d78a9d2692be1e7a32d157cf870126fe7abe6643,pmlb/pmlb.py,,fetch_data,#,198

Before Change

        if return_X_y == True: A tuple of NumPy arrays containing (features, labels)

    
    if dataset_name not in dataset_names:
        raise ValueError("Data set not found in PMLB.")

    dataset_url = "https://github.com/EpistasisLab/penn-ml-benchmarks/raw/master/datasets/{DATASET_NAME}/{DATASET_NAME}.csv.gz".format(DATASET_NAME=dataset_name)

    if local_cache_dir is None:
        dataset = pd.read_csv(dataset_url, sep="\t", compression="gzip")

After Change

        if return_X_y == True: A tuple of NumPy arrays containing (features, labels)

    
    if dataset_name in classification_dataset_names:
        data_type = "classification"
    elif dataset_name in regression_dataset_names:
        data_type = "regression"
    else:
        raise ValueError("Data set not found in PMLB.")

    dataset_url = "{GITHUB_URL}/{DATA_TYPE}/{DATASET_NAME}/{DATASET_NAME}{SUFFIX}".format(GITHUB_URL=GITHUB_URL,
                                DATA_TYPE=data_type,
                                DATASET_NAME=dataset_name,
                                SUFFIX=suffix

Instances