Set radial axis on Matplotlib polar plots

Question

I'm plotting an azimuth-elevation curve on a polar plot where the elevation is the radial component. By default, Matplotlib plots the radial value from 0 in the center to 90 on the perimeter. I want to reverse that so 90 degrees is at the center. I tried setting the limits with a call to ax.set_ylim(90,0) but this results in a LinAlgError exception being thrown. ax is the axes object obtained from a call to add_axes.

Can this be done and, if so, what must I do?

Edit: Here is what I'm using now. The basic plotting code was taken from one of the Matplotlib examples

# radar green, solid grid lines
rc('grid', color='#316931', linewidth=1, linestyle='-')
rc('xtick', labelsize=10)
rc('ytick', labelsize=10)

# force square figure and square axes looks better for polar, IMO
width, height = matplotlib.rcParams['figure.figsize']
size = min(width, height)
# make a square figure
fig = figure(figsize=(size, size))
ax = fig.add_axes([0.1, 0.1, 0.8, 0.8], projection='polar', axisbg='#d5de9c')

# Adjust radius so it goes 90 at the center to 0 at the perimeter (doesn't work)
#ax.set_ylim(90, 0)

# Rotate plot so 0 degrees is due north, 180 is due south

ax.set_theta_zero_location("N")

obs.date = datetime.datetime.utcnow()
az,el = azel_calc(obs, ephem.Sun())
ax.plot(az, el, color='#ee8d18', lw=3)
obs.date = datetime.datetime.utcnow()
az,el = azel_calc(obs, ephem.Moon())
ax.plot(az, el, color='#bf7033', lw=3)

ax.set_rmax(90.)
grid(True)

ax.set_title("Solar Az-El Plot", fontsize=10)
show()

The plot that results from this is

Answer 1

I managed to put he radial axis inverted. I had to remap the radius, in order to match the new axis:

fig = figure()
ax = fig.add_subplot(1, 1, 1, polar=True)

def mapr(r):
   """Remap the radial axis."""
   return 90 - r

r = np.arange(0, 90, 0.01)
theta = 2 * np.pi * r / 90

ax.plot(theta, mapr(r))
ax.set_yticks(range(0, 90, 10))                   # Define the yticks
ax.set_yticklabels(map(str, range(90, 0, -10)))   # Change the labels

Note that is just a hack, the axis is still with the 0 in the center and 90 in the perimeter. You will have to use the mapping function for all the variables that you are plotting.