detect and display plots in tuple/list/arrays

[gdementen]

The editor detects return values of type "matplotlib.axes._subplots.AxesSubplot" and displays them, but some functions (notably Pandas "hist" method) or any plot with argument subplots=True produces an array of AxesSubplot and are thus not displayed. This is unfortunate and confusing for our users.

>>> a = ndtest('sex=M,F;year=2000..2017')
>>> # this should work out of the box
>>> # a.plot(subplots=True)
>>> p = a.plot(subplots=True)
>>> p[0]
<matplotlib.axes._subplots.AxesSubplot at 0x20a9900ff98>
>>> type(p)
numpy.ndarray
>>> p.shape
(2,)
>>> type(p[0])
matplotlib.axes._subplots.AxesSubplot

We need to avoid examining all elements of arrays systematically, but I think checking the first element of arrays should be enough anyway.

[gdementen]

It also needs to work for arrays with ndim > 1. For example:

a = ndrange('sex=M,F;nat=BE,FO;year=2000..2017')
a.plot(subplots=True)

[alixdamman]

ndim > 1:

  a = ndtest('sex=M,F;nat=BE,FO;year=2000..2017')
  a.plot(subplots=True, layout=(2, 2))