如何添加FancyBbox作为Matplotlib图例？

问题 #

想要在Matplotlib图例中添加一系列FancyBbox，参考官方Fancybox示例，但编写的代码里图例中的FancyBbox形状和图中显示的不一致，需要修正代码。

原代码如下：

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

styles = ['darrow', 'round4', 'roundtooth', 'square']
coords = [(0.2, 0.2), (0.2, 0.8), (0.8, 0.2), (0.8, 0.8)]
colors = ['red', 'green', 'blue', 'orange']
weights = ['bold', 'ultralight', 'normal', 'heavy']

fig, ax = plt.subplots()

patches = []
for style, color, weight, (x, y) in zip(styles, colors, weights, coords):
    ax.text(x, y, style, fontsize=12, fontweight=weight, ha='center', va='center', bbox=dict(boxstyle=style, facecolor=color, alpha=0.5, edgecolor='black'))
    legend_patch = FancyBboxPatch((0, 0), width=1, height=1, boxstyle=style, facecolor=color, edgecolor='black', alpha=0.5, transform=ax.transAxes, label='Bbox')
    patches.append(legend_patch)

ax.legend(patches, styles, loc='center', handleheight=2)

plt.show()

解决方案 #

问题出在图例中FancyBboxPatch的坐标变换和尺寸设置上，直接绑定坐标轴变换会导致形状被错误缩放，默认尺寸也不匹配图例容器。

修正后的代码如下：

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

styles = ['darrow', 'round4', 'roundtooth', 'square']
coords = [(0.2, 0.2), (0.2, 0.8), (0.8, 0.2), (0.8, 0.8)]
colors = ['red', 'green', 'blue', 'orange']
weights = ['bold', 'ultralight', 'normal', 'heavy']

fig, ax = plt.subplots()

patches = []
for style, color, weight, (x, y) in zip(styles, colors, weights, coords):
    # 绘制图中的文本与Bbox
    ax.text(x, y, style, fontsize=12, fontweight=weight, 
            ha='center', va='center', 
            bbox=dict(boxstyle=style, facecolor=color, alpha=0.5, edgecolor='black'))
    
    # 创建适配图例的FancyBboxPatch：移除坐标轴变换，调整尺寸让形状居中显示
    legend_patch = FancyBboxPatch((-0.4, -0.3), width=0.8, height=0.6, 
                                  boxstyle=style, facecolor=color, 
                                  edgecolor='black', alpha=0.5)
    patches.append(legend_patch)

# 调整图例手柄参数，确保形状完整展示
ax.legend(patches, styles, loc='center', handlelength=1.5, handleheight=1.2)

plt.show()

修正说明 #

• 移除transform=ax.transAxes：图例的Patch无需绑定坐标轴变换，Matplotlib图例系统会自动处理其坐标空间。
• 调整Patch原点与尺寸：将原点设为(-0.4, -0.3)、宽高设为0.8和0.6，保证形状在图例手柄区域内居中且完整。
• 优化图例参数：通过handlelength和handleheight给不同样式的FancyBbox预留足够显示空间，避免形状被截断。

内容的提问来源于stack exchange，提问作者VERBOSE