tagged as revision as it needs tests. Probably just png for the image tests.

I vectorized and simplified quite a bit, possibly at the expense of readability. I tried to comment the tricky bits, though.

Tests will come soon.

Tests are in and Travis is happy. Anything else need to be done for this PR?

👍 from me. @efiring Should probably take a look at this.

Is there any noticeable speed up?

Thank you very much for your feature, and for vectorizing much of the code. In the initial PR, we discussed vectorizing, but the students who developed the feature could only get so far, so thank you for getting it all the way. It is much cleaner to read now.

Ok, linting done. I also did a quick benchmark using this setup code:

ax = plt.subplot(projection='3d')
x, y, z = np.meshgrid(np.arange(-0.8, 1, 0.2),
                      np.arange(-0.8, 1, 0.2),
                      np.arange(-0.8, 1, 0.8))
u = np.sin(np.pi * x) * np.cos(np.pi * y) * np.cos(np.pi * z)
v = -np.cos(np.pi * x) * np.sin(np.pi * y) * np.cos(np.pi * z)
w = (np.sqrt(2.0 / 3.0) * np.cos(np.pi * x) * np.cos(np.pi * y) *
     np.sin(np.pi * z))

Before this PR:

In [5]: %timeit ax.quiver(x, y, z, u, v, w, length=0.1)
10 loops, best of 3: 45.7 ms per loop

After this PR:

In [5]: %timeit ax.quiver(x, y, z, u, v, w, length=0.1)
100 loops, best of 3: 8.75 ms per loop

Looks like a 5x speedup! Not bad.

@perimosocordiae One last thing, can you add an entry to whats_new to advertise the new pivot feature?

Ok, added.

Thank you.