Improve docs related to reading from bytes

I’m trying to save bytes as frames of an animation of a plot but

import matplotlib.pyplot as plt
import imageio

plt.plot(1, 1)
fig = plt.gcf()
fig.canvas.draw()
data = fig.canvas.tostring_rgb() # These are bytes
imageio.imread(data)

File "<string>", line unknown
SyntaxError: cannot handle 76-bit layers

I also tried converting it to a numpy array of uint8s but imread just spits the data back into my terminal and doesn’t do anything. Is this the way to read in raw bytes? Is something missing from the docs?

Thanks

[mef51]

Ah okay, well I’ve figured it out! Just had to reshape the data a little bit before giving it to imageio. Here’s a little snippet that saves an animation:

import matplotlib.pyplot as plt
import numpy as np
import imageio
from numpy import sin, pi

print('Working...')
time = np.linspace(0, 10, 100)
with imageio.get_writer('test.gif', mode='I', fps=24) as writer:
	for t in time:
		x = np.linspace(-2*pi,2*pi, 100)
		plt.plot(x, (x/(2*pi))*sin(t+x))
		plt.grid()
		plt.xlim(-2*pi, 2*pi), plt.ylim(-1.2, 1.2)
		fig = plt.gcf()
		fig.canvas.draw()
		data = fig.canvas.tostring_rgb()
		rows, cols = fig.canvas.get_width_height()

		imarray = np.fromstring(data, dtype=np.uint8).reshape(cols, rows, 3)
		writer.append_data(imarray)
		plt.close()
print('Done.')

[almarklein]

Ah I see. I was confused by the imageio.imread(data). Glad you got it working. Nice animation!