ArtistAnimation using contourf for model plots

Question

Does ArtistAnimation work when using the "contourf" method? I'm trying to build an MP4 animation of a model forecast.

I'm using an approach similar that is similar to that shown on one of the MetPy Monday video series where you append the plots to an empty artist array (for radar plots). Radar/satellite plots seem to build the animation successfully because they use "pcolormesh", but the same doesn't hold true for model plots where I'm using "contourf".

When executing this code in JupyterLab,

plt.rcParams['animation.html'] = 'jshtml'
anim = ArtistAnimation(fig, artists, interval=100, blit=False)
anim

I see this at the end of the stack trace:

AttributeError: 'QuadContourSet' object has no attribute 'set_visible'

An old thread on stackoverflow made an attempt at solving this issue: https://github.com/matplotlib/matplotlib/issues/6139

When implementing these fixes, the set_visible attribute error no longer appears, but this appears instead:

AttributeError: 'QuadContourSet' object has no attribute 'get_figure'

Not sure where to go from here. If there's good news, I am seeing all of the data from each time step plot all at once on the map (so the data retrieval is OK), but something is up with ArtistAnimation and how it works with contourf method. Any guidance or helpful hints would be awesome!

score 3 · Answer 1 · answered May 01 '20 at 22:13

For reasons I don't understand, the object directly returned by contour, QuadContourSet, is not an Artist, unlike many other plot commands--which is why it doesn't have a set_visible method. After a bit of digging, QuadContourSet has a .collections attribute that is a list of all the line collections that are drawn, so you want to use that in something like this, which is modified from matplotlib's dynamic_image.py example:

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

fig, ax = plt.subplots()

def f(x, y):
    return np.sin(x) + np.cos(y)

x = np.linspace(0, 2 * np.pi, 120)
y = np.linspace(0, 2 * np.pi, 100).reshape(-1, 1)

artists = []
for i in range(60):
    X, Y = np.meshgrid(x, y)
    cs = ax.contour(X, Y, f(x + i * np.pi/15., y + i*np.pi/20))
    artists.append(cs.collections)

ani = animation.ArtistAnimation(fig, artists, interval=50, blit=True,
                                repeat_delay=1000)

plt.show()

ArtistAnimation using contourf for model plots

1 Answers1