如何从自定义函数中返回可后续绘制或用于网页的子图（subplot）图形？

Hey there! Let's work through this together—returning your matplotlib/seaborn figures for later use is totally straightforward once you know which objects to pass back. First, let's fix up your example code (it has a couple small syntax issues we'll sort out), then walk through exactly how to return the figure for reuse.

Step 1: Clean Up Your Base Code #

First, a quick note on your current code:

• plt.subplots(1, 2) creates a 1-row, 2-column grid, so ax is a 1-dimensional array (not 2D). You don't need ax[0][0]—it should be ax[0] for the first subplot, ax[1] for the second.
• You missed a closing parenthesis on sns.barplot(), and newer seaborn versions require explicit keyword arguments (x=x, y=y) instead of positional ones.
• Pie plots in matplotlib expect values first, then labels (I adjusted that in the example below to avoid errors).

Here's the cleaned-up base code:

import matplotlib.pyplot as plt
import seaborn as sns

def plot_g(x, y):
    # Create figure and 1x2 subplot grid
    fig, ax = plt.subplots(1, 2, figsize=(18, 14))
    
    # Bar plot on first subplot
    sns.barplot(x=x, y=y, ax=ax[0])
    # Pie plot on second subplot
    ax[1].pie(y, labels=x)
    
    # Return the figure object here!
    return fig

Step 2: Return the Right Object #

The fig variable (from plt.subplots()) is the entire container for your graphic—it holds all your subplots, styling, and layout. Returning this gives you full control over the plot later:

• Save it to a file (PNG, SVG, etc.)
• Display it in a notebook or desktop app
• Modify elements like titles, labels, or colors after calling the function
• Convert it to a format usable in web apps

Step 3: Use the Returned Figure #

Once you return fig, you can use it like this:

# Call your function and store the returned figure
my_plot = plot_g(x=["A", "B", "C"], y=[15, 30, 22])

# Example 1: Save the figure to a file (with tight cropping to avoid cut-off labels)
my_plot.savefig("my_combined_plot.png", dpi=300, bbox_inches="tight")

# Example 2: Show the figure (works in Jupyter notebooks or desktop Python environments)
plt.show()

# Example 3: Modify subplots after the fact
all_subplots = my_plot.axes  # Get list of all subplots from the figure
all_subplots[0].set_title("My Bar Chart")  # Add title to first subplot
all_subplots[1].set_title("My Pie Chart")  # Add title to second subplot
my_plot.savefig("updated_plot.png")

Bonus: Return Axes for Direct Subplot Control #

If you want to work directly with individual subplots instead of the whole figure, you can return ax (the array of subplots) too—or both objects:

def plot_g(x, y):
    fig, ax = plt.subplots(1, 2, figsize=(18, 14))
    sns.barplot(x=x, y=y, ax=ax[0])
    ax[1].pie(y, labels=x)
    return fig, ax  # Return both the figure and subplot axes

# Usage
fig, ax = plot_g(x=["A", "B", "C"], y=[15, 30, 22])
ax[0].set_xlabel("Categories")  # Modify first subplot directly
fig.savefig("final_plot.png")

That's all there is to it! Returning fig (and/or ax) lets you reuse and tweak your plot however you need, whether that's for additional edits, saving, or embedding in a web interface.

内容的提问来源于stack exchange，提问作者Darkstar Dream