Re: [matplotlib-devel] Strange bug when saving animations (with streamplot)

[Tony Yu <ts...@gm...>]

On Sun, Jun 24, 2012 at 10:02 AM, Tony Yu <ts...@gm...> wrote:

> I'm having a tough time figuring this out: Saving an animation seems to
> hang when using `streamplot`. The exact same animation runs without issue
> when calling show.
>
> On my system, the example below hangs consistently at frame 173. However,
> the number of saved frames (before hanging) varies with density of stream
> lines (i.e. number of line segments in streamplot): More frames saved for
> lower density.
>
> Seems to be a memory or file-size issue, but
>
>    * Memory use doesn't seem to grow. (this suggests that ffmpeg adds
> frames to disk as opposed to memory.)
>    * Saving the animation up to the frame just before it hangs gives
> modest (~1MB) videos.
> Maybe this is a limitation of `ffmpeg`?
>
> Can someone confirm that the example hangs on their system? (Somehow, I
> often run into bugs that are not reproducible.)
>
> In case this is system dependent:
> * OSX 10.6
> * Python 2.6.1 (system install)
> * matplotlib master
> * ffmpeg 0.10
>
> Simple plots seem to work fine, but I should probably try to reproduce
> using something other than streamplot (maybe create the equivalent number
> of line and arrow collections).
>
> -Tony
>
> #~~~ Example
> import numpy as np
> import matplotlib.pyplot as plt
> import matplotlib.animation as animation
>
> fig, ax = plt.subplots()
>
> # do nothing function that prints frame
> def update(frame_num):
>     print frame_num
>     return []
>
> Y, X = np.mgrid[:1:100j, :1:100j]
> U = np.ones_like(X)
> V = np.ones_like(X)
> ax.streamplot(X, Y, U, V, density=1)
>
> ani = animation.FuncAnimation(fig, update, save_count=500)
>
> # save hangs at frame 173 (this probably varies by system, if it's
> reproducible).
> ani.save('animation.avi')
> # show animates all frames w/o issue.
> #plt.show()
>
>
>
I finally had some time to revisit this issue. It seems that subprocess
PIPEs have fairly limited buffers [1] (also see docs for
`subprocess.Popen.communicate` [2]) and ffmpeg tends to generate a decent
amount of output to stderr. A simple solution is to redirect stderr to a
temporary file. This change fixes the issue on my system.

If this bug is reproducible, I can submit a PR. The temporary file here
could be created using tempfile so it gets cleaned up automatically, or
maybe a more permanent log file for debugging. Thoughts?

-Tony

[1]
http://thraxil.org/users/anders/posts/2008/03/13/Subprocess-Hanging-PIPE-is-your-enemy/

[2]
http://docs.python.org/library/subprocess.html#subprocess.Popen.communicate