Creating Reusable AxesSubplot Objects in Matplotlib

Matplotlib provides the Figure.add_subplot method as the standard approach for adding AxesSubplot objects to a figure. While this is effective, there may be scenarios where creating AxesSubplot objects independently of the figure is desirable.

To decouple AxesSubplot creation from figure instances, one can harness the power of passing axes instances to functions. For instance:

def plot(x, y, ax=None):
    if ax is None:
        ax = plt.gca()  # Get the current axes instance (default)
    ax.plot(x, y, 'go')
    ax.set_ylabel('Yabba dabba do!')

In this example, the plot function takes an optional axes argument, providing flexibility in its usage:

# Create a figure with two subplots
fig1, (ax1, ax2) = plt.subplots(nrows=2)
plot(x, np.sin(x), ax1)  # Use the first axes instance
plot(x, np.random.random(100), ax2)  # Use the second axes instance

# Create a new figure
fig2 = plt.figure()
plot(x, np.cos(x))  # Use the new figure's axes instance

plt.show()

Additionally, axes instances can be appended to existing figures, allowing for reuse:

import matplotlib.pyplot as plt

# Create an axes instance
ax = plt.gca()
ax.plot(range(10))

# Create a new figure
fig2 = plt.figure()
fig2.axes.append(ax)  # Add the existing axes instance to the new figure

plt.show()

While it may be possible to further customize the axes instance to fit specific "shapes," the ease of passing around figures and axes instances or lists of instances is generally more practical and efficient for complex scenarios.