This has been on my radar for a while, but Iikely won't have time to think about this for several months. If you want to take this on, I think that would be great.

As for difficulty, I'd say moderate to hard. There would need to be a design discussion about the proper way to hook into matplotlib and select a backend. This needs to happen using matplotlib's OO interface, outside of pyplot, unless we can figure out a way to not crash Python when we try to use an interactive backend on a headless session.

Once that happens, it should only be necessary to select the correct FigureCanvas subclass and attach it to the plot object.

FWIW, there's an issue for this here:

https://bitbucket.org/yt_analysis/yt/issue/660/support-for-interactive-mpl-backends

On Tue, Nov 25, 2014 at 8:04 PM John ZuHone <jzuhone@gmail.com> wrote:

Hi gang,

I got an inquiry about using the PlotWindow plot show() method outside of the notebook. How hard would it be to extend this to other matplotlib backends, so that a window pops up on the screen when you want to see a plot, but aren't in the notebook?

Obviously matplotlib does this already, so it's a matter of figuring out how to get it hooked up.

Best,

John Z

Sent from John ZuHone's iPad
_______________________________________________
yt-dev mailing list
yt-dev@lists.spacepope.org
http://lists.spacepope.org/listinfo.cgi/yt-dev-spacepope.org