Show plots in the Plots pane and in an interactive backend simultaneously

The plot panel is useful but I have a few problems with it:

• The images are saved by default as low quality png. I would prefer svg and therefore need to:

from IPython.display import set_matplotlib_formats
set_matplotlib_formats('svg')

before plotting. I think it would be nice to have a setting in the plot widget to choose one of 'png', 'retina', 'jpeg', 'svg', 'pdf'.

• The images are not shown if the backend is interactive (For example qt5). I think it would be great if there was a button to change the backend between inline and interactive, and if the interactive backend plots where shown in the widget. This works by:

%matplotlib qt5
from IPython.core.display import display
import matplotlib.pyplot as plt
fig = plt.figure()
ax = fig.add_subplot()
ax.plot([1, 2, 4])
display(fig)

But this also displays the plot inline. One implementation could patch the figure show method and get the image data by:

%matplotlib qt5
import matplotlib.pyplot as plt
fig = plt.figure()
ax = fig.add_subplot()
ax.plot([1, 2, 4])
format = get_ipython().kernel.shell.display_formatter.format
format_dict, md_dict = format(fig)

and then send format_dict by the comm channel.

[impact27]

[impact27]

I can do the communication with the kernel: