Implement norm with `np.hypot`

[jankrepl]

Would be nice to check whether it is faster:)

[Stannislav]

I think np.linalg.norm and np.hypot have slightly different interfaces that correspond to different usecases:

• np.linalg.norm(vecs, axis=1) - all coordinates in the same array
• np.hypot(xs, ys) - coordinates in separate arrays, more like a mesh-grid; only 2D.

By in the case of DisplacementField.norm

https://github.com/BlueBrain/Atlas-Download-Tools/blob/8386f2bd0baf4d549dea48b00dae72379f5c6fbf/src/atldld/base.py#L215-L218

we could definitely improve by using np.hypot(self.delta_x, self.delta_y)

[jankrepl]

Heh, for some reason I thought we had np.norm in the code but it is clearly not the case. I will adjust the name of this issue.