Bug report

Bug summary

The smith chart grid is distorted in 2.1.2, correct in 2.0.2.
Updating the polar plot data blanks the plot in 2.1.2, updates properly in 2.0.2.

In both versions when I create the plots/lines I see:

/Users/chrisgmorton/usr/tcad/fusion/prod/lib/python3.6/site-packages/pysmithplot-0.2.0-py3.6.egg/smithplot/smithhelper.py:46: RuntimeWarning: divide by zero encountered in true_divide
/Users/chrisgmorton/usr/tcad/fusion/prod/lib/python3.6/site-packages/pysmithplot-0.2.0-py3.6.egg/smithplot/smithhelper.py:46: RuntimeWarning: invalid value encountered in true_divide
/Users/chrisgmorton/usr/tcad/fusion/prod/lib/python3.6/site-packages/matplotlib/projections/polar.py:58: RuntimeWarning: invalid value encountered in less
mask = r < 0

On the first update of the plots I see:

Users/chrisgmorton/usr/tcad/fusion/prod/lib/python3.6/site-packages/matplotlib/projections/polar.py:58: RuntimeWarning: invalid value encountered in less
mask = r < 0

and the polar plots are then blanked out (not so with 2.0.2).

Code for reproduction

Plots are built with:

self.axes=[]n=self.numberOfPortsfori,jinself.indices:ifi==j:self.axes.append(self.fig.add_subplot(n,n,1+i*2+j,projection='smith'))else:self.axes.append(self.fig.add_subplot(n,n,1+i*2+j,projection='polar'))fork,(i,j)inenumerate(self.indices):self.axes[k].tick_params(which='major',labelsize=10)self.fig.subplots_adjust(top=0.9)

Lines initialized with:

self.lines=[]form,(i,j)inenumerate(self.indices):ifi==j:# Smith ChartS=self.Splt.T[j,i].TZ= (1+S)/(1-S)# normalized impedance on Smith Chartself.lines.append(self.axes[m].plot(Z.T,color='b',marker='',markerfacecolor='b',markeredgecolor='b'))else:# Polar PlotS=self.Splt.T[j,i].Tself.lines.append(self.axes[m].plot(np.angle(S.T),np.absolute(S.T),color='r',marker='',markerfacecolor='r',markeredgecolor='r'))

Lines updated with:

fork,(i,j)inenumerate(self.indices):ifi==j:# Smith ChartS=self.Splt.T[j,i].TZ= (1+S)/(1-S)# normalized impedance on Smith Chart                [line.set_xdata(np.real(Z[idx]))foridx,lineinenumerate(self.lines[k])]                [line.set_ydata(np.imag(Z[idx]))foridx,lineinenumerate(self.lines[k])]else:# Polar PlotS=self.Splt.T[j,i].T                [line.set_xdata(np.angle(S[idx]))foridx,lineinenumerate(self.lines[k])]                [line.set_ydata(np.absolute(S[idx]))foridx,lineinenumerate(self.lines[k])]

Actual outcome

Matplotlib 2.1.2 output, on completion of simulation:

Expected outcome

Matplotlib 2.0.2 output, on completion of simulation:

Matplotlib version

• Operating system: Ubuntu 16.04 and OSX 10.13.3
• Matplotlib version: 2.1.2
• Matplotlib backend: Qt5Agg (Qt5.10)
• Python version: 3.6.4
• PySmithPlot version: 0.2

Matplotlib installed using pip. Qt5 and PyQt5 built from source. Python 3.6.4 built from source.