我试图使用matplotlib.pyplot.streamplot来显示使用magpylib生成的磁场。我从他们的一个例子中改编了下面的内容，以显示具有受控间距的完整线条，但我有一些问题，即一些理论上闭合的磁力线实际上并不闭合，而是螺旋形的，可能是由于数值不稳定或近似，来自matplotlib，从magpylib或从两者。有没有一种方法可以防止不间断的流线螺旋，或者更好的是，告诉matplotlib在可能的情况下关闭流线本身？
代码：

import matplotlib.pyplot as plt
import numpy as np
import magpylib as magpy

# Create a Matplotlib figure
fig, ax = plt.subplots(figsize=(10, 10))

# Create an observer grid in the xz-symmetry plane
ts = np.linspace(-5,5,200)
grid = np.array([[(x,0,z) for x in ts] for z in ts]) # slow Python loop

# Compute the B-field of a cube magnet on the grid
cube = magpy.magnet.Cuboid(magnetization=(0, 0, -1000), dimension=(2, 2, 2))
B = cube.getB(grid)
x, y, u, v = grid[:,:,0], grid[:,:,2], B[:, :, 0], B[:, :, 2],

# Mask for breaking the streamlines
u[98:102, 80:120] = np.nan
v[98:102, 80:120] = np.nan

top_seed_points = np.array([np.linspace(-1.0, +1.0, 21), np.ones(21)])[:,:-3]
bottom_seed_points = np.array([np.linspace(-1.0, +1.0, 21), -np.ones(21)])[:,3:]
left_seed_points = np.array([-np.ones(4), 1-(0.05+(np.linspace(0, 1, 4, endpoint=False))**2)])[:,:3]
right_seed_points = np.array([np.ones(4), (0.05+(np.linspace(0, 1, 4, endpoint=False))**2-1)])[:,:3]
seed_points = np.concatenate((top_seed_points, bottom_seed_points, left_seed_points, right_seed_points[:, :3]), axis=1)

# Display the B-field with streamplot using log10-scaled
# color function and linewidth
splt = ax.streamplot(
    x, y, u, v,
    start_points=seed_points.T,
    broken_streamlines=False,
)
ticks = np.array([3,10,30,100,300])

# Outline magnet boundary
ax.plot([1, 1, -1, -1, 1], [1, -1, -1, 1, 1], "k-", lw=2)
ax.set_xlim([-5, 5])
ax.set_ylim([-5, 5])

# Figure styling
ax.set(
    xlabel="x-position (mm)",
    ylabel="z-position (mm)",
)

plt.tight_layout()
plt.show()

字符串
结果：

我试图打破螺旋使用面具，但一些流线设法使它通过面具地区（我会认为是一个错误）和螺旋无论如何。
注意：如果我以这种方式进行掩蔽，结果是相同的：

# Mask for breaking the streamlines
mask = np.zeros(u.shape, dtype=bool)
mask[98:102, 80:120] = True
u = np.ma.array(u, mask=mask)
v = np.ma.array(v, mask=mask)

型