sgrid() and zgrid() call matplotlib.pyplot.figure()

While debugging some code I noticed that examples/pvtol-nested.py was producing a blank figure instead of a pole-zero map that was expected. I chased this down tto the fact that the sgrid() function calls matplotlib.pyplot.figure(). This is in contrast to all other plotting routines and the consequence is that you get a new set of axes. This causes the following code (from examples/pvtol-nested.py) not to work as expected:

figure(10); clf();
(P, Z) = pzmap(T, Plot=True)
print("Closed loop poles and zeros: ", P, Z)

# Gang of Four
figure(11); clf();
gangof4(Hi*Po, Co);

Figure 10, which should have the output of pzmap will be blank since pzmap calls figure() (through sgrid) and creates a new set of axis (figure #11). Figure #11 will then get erased (away goes the pole/zero map) and be replace by the Gang of 4.

[bnavigator]

You already mentioned the Pandas way of providing .plot() methods to the data objects. I would consider this together with kwargs such as ax= and grid='s' to be the most pythonic.

If you want to retain the MATLAB way of calling sgrid() after creating the plot, I have added a comment at StackOverflow with another idea to add the grid axis.

Using the pylab API in matplotlib has been deprecated for a reason, though. Likewise python-control should not try to stick to that coding paradigm either.

[billtubbs]

B.t.w. I haven’t forgotten about this. Had to focus on another project. Hopefully the delay is not impacting the project.