plt.plot(...) doesn't show a window. what am I missing?

I am practicing with open cv on VScode.

I am used to display images by using the following structure:

import cv2
import numpy as np
import matplotlib.pyplot as plt

puppy = cv2.imread('./DATA/00-puppy.jpg')
horse = cv2.imread('./DATA/horse.jpg')
rainbow = cv2.imread('./DATA/rainbow.jpg')

# "inplace" operations on these images such as
# puppy = cv2.cvtColor(puppy, cv2.COLOR_BGR2RGB)
# etc...

img_to_show_1 = puppy
img_to_show_2 = horse
img_to_show_3 = rainbow

# ---> beginning of the "while True ... - if 0xFF == 27: break" structure

cv2.namedWindow(winname='window_1')
cv2.namedWindow(winname='window_2')
cv2.namedWindow(winname='window_3')


while True:
	cv2.imshow('window_1', img_to_show_1)
	cv2.imshow('window_2', img_to_show_2)    
	cv2.imshow('window_3', img_to_show_3)

	if cv2.waitKey(1) & 0xFF == 27: # stop loop when 'esc' key is pressed
		break

cv2.destroyAllWindows()

# ---> end of the "while True ... - if 0xFF == 27: break" structure

Now I have calculated the histogram of the values of an image stored in puppy variable

hist_values = cv2.calcHist([puppy], channels=[0], mask=None, histSize=[256], ranges=[0,256])

On jupyter notebooks, this can be showed by simply doing:

plt.plot(hist_values)

But on VScode, this does not show anything.

Is there a way to show this image on VScode and without intalling any other graphical extention?

Is there a way to exploit the same "while True ... - if 0xFF == 27: break" structure, in order to have this image also disappear with the others as I press 'esc' key?

I found a workaround to do it.

This is based on saving the image with opencv module method and to reload it with the matplotlib method:

import cv2
import numpy as np
import matplotlib.pyplot as plt

puppy = cv2.imread('./DATA/00-puppy.jpg')
horse = cv2.imread('./DATA/horse.jpg')
rainbow = cv2.imread('./DATA/rainbow.jpg')

# "inplace" operations on these images such as
# puppy = cv2.cvtColor(puppy, cv2.COLOR_BGR2RGB)
# etc...

hist_values = cv2.calcHist([puppy], channels=[0], mask=None, histSize=[256], ranges=[0,256])

plt.plot(hist_values)
plt.plot([1, 2])
mypath = 'my/path/hist_values.jpg'
plt.savefig(mypath)

hist_values_img = cv2.imread(mypath)

import os
try: 
    os.remove(hist_values_path)
except: 
    print("could not remove {}".format(hist_values_path))

img_to_show_1 = puppy
img_to_show_2 = hist_values_img
img_to_show_3 = rainbow

# ---> beginning of the "while True ... - if 0xFF == 27: break" structure

cv2.namedWindow(winname='window_1')
cv2.namedWindow(winname='window_2')
cv2.namedWindow(winname='window_3')

while True:
	cv2.imshow('window_1', img_to_show_1)
	cv2.imshow('window_2', img_to_show_2)    
	cv2.imshow('window_3', img_to_show_3)
	plt.show()

	if cv2.waitKey(1) & 0xFF == 27: # se premo esc
		break

cv2.destroyAllWindows()

# ---> end of the "while True ... - if 0xFF == 27: break" structure

In this way I am not mixing GUIs and as I press 'esc', the histogram image closes together with the other images.

Credits to @ChristophRackwitz

Also, the snippet to convert a matplotlib-format element to a opencv-format element can be shrinked to this function:

def convert_matplotlib_to_opencv_img(matplotlib_format_img, clear=False, transitory_img_path='my_transitory_img.jpg'):

	if clear:
		plt.clf()

	plt.plot(matplotlib_format_img)

	try:
		plt.savefig(transitory_img_path)
	except: 
		raise OSError("could not save img to {}".format(transitory_img_path))

	cv2_imread_format_img = cv2.imread(transitory_img_path)

	import os
	try: 
		os.remove(transitory_img_path)
	except: 
		print("could not remove {}".format(transitory_img_path))

	return cv2_imread_format_img