如何使用 matplotlib 将 inset_axes 添加到子图

Question

I am trying to plot a number of subplots in matplotlib, each subplot should have an inset axes. I can get the code examples to work for a single axis with a inset axes added using mpl_toolkits.axes_grid.inset_locator.inset_axes() , and I can plot subplots fine without the inset axes, but when trying to do the same for subplots in a loop, I get TypeError: 'AxesHostAxes' object is not callable on the second subplot . This seems a bit strange that it should work when number_of_plots is ==1, but not >1. How should I be doing it, or is it a bug? ( matplotlib.__version__ is '1.5.1')

from matplotlib import pyplot as plt
from mpl_toolkits.axes_grid.inset_locator import inset_axes
import numpy as np
x = np.linspace(0, 2 * np.pi, 100)
y = np.sin(x)
n_row, n_col = 4, 4
fig = plt.figure(1,(10,10))
#number_of_plots = 1 Works!
number_of_plots = n_row * n_col # Does not work!
for idx in range(number_of_plots):
    ax = fig.add_subplot(n_row, n_col, idx + 1)
    ax.plot(x, y)
    inset_axes = inset_axes(ax,
                            width="30%", # width = 30% of parent_bbox
                            height="30%", # height : 1 inch
                            )

Answer 1

For future readers: This post shows methods to create insets in matplotlib.

---

Concrete answer to the question here: It's not a bug. You are redefining inset_axes .

Before the line inset_axes = inset_axes(...) , inset_axes is a function from mpl_toolkits.axes_grid.inset_locator . After that, inset_axes is the return of that function, which is an AxesHostAxes .

The general advice is of course: Never call a variable by the same name as a function you import or use in your code.

The concrete solution:

 ax_ins = inset_axes(ax, width="30%", height="30%") 

Answer 2

Good looking out big dawg, that helped me with a similar issue.