MWE for plot widget

[ebatz]

The following code used to work in previous versions,

import ipympl
import matplotlib.pyplot as plt

import ipywidgets as widgets
from IPython.display import display

aFig = plt.figure()
aBox = widgets.Box(children=[aFig.canvas,])

display(aBox)

and the figure content could then be updated later by plotting onto its axes and subsequently calling canvas's draw(). After upgrading to the current version this snippet fails with

TraitError: Element of the 'children' trait of a Box instance must be a Widget, but a value of <matplotlib.backends.backend_agg.FigureCanvasAgg object at ...> <class 'matplotlib.backends.backend_agg.FigureCanvasAgg'> was specified.

Using the %matplotlib ipympl magic command, there is no such error message, but no plot gets rendered. Based on #39 my understanding was that the only difference between importing ipympl and using the magic command would be when the plot gets shown, but now I cannot seem to get either to work.

What is the preferred way to set up an interactive figure and retain control over its positioning?

[ebatz]

Using the %matplotlib widget magic the code snippet does not throw an error, but it also does not produce a plot,

%matplotlib widget

import matplotlib.pyplot as plt
import ipywidgets as widgets

from IPython.display import display

fig = plt.figure()
fig.add_subplot(1,1,1).plot([1,2,3])

fig.canvas.draw()

box = widgets.Box(children=[fig.canvas,])
display(box)

fig.canvas.draw()

What's the missing piece?

[ebatz]

Thanks, commits fba1226 and 0b775d1 now demonstrate the use of the canvas as a widget in the example notebook.

The reason that the previous code snippet did not produce a plot remains unclear to me -- maybe similar to #45 -- but it works now after upgrading all the relevant modules.