def test_cart2pol(self, batch_shape, device, dtype):
# generate input data
x = torch.rand(batch_shape, dtype=dtype)
y = torch.rand(batch_shape, dtype=dtype)
x = x.to(device)
y = y.to(device)
# convert cart/pol
rho_cart2pol, phi_cart2pol = kornia.cart2pol(x, y, 0)
x_cart2pol, y_cart2pol = kornia.pol2cart(rho_cart2pol, phi_cart2pol)
assert_allclose(x, x_cart2pol)
assert_allclose(y, y_cart2pol)
assert gradcheck(kornia.cart2pol, (tensor_to_gradcheck_var(x),
tensor_to_gradcheck_var(y), ), raise_exception=True)
def test_pol2cart(self, batch_shape, device, dtype):
# generate input data
rho = torch.rand(batch_shape, dtype=dtype)
phi = kornia.pi * torch.rand(batch_shape, dtype=dtype)
rho = rho.to(device)
phi = phi.to(device)
# convert pol/cart
x_pol2cart, y_pol2cart = kornia.pol2cart(rho, phi)
rho_pol2cart, phi_pol2cart = kornia.cart2pol(x_pol2cart, y_pol2cart, 0)
assert_allclose(rho, rho_pol2cart)
assert_allclose(phi, phi_pol2cart)
assert gradcheck(kornia.pol2cart, (tensor_to_gradcheck_var(rho),
tensor_to_gradcheck_var(phi), ), raise_exception=True)
def test_smoke(self, device, dtype):
x = torch.ones(1, 1, 1, 1, device=device, dtype=dtype)
assert kornia.pol2cart(x, x) is not None
assert kornia.cart2pol(x, x) is not None
