How to use the alibi.utils.data.Bunch function in alibi

categorical_features = [f for f in features if data[f].dtype == 'O']
    category_map = {}
    for f in categorical_features:
        le = LabelEncoder()
        data_tmp = le.fit_transform(data[f].values)
        data[f] = data_tmp
        category_map[features.index(f)] = list(le.classes_)

    # only return data values
    data = data.values
    target_names = ['<=50K', '>50K']

    if return_X_y:
        return data, labels

    return Bunch(data=data, target=labels, feature_names=features, target_names=target_names, category_map=category_map)

labels = []
    for i, member in enumerate(tar.getnames()[1:]):
        f = tar.extractfile(member)
        for line in f.readlines():
            try:
                line.decode('utf8')
            except UnicodeDecodeError:
                continue
            data.append(line.decode('utf8').strip())
            labels.append(i)
    tar.close()
    if return_X_y:
        return data, labels

    target_names = ['negative', 'positive']
    return Bunch(data=data, target=labels, target_names=target_names)

resp = requests.get(url_labels)
        resp.raise_for_status()
        label_dict = pickle.load(BytesIO(resp.content))
    except RequestException:
        logger.exception("Could not download labels, URL may be out of service")
        raise

    inv_label = {v: k for k, v in label_dict.items()}
    label_idx = inv_label[category]
    labels = np.array([label_idx for _ in range(nb_images)])

    if return_X_y:
        return data, labels

    target_names = [category for _ in range(nb_images)]
    return Bunch(data=data, target=labels, target_names=target_names)