def parse_args()

in single-rai-job/src/run_rai.py [0:0]

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

def parse_args():
    # setup arg parser
    parser = argparse.ArgumentParser()

    # Constructor arguments
    parser.add_argument("--title", type=str, required=True)
    parser.add_argument(
        "--task_type", type=str, required=True, choices=["classification", "regression", "forecasting"]
    )
    parser.add_argument("--model_id", type=str, help="name:version", required=True)
    parser.add_argument("--train_dataset_id", type=str, required=True)
    parser.add_argument("--test_dataset_id", type=str, required=True)
    parser.add_argument("--target_column_name", type=str, required=True)
    parser.add_argument("--maximum_rows_for_test_dataset", type=int, default=5000)
    parser.add_argument(
        "--categorical_column_names", type=str, help="Optional[List[str]]"
    )
    parser.add_argument("--classes", type=str, help="Optional[List[str]]")

    # Causal arguments
    parser.add_argument("--enable_causal", type=boolean_parser, required=True)
    parser.add_argument(
        "--causal_treatment_features", type=json.loads, help="List[str]"
    )
    parser.add_argument(
        "--causal_heterogeneity_features",
        type=json.loads,
        help="Optional[List[str]] use 'null' to skip",
    )
    parser.add_argument(
        "--causal_nuisance_model", type=str, choices=["linear", "automl"]
    )
    parser.add_argument(
        "--causal_heterogeneity_model", type=str, choices=["linear", "forest"]
    )
    parser.add_argument("--causal_alpha", type=float)
    parser.add_argument("--causal_upper_bound_on_cat_expansion", type=int)
    parser.add_argument(
        "--causal_treatment_cost",
        type=float_or_json_parser,
        help="Union[float, List[Union[float, np.ndarray]]]",
    )
    parser.add_argument("--causal_min_tree_leaf_samples", type=int)
    parser.add_argument("--causal_max_tree_depth", type=int)
    parser.add_argument("--causal_skip_cat_limit_checks", type=boolean_parser)
    parser.add_argument("--causal_categories", type=str_or_list_parser)
    parser.add_argument("--causal_n_jobs", type=int)
    parser.add_argument("--causal_verbose", type=int)
    parser.add_argument("--causal_random_state", type=int_or_none_parser)

    # Counterfactual arguments
    parser.add_argument("--enable_counterfactual", type=boolean_parser, required=True)
    parser.add_argument("--counterfactual_total_CFs", type=int, required=True)
    parser.add_argument("--counterfactual_method", type=str)
    parser.add_argument("--counterfactual_desired_class", type=str_or_int_parser)
    parser.add_argument(
        "--counterfactual_desired_range", type=json_empty_is_none_parser, help="List"
    )
    parser.add_argument(
        "--counterfactual_permitted_range", type=json_empty_is_none_parser, help="Dict"
    )
    parser.add_argument("--counterfactual_features_to_vary", type=str_or_list_parser)
    parser.add_argument("--counterfactual_feature_importance", type=boolean_parser)

    # Error analysis arguments
    parser.add_argument("--enable_error_analysis", type=boolean_parser, required=True)
    parser.add_argument("--error_analysis_max_depth", type=int)
    parser.add_argument("--error_analysis_num_leaves", type=int)
    parser.add_argument("--error_analysis_min_child_samples", type=int)
    parser.add_argument(
        "--error_analysis_filter_features", type=json.loads, help="List"
    )

    # Explanation arguments
    parser.add_argument("--enable_explanation", type=boolean_parser, required=True)

    # Output arguments
    # parser.add_argument("--dashboard", type=str, required=True)
    # parser.add_argument("--ux_json", type=str, required=True)

    # parse args
    args = parser.parse_args()

    # return args
    return args