Alphashape与PolygonPatch基础示例运行报错，请教原因

报错信息 #

File "/anaconda3/lib/python3.8/site-packages/descartes/patch.py", line 87, in PolygonPatch
    return PathPatch(PolygonPath(polygon), **kwargs)
File "/anaconda3/lib/python3.8/site-packages/descartes/patch.py", line 62, in PolygonPath
    vertices = concatenate([
File "/anaconda3/lib/python3.8/site-packages/descartes/patch.py", line 63, in <listcomp>
    concatenate([asarray(t.exterior)[:, :2]] +
IndexError: too many indices for array: array is 0-dimensional, but 2 were indexed

运行代码 #

import numpy as np
from descartes import PolygonPatch
import matplotlib.pyplot as plt
import alphashape


points_2d = [(0., 0.), (0., 1.), (1., 1.), (1., 0.),
(0.5, 0.25), (0.5, 0.75), (0.25, 0.5), (0.75, 0.5)]

fig, ax = plt.subplots()
ax.scatter(*zip(*points_2d))
alpha_shape = alphashape.alphashape(points_2d, 0.)
ax.add_patch(PolygonPatch(alpha_shape, alpha=0.2))
plt.savefig(f"./test.png")

原因分析 #

这个错误的核心是descartes库与新版shapely/alphashape的兼容性问题：

• descartes是已停止维护的库，它的PolygonPatch方法无法正确处理新版shapely返回的几何对象结构，当尝试读取几何对象的exterior属性时，得到的是0维数组，触发索引报错。
• 即使alpha=0时理论上应返回凸包多边形，但库版本不匹配会导致返回的几何对象类型或结构不符合descartes的预期。

解决办法 #

方法1：替换descartes，使用shapely原生绘图接口 #

放弃descartes，改用shapely自带的plot方法，代码修改如下：

import numpy as np
import matplotlib.pyplot as plt
import alphashape
from shapely.plotting import plot_polygon

points_2d = [(0., 0.), (0., 1.), (1., 1.), (1., 0.),
(0.5, 0.25), (0.5, 0.75), (0.25, 0.5), (0.75, 0.5)]

fig, ax = plt.subplots()
ax.scatter(*zip(*points_2d))
alpha_shape = alphashape.alphashape(points_2d, 0.)
# 使用shapely的plot_polygon绘制
plot_polygon(alpha_shape, ax=ax, alpha=0.2, facecolor='lightblue')
plt.savefig("./test.png")

方法2：使用matplotlib原生Polygon类 #

手动提取多边形的坐标，用matplotlib的Polygon创建补丁：

import numpy as np
import matplotlib.pyplot as plt
from matplotlib.patches import Polygon
import alphashape

points_2d = [(0., 0.), (0., 1.), (1., 1.), (1., 0.),
(0.5, 0.25), (0.5, 0.75), (0.25, 0.5), (0.75, 0.5)]

fig, ax = plt.subplots()
ax.scatter(*zip(*points_2d))
alpha_shape = alphashape.alphashape(points_2d, 0.)
# 提取多边形 exterior 的坐标
coords = list(alpha_shape.exterior.coords)
ax.add_patch(Polygon(coords, alpha=0.2, facecolor='lightblue'))
plt.savefig("./test.png")

方法3：降级到兼容的库版本 #

如果必须使用descartes，可以安装旧版本的shapely（2.0以下版本），因为descartes仅支持旧版shapely：

pip install shapely==1.8.5 descartes==1.1.0

内容的提问来源于stack exchange，提问作者Mik Kim