How to require equal axes in matplotlib

Question

I was used to using a version of python where the code below generated exactly proportional axes.

    plt.ylabel("y(Km)")
    plt.gca().set_aspect('equal')
    ax = plt.axes()
    ax.set_facecolor("black")
    circle = Circle((0, 0), 2, color='dimgrey')
    plt.gca().add_patch(circle)
    plt.axis([-1 / u1 , 1 / u1 , -1 / u1 , 1 / u1 ])

When I switched computers and started using python 3.7, the same code started to generate a desconfigured picture. Why did this happen and how can I resolve it? The before and after photos are below.

Answer 1

Your code is creating two overlapping axes (as you point out in the red circles in the second image). You can see that also by running the following lines at the end of your code:

print(plt.gcf().get_axes()) # (get current figure)

Which gives:

[<AxesSubplot:ylabel='y(Km)'>, <AxesSubplot:>]

This happens because while you have already an axes, you add another by the line ax = plt.axes(). From documentation:

Add an axes to the current figure and make it the current axes.

Maybe you meant ax = plt.gca() ?

Anyway, tidying up the code:

import matplotlib.pyplot as plt
from matplotlib.patches import Circle

u1 = 0.05

fig, ax = plt.subplots() # figure fig with one axes object ax
ax.set_facecolor("black")
ax.set_aspect("equal")
circle = Circle((0, 0), 2, color='dimgrey')
ax.add_patch(circle)
ax.set_xlim([-1 / u1, 1 / u1 ])
ax.set_ylim([-1 / u1, 1 / u1 ])
ax.set_ylabel("y (km)")

print(fig.get_axes()) # not needed for anything, just for comparison with original code

plt.show()

Which should give: