plt.show()


if __name__ == '__main__':
    model_path = sys.argv[1]
    df_path = sys.argv[2]
    y_path = sys.argv[3]

    # get model
    model = pickle.load( open(model_path, 'rb') )

    # get_features
    df = pd.read_pickle(df_path)
    y = pd.read_pickle(y_path)

    features = np.array(df.columns.tolist())

    # Make same splits as training
    X_train, X_test, y_train, y_test = train_test_split(df, y, test_size=0.1, random_state=RANDOM_NUM)


    #make_feat_plot(model,features)


    #make_part_plot(model, features)

    #make_pred_plot(model, X_test, y_test)


    predicted_value_plot(model=model, df=df, column='number_topics', classification=False, discrete_col=False, freq=False, response_label='Subsrciber_Count', xlim=(3,23))
    plt.subplots_adjust(top=0.9)  # tight_layout causes overlap with suptitle

    plt.show()
    return fig, axs



if __name__ == '__main__':
    model_path = sys.argv[1]
    df_path = sys.argv[2]

    # get model
    model = pickle.load( open(model_path, 'rb') )

    # get_features
    df = pd.read_pickle(df_path)





    features = np.array(df.columns.tolist())


    #make_feat_plot(model,features)


    #fig, axs = make_part_plot(model, features)

    predicted_value_plot(model=model, df=df, column='day_of_week', classification=True, class_pred='great traction', discrete_col=False, freq=False, response_label='Likelehood')