[ENH]: Add click to enlarge subplots in a figure

[Coderx7]

Problem

When we add a few subplots to a figure, the detail may not be very visible, the zooming capability wouldn't work much, because it works inside each subplot, and is not of much help when the subplots itself is small.

It would be great if we could click on any subplots, and a popup window displays the said plot, as a separate window, in full size, as if it was plotted individually.

This way, we can have all subplots neatly sit next to each other for visualizations, and when one or more subplots need better investigation, simply clicking on them would display them in a separate window each, which can then be closed (the main figure stays open with all the subplots by the way).

Currently, when I want to investigate anything, in addition to creating the normal figure with all the plots, I use cv2.imshow to load the images that need investigation, or use separate individual plots in combination to the aggregated one which is a lot of work honestly.

I also think this can be done using events, but I guess its not doable if one uses the old/functional form i.e. sth like :

plt.subplot(5,6,1), plt.imshow(img_gentleman[:,:,::-1]), plt.title('original')
plt.subplot(5,6,2), plt.imshow(hist_img_gentleman[...,::-1]),plt.title('global hist eq')
...

aside from that, just having such ability goes a long way IMHO.

As an example see the following figure which consists of several subplots:

They show the overall differences between different methods, while some samples are easy to diffentiate, others need to be viewed in larger form to see the overall quality and check for stuff such as noises, etc.

Proposed solution

I haven't looked at the code, so take this with a pinch of salt! but creating a new popup window, shouldn't be an issue, provided we have access to the underlying data

[jklymak]

This seems useful, but hard to generalize. I am skeptical that it would be made default part of Matplotlib proper. We have a pretty straightforward GUI event handler that would be just a few lines of code to make something like this work:

import matplotlib.pyplot as plt
import numpy as np

def on_click(event):
    for i, ax in enumerate(axs.flat):
        if event.inaxes == ax:
            fig, new_ax = plt.subplots()
            new_ax.imshow(images[i], cmap='gray')
            new_ax.set_title(f'Image {i+1}')
            new_ax.axis('off')
            fig.show()

def create_random_image():
    return np.random.rand(100, 100)

# Generate four random images
images = [create_random_image() for _ in range(4)]

# Create a main figure to display the images
main_fig, axs = plt.subplots(2, 2, layout='compressed')

for i, ax in enumerate(axs.flat):
    ax.imshow(images[i], cmap='gray')
    ax.set_title(f'Image {i+1}')
    ax.axis('off')

# Connect the click event
main_fig.canvas.mpl_connect('button_press_event', on_click)

plt.show()

[tacaswell]

Similarly, you could change the position of clicked axes

import matplotlib.pyplot as plt
import numpy as np

cache = {}

def on_click(event):
    ax = event.inaxes
    if ax is None:
        return
    if ax not in cache:
        cache[ax] = ax.get_subplotspec()
        ax.set_position([0, 0, 1, 1])
        ax.set_zorder(5)
    else:
        sp = cache.pop(ax)
        ax.set_subplotspec(sp)
        ax.set_zorder(1)        
        ax.set_in_layout(True)
    print(ax.get_subplotspec())
    ax.figure.canvas.draw_idle()
    
def create_random_image():
    return np.random.rand(100, 100)

# Generate four random images
images = [create_random_image() for _ in range(4)]

# Create a main figure to display the images
main_fig, axs = plt.subplots(2, 2, layout='compressed')

for i, ax in enumerate(axs.flat):
    ax.imshow(images[i], cmap='gray')
    ax.set_title(f'Image {i+1}')
    ax.axis('off')

# Connect the click event
main_fig.canvas.mpl_connect('button_press_event', on_click)

plt.show()