Re: [Matplotlib-users] Copying a figure instance?

Thanks Tony, I think this will work! 

Cheers, 
Joe
On Mar 31, 2012, at 11:33 AM, Tony Yu wrote:

> 
> 
> On Fri, Mar 30, 2012 at 12:25 PM, eoj <jos...@gm...> wrote:
> 
> Basically the problem is like this. I have a relatively expensive (time wise)
> figure to create, specifically a map with lots of detail in it. On top of
> that, I'm  making an inset set of axes to highlight some regions. I'm having
> to recreate the base of the map, which is static, every time I want to make
> an inset for a different region which is taking forever. What I want to do
> is something like this:
> 
> fig = figure()
> 
> #do stuff that makes the expensive figure
> 
> for region in regions:
>    fig2 = copy.copy(fig)
>    ax2 = fig.add_axes([0.1, 0.15, 0.25, 0.25])
>    ax2.fill(x1,y1)
>    savefig(region_name)
>    close()
> 
> 
> 
> The problem is that the clf() clear figure function seems to clear the base
> of the map off, even if I try to make a copy of it inside a loop that is
> generating the insets. Also, close() seems to close fig and not fig2. Does
> this rambling make any sense, and if so, any suggestions?
> 
> 
> If all your additions are confined to the inset axes, you can use `fig.delaxes` to remove the axes from the figure (see example below). If you want to save and restore, as you describe above, you should (in principal) be able to do so with `canvas.copy_from_bbox` and `canvas.restore_region`, but I couldn't get it to work (see bottom of the animation cookbook).
> 
> Cheers,
> -Tony
> 
> # ~~~ example
> import numpy as np
> import matplotlib.pyplot as plt
> 
> fig, ax = plt.subplots()
> ax.imshow(np.random.uniform(size=(10, 10)))
> 
> ax_inset = fig.add_axes([0.3, 0.3, 0.2, 0.2])
> ax_inset.plot([0, 1])
> plt.savefig('plot0')
> 
> fig.delaxes(ax_inset)
> ax_inset = fig.add_axes([0.5, 0.5, 0.2, 0.2])
> ax_inset.plot([1, 0])
> plt.savefig('plot1')
> # ~~~
> 

=======================
Joseph D. Meiring
Department of Astronomy
LGRT 517
University of Massachusetts
Amherst MA 01003
=======================