# Vector Calculus¶

findiff implements the standard vector calculus operations by the convenience classes Gradient, Divergence, Laplace and Curl, respectively.

import numpy as np
from findiff import Gradient, Divergence, Laplacian, Curl


First, we want to apply the gradient, the divergence and the Laplacian to some scalar function We set up our grid and fill the array f:

x, y, z = [np.linspace(0, 10, 100)] * 3
dx, dy, dz = [c - c for c in (x, y, z)]
X, Y, Z = np.meshgrid(x, y, z, indexing='ij')
f = np.sin(X) * np.cos(Y) * np.sin(Z) is a function of three variables, so the array f has three axes:

f.shape

(100, 100, 100)


It is a scalar function, so we can apply the gradient:

grad = Gradient(h=[dx, dy, dz])


Applying the gradient yields a vector function, where each component is a function of three variables. So the shape of the gradient array is:

grad_f.shape

(3, 100, 100, 100)


Applying the Laplacian to a scalar function yields another scalar function:

laplace = Laplacian(h=[dx, dy, dz])
laplace_f = laplace(f)
laplace_f.shape

(100, 100, 100)


Now we define a vector function g = np.array([f, 2*f, 3*f])
g.shape

(3, 100, 100, 100)


Applying the divergence yields a scalar function:

div = Divergence(h=[dx, dy, dz])
div_g = div(g)
div_g.shape

(100, 100, 100)


Applying the curl yields another vector function:

curl = Curl(h=[dx, dy, dz])
curl_g = curl(g)
curl_g.shape

(3, 100, 100, 100)


Note that the curl is only defined for three dimensions. Defining the operator on some other dimension raises an exception.