffle_seed', type=int, default=0, help="Shuffling seed")
    return parser.parse_args()