The core of that Life expression works by taking the array of cells and shifting it in the eight different directions, then summing those arrays to get the population counts.

I tried translating this into Python — but I've never written numpy code before, so this is probably less efficient than it could be. But it does work and you can see a glider move through a few generations.

The array-shifting logic is in the populations function, with np.roll being the equivalent of the APL rotate operation (written as ⌽ and ⊖ in the original).

import numpy as np

def alive(popu, cell):
    return int(popu == 3 or (popu == 4 and cell))
alive = np.frompyfunc(alive, 2, 1)

def populations(grid):
    return np.array(
            [ np.roll(r1, shift, 1)
                for shift in [-1, 0, 1]
                for r1 in (np.roll(grid, shift, 0) for shift in [-1, 0, 1]) ]
            ).sum(axis=0)

def nextgen(grid):
    return alive(populations(grid), grid)

grid = np.array([[0, 0, 0, 0, 0, 0],
                 [0, 0, 1, 0, 0, 0],
                 [0, 0, 0, 1, 0, 0],
                 [0, 1, 1, 1, 0, 0],
                 [0, 0, 0, 0, 0, 0],
                 [0, 0, 0, 0, 0, 0]])

for gens in range(5):
    print(grid)
    grid = nextgen(grid)
print(grid)