腌制的Matplotlib 3D缺乏交互功能。

Question

在Windows系统上，当我使用pickle模块保存并重新加载3D matplotlib表面图时，这些图缺乏任何交互功能，例如旋转或缩放。然而，如果我保存并重新加载2D的matplotlib图，我仍然能够与它们进行交互，至少可以使用缩放工具。以下代码重现了成功的2D图案例。

import matplotlib.pyplot as plt
import numpy as np
import pickle


xs = np.linspace(0, 4*np.pi, 100)
ys = np.sin(xs)

fig, ax = plt.subplots()
ax.plot(xs, ys)

with open("plot.pickle", 'wb') as file:
    pickle.dump(fig, file)

plt.show()

plt.close(fig)

with open("plot.pickle", 'rb') as file:
    fig = pickle.load(file)

plt.show()

下面的代码是一个小例子，重新创建了3D绘图问题。

import matplotlib.pyplot as plt
from mpl_toolkits.mplot3d import Axes3D
import numpy as np
import pickle


xs = np.linspace(0, 1, 100)
ys = np.linspace(0, 4*np.pi, 100)
zs = np.zeros((100, 100))

for i, amp in enumerate(xs):
    for j, phase in enumerate(ys):
        zs[j, i] = amp*np.sin(phase)

xs, ys = np.meshgrid(xs, ys)
fig, ax = plt.subplots(subplot_kw={'projection': '3d'})
ax.plot_surface(xs, ys, zs)

with open("plot.pickle", 'wb') as file:
    pickle.dump(fig, file)

plt.close(fig)

with open("plot.pickle", 'rb') as file:
    fig = pickle.load(file)

plt.show()

我该如何修改代码，使得我的3D plot有交互式的旋转和缩放功能？如果3D plot无法实现此功能，是否有其他方法可以保存（而不必在窗口中显示图形）并重新加载交互式的3D matplotlib曲面图？

- scruffaluff

---

这个功能请求：https://github.com/matplotlib/matplotlib/issues/10843。 - ImportanceOfBeingErnest

Answer 1

我找到的不是解决方案，而是一个变通方法。 这个变通方法是将绘制图形所需的所有数据以及绘制它的函数保存在pickle文件中。

为了做到这一点，我需要使用dill而不是picke，正如@JoshRosen在他的回答Is there an easy way to pickle a python function (or otherwise serialize its code)?中建议的那样。

创建pickle：

import matplotlib.pyplot as plt
from mpl_toolkits.mplot3d import Axes3D
import numpy as np
import dill as pickle


def plotfun(plotdata):
    ax = plt.gca()
    ax.plot_surface(plotdata['xs'], plotdata['ys'], plotdata['zs'])


xs = np.linspace(0, 1, 100)
ys = np.linspace(0, 4*np.pi, 100)
zs = np.zeros((100, 100))
for j, phase in enumerate(ys):
    zs[j, :] = xs*np.sin(phase)
xs, ys = np.meshgrid(xs, ys)

fig, ax = plt.subplots(subplot_kw={'projection': '3d'})
plotdata = {'xs': xs, 'ys': ys, 'zs': zs, 'plotfun': plotfun}
plotdata['plotfun'](plotdata)
with open("test.pickle", 'wb') as file:
    pickle.dump(plotdata, file)

加载pickle文件

import dill as pickle
from mpl_toolkits.mplot3d import Axes3D
import matplotlib.pyplot as plt


plt.figure().add_subplot(111, projection='3d'),
plotdata = pickle.load(open('test.pickle', 'rb'))
plotdata['plotfun'](plotdata)
plt.show()

- Nathan