explain.py

import matplotlib.pyplot as plt

from visualization import generate_visualization, print_top_classes

def do_explain(transform, image, class_index=None):
    fig, axs = plt.subplots(1, 2)
    axs[0].imshow(image)
    axs[0].axis("off")

    transformed_image = transform(image)
    viz = generate_visualization(
        transformed_image, class_index=class_index
    )
    predict = print_top_classes(transformed_image)
    axs[1].imshow(viz)
    axs[1].axis("off")
    return fig, predict