def test_get_internals():
parametric_circle = lambda x1, x2, t: [diff(x1, t) - x2, diff(x2, t) + x1]
init_vals_pc = [
IVP(t_0=0.0, u_0=0.0),
IVP(t_0=0.0, u_0=1.0),
]
solver = Solver1D(
ode_system=parametric_circle,
conditions=init_vals_pc,
t_min=0.0,
t_max=2 * np.pi,
)
solver.fit(max_epochs=1)
internals = solver.get_internals()
assert isinstance(internals, dict)
internals = solver.get_internals(return_type='list')
assert isinstance(internals, dict)
internals = solver.get_internals(return_type='dict')
assert isinstance(internals, dict)
internals = solver.get_internals(['generator', 'n_batches'], return_type='dict')
assert isinstance(internals, dict)
internals = solver.get_internals(['generator', 'n_batches'], return_type='list')
assert isinstance(internals, list)
internals = solver.get_internals('train_generator')
assert isinstance(internals, BaseGenerator)
def test_monitor():
laplace = lambda u, x, y: diff(u, x, order=2) + diff(u, y, order=2)
bc = DirichletBVP2D(x_min=0,
x_min_val=lambda y: torch.sin(np.pi * y),
x_max=1,
x_max_val=lambda y: 0,
y_min=0,
y_min_val=lambda x: 0,
y_max=1,
y_max_val=lambda x: 0)
net = FCNN(n_input_units=2, hidden_units=(32, 32))
solution_neural_net_laplace, _ = solve2D(
pde=laplace,
condition=bc,
xy_min=(0, 0),
xy_max=(1, 1),
net=net,
max_epochs=3,
train_generator=Generator2D((32, 32), (0, 0), (1, 1),
method='equally-spaced-noisy'),
batch_size=64,
monitor=Monitor2D(check_every=1, xy_min=(0, 0), xy_max=(1, 1)))
print('Monitor test passed.')
def test_laplace():
laplace = lambda u, x, y: diff(u, x, order=2) + diff(u, y, order=2)
bc = DirichletBVP2D(x_min=0,
x_min_val=lambda y: torch.sin(np.pi * y),
x_max=1,
x_max_val=lambda y: 0,
y_min=0,
y_min_val=lambda x: 0,
y_max=1,
y_max_val=lambda x: 0)
net = FCNN(n_input_units=2, hidden_units=(32, 32))
solution_neural_net_laplace, _ = solve2D(pde=laplace,
condition=bc,
xy_min=(0, 0),
xy_max=(1, 1),
net=net,
max_epochs=300,
train_generator=Generator2D(
(32, 32), (0, 0), (1, 1),
method='equally-spaced-noisy',
xy_noise_std=(0.01, 0.01)),
batch_size=64)
solution_analytical_laplace = lambda x, y: np.sin(np.pi * y) * np.sinh(
np.pi * (1 - x)) / np.sinh(np.pi)
xs, ys = np.linspace(0, 1, 101), np.linspace(0, 1, 101)
xx, yy = np.meshgrid(xs, ys)
sol_net = solution_neural_net_laplace(xx, yy, as_type='np')
sol_ana = solution_analytical_laplace(xx, yy)
assert isclose(sol_net, sol_ana, atol=0.01).all()
print('Laplace test passed.')
def test_lotka_volterra():
alpha, beta, delta, gamma = 1, 1, 1, 1
lotka_volterra = lambda x, y, t : [diff(x, t) - (alpha*x - beta*x*y),
diff(y, t) - (delta*x*y - gamma*y)]
init_vals_lv = [
IVP(t_0=0.0, x_0=1.5),
IVP(t_0=0.0, x_0=1.0)
]
nets_lv = [
FCNN(hidden_units=(32, 32), actv=SinActv),
FCNN(hidden_units=(32, 32), actv=SinActv),
]
solution_lv, _ = solve_system(ode_system=lotka_volterra, conditions=init_vals_lv,
t_min=0.0, t_max=12, nets=nets_lv, max_epochs=12000,
monitor=Monitor(t_min=0.0, t_max=12, check_every=100))
ts = np.linspace(0, 12, 100)
prey_net, pred_net = solution_lv(ts, as_type='np')
def dPdt(P, t):
return [P[0]*alpha - beta*P[0]*P[1], delta*P[0]*P[1] - gamma*P[1]]
P0 = [1.5, 1.0]
Ps = odeint(dPdt, P0, ts)
prey_num = Ps[:,0]
pred_num = Ps[:,1]
assert isclose(prey_net, prey_num, atol=0.1).all()
assert isclose(pred_net, pred_num, atol=0.1).all()
print('Lotka Volterra test passed.')
def spherical_curl(u_r, u_theta, u_phi, r, theta, phi):
r"""Derives and evaluates the spherical curl of a spherical vector field :math:`u`
:param u_r: :math:`r`-component of the vector field :math:`u`, must have shape (n_samples, 1)
:type u_r: `torch.Tensor`
:param u_theta: :math:`\theta`-component of the vector field :math:`u`, must have shape (n_samples, 1)
:type u_theta: `torch.Tensor`
:param u_phi: :math:`\phi`-component of the vector field :math:`u`, must have shape (n_samples, 1)
:type u_phi: `torch.Tensor`
:param r: a vector of :math:`r`-coordinate values, must have shape (n_samples, 1)
:type r: `torch.Tensor`
:param theta: a vector of :math:`\theta`-coordinate values, must have shape (n_samples, 1)
:type theta: `torch.Tensor`
:param phi: a vector of :math:`\phi`-coordinate values, must have shape (n_samples, 1)
:type phi: `torch.Tensor`
:return: :math:`r`, :math:`\theta`, and :math:`\phi` components of the curl, each with shape (n_samples, 1)
:rtype: tuple[`torch.Tensor`]
"""
d_r = lambda u: diff(u, r)
d_theta = lambda u: diff(u, theta)
d_phi = lambda u: diff(u, phi)
curl_r = (d_theta(u_phi * sin(theta)) - d_phi(u_theta)) / (r * sin(theta))
curl_theta = (d_phi(u_r) / sin(theta) - d_r(u_phi * r)) / r
curl_phi = (d_r(u_theta * r) - d_theta(u_r)) / r
return curl_r, curl_theta, curl_phi
def test_laplace():
laplace = lambda u, x, y: diff(u, x, order=2) + diff(u, y, order=2)
bc = DirichletBVP2D(x_min=0,
x_min_val=lambda y: torch.sin(np.pi * y),
x_max=1,
x_max_val=lambda y: 0,
y_min=0,
y_min_val=lambda x: 0,
y_max=1,
y_max_val=lambda x: 0)
net = FCNN(n_input_units=2, hidden_units=(32, 32))
solution_neural_net_laplace, loss_history = solve2D(
pde=laplace,
condition=bc,
xy_min=(0, 0),
xy_max=(1, 1),
net=net,
max_epochs=3,
train_generator=Generator2D((32, 32), (0, 0), (1, 1),
method='equally-spaced-noisy',
xy_noise_std=(0.01, 0.01)),
batch_size=64)
assert isinstance(solution_neural_net_laplace, Solution2D)
assert isinstance(loss_history, dict)
keys = ['train_loss', 'valid_loss']
for key in keys:
assert key in loss_history
assert isinstance(loss_history[key], list)
assert len(loss_history[keys[0]]) == len(loss_history[keys[1]])
def laplacian_spherical(u, r, theta, phi):
"""a helper function that computes the Laplacian in spherical coordinates
"""
r_component = diff(u * r, r, order=2) / r
theta_component = diff(torch.sin(theta) * diff(u, theta),
theta) / (r**2 * torch.sin(theta))
phi_component = diff(u, phi, order=2) / (r**2 * torch.sin(theta)**2)
return r_component + theta_component + phi_component
def test_legacy_signature():
u, t = get_data(flatten_u=False, flatten_t=False)
with pytest.warns(FutureWarning):
diff(x=u, t=t)
with pytest.warns(FutureWarning):
safe_diff(x=u, t=t)
with pytest.warns(FutureWarning):
unsafe_diff(x=u, t=t)
def test_higher_order_derivatives():
u, t = get_data(flatten_u=False, flatten_t=False, f=lambda x: x**2)
assert torch.isclose(diff(u, t), t * 2).all()
assert torch.isclose(diff(u, t, order=2), 2 * torch.ones_like(t)).all()
for order in range(3, 10):
assert torch.isclose(diff(u, t, order=order),
torch.zeros_like(t)).all()
u, t = get_data(flatten_u=False, flatten_t=False, f=torch.exp)
for order in range(1, 10):
assert torch.isclose(diff(u, t, order=order), u).all()
def test_ibvp_ode():
x0, x1 = random.random(), random.random() + 1
u0, u0_prime = random.random(), random.random()
u1, u1_prime = random.random(), random.random()
net = FCNN(1, 1)
# test Dirichlet-Dirichlet
condition = DoubleEndedBVP1D(x_min=x0,
x_max=x1,
x_min_val=u0,
x_max_val=u1)
x = x0 * ones
assert all_close(condition.enforce(net, x),
u0), 'left Dirichlet BC not satisfied'
x = x1 * ones
assert all_close(condition.enforce(net, x),
u1), 'right Dirichlet BC not satisfied'
# test Dirichlet-Neumann
condition = DoubleEndedBVP1D(x_min=x0,
x_max=x1,
x_min_val=u0,
x_max_prime=u1_prime)
x = x0 * ones
assert all_close(condition.enforce(net, x),
u0), 'left Dirichlet BC not satisfied'
x = x1 * ones
assert all_close(diff(condition.enforce(net, x), x),
u1_prime), 'right Neumann BC not satisfied'
# test Neumann-Dirichlet
condition = DoubleEndedBVP1D(x_min=x0,
x_max=x1,
x_min_prime=u0_prime,
x_max_val=u1)
x = x0 * ones
assert all_close(diff(condition.enforce(net, x), x),
u0_prime), 'left Neumann BC not satisfied'
x = x1 * ones
assert all_close(condition.enforce(net, x),
u1), 'right Dirichlet BC not satisfied'
# test Neumann-Neumann
condition = DoubleEndedBVP1D(x_min=x0,
x_max=x1,
x_min_prime=u0_prime,
x_max_prime=u1_prime)
x = x0 * ones
assert all_close(diff(condition.enforce(net, x), x),
u0_prime), 'left Neumann BC not satisfied'
x = x1 * ones
assert all_close(diff(condition.enforce(net, x), x),
u1_prime), 'right Neumann BC not satisfied'
def test_double_ended_bvp_1d(x0, x1, u00, u01, u10, u11, ones, net11):
u0, u0_prime = u00, u01
u1, u1_prime = u10, u11
# test Dirichlet-Dirichlet
condition = DoubleEndedBVP1D(x_min=x0,
x_max=x1,
x_min_val=u0,
x_max_val=u1)
x = x0 * ones
assert all_close(condition.enforce(net11, x),
u0), 'left Dirichlet BC not satisfied'
x = x1 * ones
assert all_close(condition.enforce(net11, x),
u1), 'right Dirichlet BC not satisfied'
# test Dirichlet-Neumann
condition = DoubleEndedBVP1D(x_min=x0,
x_max=x1,
x_min_val=u0,
x_max_prime=u1_prime)
x = x0 * ones
assert all_close(condition.enforce(net11, x),
u0), 'left Dirichlet BC not satisfied'
x = x1 * ones
assert all_close(diff(condition.enforce(net11, x), x),
u1_prime), 'right Neumann BC not satisfied'
# test Neumann-Dirichlet
condition = DoubleEndedBVP1D(x_min=x0,
x_max=x1,
x_min_prime=u0_prime,
x_max_val=u1)
x = x0 * ones
assert all_close(diff(condition.enforce(net11, x), x),
u0_prime), 'left Neumann BC not satisfied'
x = x1 * ones
assert all_close(condition.enforce(net11, x),
u1), 'right Dirichlet BC not satisfied'
# test Neumann-Neumann
condition = DoubleEndedBVP1D(x_min=x0,
x_max=x1,
x_min_prime=u0_prime,
x_max_prime=u1_prime)
x = x0 * ones
assert all_close(diff(condition.enforce(net11, x), x),
u0_prime), 'left Neumann BC not satisfied'
x = x1 * ones
assert all_close(diff(condition.enforce(net11, x), x),
u1_prime), 'right Neumann BC not satisfied'
def test_monitor():
exponential = lambda x, t: diff(x, t) - x
init_val_ex = IVP(t_0=0.0, x_0=1.0)
solution_ex, _ = solve(ode=exponential, condition=init_val_ex,
t_min=0.0, t_max=2.0,
max_epochs=3,
monitor=Monitor(t_min=0.0, t_max=2.0, check_every=1))
print('Monitor test passed.')
def spherical_grad(u, r, theta, phi):
r"""Derives and evaluates the spherical gradient of a spherical scalar field :math:`u`
:param u: a scalar field :math:`u`, must have shape (n_samples, 1)
:type u: `torch.Tensor`
:param r: a vector of :math:`r`-coordinate values, must have shape (n_samples, 1)
:type r: `torch.Tensor`
:param theta: a vector of :math:`\theta`-coordinate values, must have shape (n_samples, 1)
:type theta: `torch.Tensor`
:param phi: a vector of :math:`\phi`-coordinate values, must have shape (n_samples, 1)
:type phi: `torch.Tensor`
:return: :math:`r`, :math:`\theta`, and :math:`\phi` components of the gradient, each with shape (n_samples, 1)
:rtype: tuple[`torch.Tensor`]
"""
grad_r = diff(u, r)
grad_theta = diff(u, theta) / r
grad_phi = diff(u, phi) / (r * sin(theta))
return grad_r, grad_theta, grad_phi
def test_train_generator():
laplace = lambda u, x, y: diff(u, x, order=2) + diff(u, y, order=2)
bc = DirichletBVP2D(
x_min=0, x_min_val=lambda y: torch.sin(np.pi * y),
x_max=1, x_max_val=lambda y: 0,
y_min=0, y_min_val=lambda x: 0,
y_max=1, y_max_val=lambda x: 0
)
net = FCNN(n_input_units=2, hidden_units=(32, 32))
solution_neural_net_laplace, _ = solve2D(
pde=laplace, condition=bc, xy_min=(0, 0), xy_max=(1, 1),
net=net, max_epochs=3,
train_generator=Generator2D((32, 32), (0, 0), (1, 1), method='equally-spaced-noisy'),
batch_size=64,
monitor=Monitor2D(check_every=1, xy_min=(0, 0), xy_max=(1, 1))
)
train_gen = Generator2D((32, 32), (0, 0), (1, 1), method='equally-spaced')
solution_neural_net_laplace, _ = solve2D(
pde=laplace, condition=bc, xy_min=(0, 0), xy_max=(1, 1),
net=net, max_epochs=3, train_generator=train_gen, batch_size=64
)
train_gen = Generator2D((32, 32), (0, 0), (1, 1), method='equally-spaced-noisy')
solution_neural_net_laplace, _ = solve2D(
pde=laplace, condition=bc, xy_min=(0, 0), xy_max=(1, 1),
net=net, max_epochs=3, train_generator=train_gen, batch_size=64
)
with raises(ValueError):
train_gen = Generator2D((32, 32), (0, 0), (1, 1), method='magic')
valid_gen = Generator2D((32, 32), (0, 0), (1, 1), method='equally-spaced-noisy')
train_gen = Generator2D((32, 32), (0, 0), (1, 1), method='equally-spaced')
solution_neural_net_laplace, _ = solve2D(
pde=laplace, condition=bc,
net=net, max_epochs=3, train_generator=train_gen, valid_generator=valid_gen, batch_size=64
)
with raises(RuntimeError):
solution_neural_net_laplace, _ = solve2D(
pde=laplace, condition=bc,
net=net, max_epochs=3, batch_size=64
)
def test_ivp(x0, y0, y1, ones, net11):
x = x0 * ones
cond = IVP(x0, y0)
y = cond.enforce(net11, x)
assert torch.isclose(y, y0 * ones).all(), "y(x_0) != y_0"
cond = IVP(x0, y0, y1)
y = cond.enforce(net11, x)
assert all_close(y, y0), "y(x_0) != y_0"
assert all_close(diff(y, x), y1), "y'(x_0) != y'_0"
def test_train_generator():
laplace = lambda u, x, y: diff(u, x, order=2) + diff(u, y, order=2)
bc = DirichletBVP2D(x_min=0,
x_min_val=lambda y: torch.sin(np.pi * y),
x_max=1,
x_max_val=lambda y: 0,
y_min=0,
y_min_val=lambda x: 0,
y_max=1,
y_max_val=lambda x: 0)
net = FCNN(n_input_units=2, hidden_units=(32, 32))
with pytest.raises(ValueError), pytest.warns(FutureWarning):
solution_neural_net_laplace, _ = solve2D(pde=laplace,
condition=bc,
net=net,
max_epochs=3,
batch_size=64)
def test_ode_system():
parametric_circle = lambda u1, u2, t: [diff(u1, t) - u2, diff(u2, t) + u1]
init_vals_pc = [IVP(t_0=0.0, u_0=0.0), IVP(t_0=0.0, u_0=1.0)]
solution_pc, loss_history = solve_system(
ode_system=parametric_circle,
conditions=init_vals_pc,
t_min=0.0,
t_max=2 * np.pi,
max_epochs=10,
)
assert isinstance(solution_pc, Solution)
assert isinstance(loss_history, dict)
keys = ['train_loss', 'valid_loss']
for key in keys:
assert key in loss_history
assert isinstance(loss_history[key], list)
assert len(loss_history[keys[0]]) == len(loss_history[keys[1]])
def test_monitor():
exponential = lambda u, t: diff(u, t) - u
init_val_ex = IVP(t_0=0.0, u_0=1.0)
solution_ex, _ = solve(ode=exponential,
condition=init_val_ex,
t_min=0.0,
t_max=2.0,
max_epochs=3,
monitor=Monitor(t_min=0.0, t_max=2.0,
check_every=1))
def test_ode_ivp():
oscillator = lambda x, t: diff(x, t, order=2) + x
init_val_ho = IVP(t_0=0.0, x_0=0.0, x_0_prime=1.0)
solution_ho, _ = solve(ode=oscillator, condition=init_val_ho,
max_epochs=3000,
t_min=0.0, t_max=2*np.pi)
ts = np.linspace(0, 2*np.pi, 100)
x_net = solution_ho(ts, as_type='np')
x_ana = np.sin(ts)
assert isclose(x_net, x_ana, atol=0.1).all()
print('IVP basic test passed.')
def test_ode_bvp():
oscillator = lambda x, t: diff(x, t, order=2) + x
bound_val_ho = DirichletBVP(t_0=0.0, x_0=0.0, t_1=1.5*np.pi, x_1=-1.0)
solution_ho, _ = solve(ode=oscillator, condition=bound_val_ho,
max_epochs=3000,
t_min=0.0, t_max=1.5*np.pi)
ts = np.linspace(0, 1.5*np.pi, 100)
x_net = solution_ho(ts, as_type='np')
x_ana = np.sin(ts)
assert isclose(x_net, x_ana, atol=0.1).all()
print('BVP basic test passed.')
def test_ode_system():
parametric_circle = lambda x1, x2, t : [diff(x1, t) - x2,
diff(x2, t) + x1]
init_vals_pc = [
IVP(t_0=0.0, x_0=0.0),
IVP(t_0=0.0, x_0=1.0)
]
solution_pc, _ = solve_system(ode_system=parametric_circle,
conditions=init_vals_pc,
t_min=0.0, t_max=2*np.pi,
max_epochs=5000,)
ts = np.linspace(0, 2*np.pi, 100)
x1_net, x2_net = solution_pc(ts, as_type='np')
x1_ana, x2_ana = np.sin(ts), np.cos(ts)
assert isclose(x1_net, x1_ana, atol=0.1).all()
assert isclose(x2_net, x2_ana, atol=0.1).all()
print('solve_system basic test passed.')
def spherical_vector_laplacian(u_r, u_theta, u_phi, r, theta, phi):
r"""Derives and evaluates the spherical laplacian of a spherical vector field :math:`u`
:param u_r: :math:`r`-component of the vector field :math:`u`, must have shape (n_samples, 1)
:type u_r: `torch.Tensor`
:param u_theta: :math:`\theta`-component of the vector field :math:`u`, must have shape (n_samples, 1)
:type u_theta: `torch.Tensor`
:param u_phi: :math:`\phi`-component of the vector field :math:`u`, must have shape (n_samples, 1)
:type u_phi: `torch.Tensor`
:param r: a vector of :math:`r`-coordinate values, must have shape (n_samples, 1)
:type r: `torch.Tensor`
:param theta: a vector of :math:`\theta`-coordinate values, must have shape (n_samples, 1)
:type theta: `torch.Tensor`
:param phi: a vector of :math:`\phi`-coordinate values, must have shape (n_samples, 1)
:type phi: `torch.Tensor`
:return: the laplacian evaluated at :math:`(r, \theta, \phi)`, with shape (n_samples, 1)
:rtype: `torch.Tensor`
"""
d_theta = lambda u: diff(u, theta)
d_phi = lambda u: diff(u, phi)
scalar_lap = lambda u: spherical_laplacian(u, r, theta, phi)
lap_r = \
scalar_lap(u_r) \
- 2 * u_r / r ** 2 \
- 2 * d_theta(u_theta * sin(theta)) / (r ** 2 * sin(theta)) \
- 2 * d_phi(u_phi) / (r ** 2 * sin(theta))
lap_theta = \
scalar_lap(u_theta) \
- u_theta / (r ** 2 * sin(theta) ** 2) \
+ 2 * d_theta(u_r) / r ** 2 \
- 2 * cos(theta) * d_phi(u_phi) / (r ** 2 * sin(theta) ** 2)
lap_phi = \
scalar_lap(u_phi) \
- u_phi / (r ** 2 * sin(theta) ** 2) \
+ 2 * d_phi(u_r) / (r ** 2 * sin(theta)) \
+ 2 * cos(theta) * d_phi(u_theta) / (r ** 2 * sin(theta) ** 2)
return lap_r, lap_theta, lap_phi
def test_ivp():
x = x0 * ones
net = FCNN(1, 1)
cond = IVP(x0, y0)
y = cond.enforce(net, x)
assert torch.isclose(y, y0 * ones).all(), "y(x_0) != y_0"
cond = IVP(x0, y0, y1)
y = cond.enforce(net, x)
assert all_close(y, y0), "y(x_0) != y_0"
assert all_close(diff(y, x), y1), "y'(x_0) != y'_0"
def test_heat():
k, L, T = 0.3, 2, 3
heat = lambda u, x, t: diff(u, t) - k * diff(u, x, order=2)
ibvp = IBVP1D(x_min=0,
x_min_val=lambda t: 0,
x_max=L,
x_max_val=lambda t: 0,
t_min=0,
t_min_val=lambda x: torch.sin(np.pi * x / L))
net = FCNN(n_input_units=2, hidden_units=(32, 32))
def mse(u, x, y):
true_u = torch.sin(np.pi * y) * torch.sinh(np.pi *
(1 - x)) / np.sinh(np.pi)
return torch.mean((u - true_u)**2)
solution_neural_net_heat, _ = solve2D(pde=heat,
condition=ibvp,
xy_min=(0, 0),
xy_max=(L, T),
net=net,
max_epochs=300,
train_generator=Generator2D(
(32, 32), (0, 0), (L, T),
method='equally-spaced-noisy'),
batch_size=64,
metrics={'mse': mse})
solution_analytical_heat = lambda x, t: np.sin(np.pi * x / L) * np.exp(
-k * np.pi**2 * t / L**2)
xs = np.linspace(0, L, 101)
ts = np.linspace(0, T, 101)
xx, tt = np.meshgrid(xs, ts)
make_animation(solution_neural_net_heat, xs, ts) # test animation
sol_ana = solution_analytical_heat(xx, tt)
sol_net = solution_neural_net_heat(xx, tt, as_type='np')
assert isclose(sol_net, sol_ana, atol=0.01).all()
print('Heat test passed.')
def spherical_div(u_r, u_theta, u_phi, r, theta, phi):
r"""Derives and evaluates the spherical divergence of a spherical vector field :math:`u`
:param u_r: :math:`r`-component of the vector field :math:`u`, must have shape (n_samples, 1)
:type u_r: `torch.Tensor`
:param u_theta: :math:`\theta`-component of the vector field :math:`u`, must have shape (n_samples, 1)
:type u_theta: `torch.Tensor`
:param u_phi: :math:`\phi`-component of the vector field :math:`u`, must have shape (n_samples, 1)
:type u_phi: `torch.Tensor`
:param r: a vector of :math:`r`-coordinate values, must have shape (n_samples, 1)
:type r: `torch.Tensor`
:param theta: a vector of :math:`\theta`-coordinate values, must have shape (n_samples, 1)
:type theta: `torch.Tensor`
:param phi: a vector of :math:`\phi`-coordinate values, must have shape (n_samples, 1)
:type phi: `torch.Tensor`
:return: the divergence evaluated at :math:`(r, \theta, \phi)`, with shape (n_samples, 1)
:rtype: `torch.Tensor`
"""
div_r = diff(u_r * r**2, r) / r**2
div_theta = diff(u_theta * sin(theta), theta) / (r * sin(theta))
div_phi = diff(u_phi, phi) / (r * sin(theta))
return div_r + div_theta + div_phi
def spherical_laplacian(u, r, theta, phi):
r"""Derives and evaluates the spherical laplacian of a spherical scalar field :math:`u`
:param u: a scalar field :math:`u`, must have shape (n_samples, 1)
:type u: `torch.Tensor`
:param r: a vector of :math:`r`-coordinate values, must have shape (n_samples, 1)
:type r: `torch.Tensor`
:param theta: a vector of :math:`\theta`-coordinate values, must have shape (n_samples, 1)
:type theta: `torch.Tensor`
:param phi: a vector of :math:`\phi`-coordinate values, must have shape (n_samples, 1)
:type phi: `torch.Tensor`
:return: the laplacian evaluated at :math:`(r, \theta, \phi)`, with shape (n_samples, 1)
:rtype: `torch.Tensor`
"""
d_r = lambda u: diff(u, r)
d_theta = lambda u: diff(u, theta)
d_phi = lambda u: diff(u, phi)
lap_r = d_r(d_r(u) * r**2) / r**2
lap_theta = d_theta(d_theta(u) * sin(theta)) / (r**2 * sin(theta))
lap_phi = d_phi(d_phi(u)) / (r**2 * sin(theta)**2)
return lap_r + lap_theta + lap_phi
def test_neumann_boundaries_2():
k, L, T = 0.3, 2, 3
heat = lambda u, x, t: diff(u, t) - k * diff(u, x, order=2)
solution_analytical_heat = lambda x, t: np.sin(np.pi * x / L) * np.exp(
-k * np.pi**2 * t / L**2)
# Neumann on the left Dirichlet on the right
ibvp = IBVP1D(
x_min=0,
x_min_prime=lambda t: np.pi / L * torch.exp(-k * np.pi**2 * t / L**2),
x_max=L,
x_max_val=lambda t: 0,
t_min=0,
t_min_val=lambda x: torch.sin(np.pi * x / L))
net = FCNN(n_input_units=2, hidden_units=(32, 32))
solution_neural_net_heat, _ = solve2D(pde=heat,
condition=ibvp,
xy_min=(0, 0),
xy_max=(L, T),
net=net,
max_epochs=300,
train_generator=Generator2D(
(32, 32), (0, 0), (L, T),
method='equally-spaced-noisy'),
batch_size=64)
xs = np.linspace(0, L, 101)
ts = np.linspace(0, T, 101)
xx, tt = np.meshgrid(xs, ts)
make_animation(solution_neural_net_heat, xs, ts) # test animation
sol_ana = solution_analytical_heat(xx, tt)
sol_net = solution_neural_net_heat(xx, tt, as_type='np')
assert isclose(sol_net, sol_ana, atol=0.1).all()
print('Neumann on the left Dirichlet on the right test passed.')
def test_ode():
def mse(x, t):
true_x = torch.sin(t)
return torch.mean((x - true_x) ** 2)
exponential = lambda x, t: diff(x, t) - x
init_val_ex = IVP(t_0=0.0, x_0=1.0)
solution_ex, _ = solve(ode=exponential, condition=init_val_ex,
t_min=0.0, t_max=2.0, shuffle=False,
max_epochs=2000, return_best=True, metrics={'mse': mse})
ts = np.linspace(0, 2.0, 100)
x_net = solution_ex(ts, as_type='np')
x_ana = np.exp(ts)
assert isclose(x_net, x_ana, atol=0.1).all()
print('solve basic test passed.')
def test_ode():
def mse(u, t):
true_u = torch.sin(t)
return torch.mean((u - true_u) ** 2)
exponential = lambda u, t: diff(u, t) - u
init_val_ex = IVP(t_0=0.0, u_0=1.0)
solution_ex, loss_history = solve(ode=exponential, condition=init_val_ex,
t_min=0.0, t_max=2.0, shuffle=False,
max_epochs=10, return_best=True, metrics={'mse': mse})
assert isinstance(solution_ex, Solution1D)
assert isinstance(loss_history, dict)
keys = ['train_loss', 'valid_loss']
for key in keys:
assert key in loss_history
assert isinstance(loss_history[key], list)
assert len(loss_history[keys[0]]) == len(loss_history[keys[1]])
def test_train_generator():
exponential = lambda x, t: diff(x, t) - x
init_val_ex = IVP(t_0=0.0, x_0=1.0)
train_gen = Generator1D(size=32, t_min=0.0, t_max=2.0, method='uniform')
solution_ex, _ = solve(ode=exponential, condition=init_val_ex,
t_min=0.0, t_max=2.0,
train_generator=train_gen,
max_epochs=3)
train_gen = Generator1D(size=32, t_min=0.0, t_max=2.0, method='equally-spaced')
solution_ex, _ = solve(ode=exponential, condition=init_val_ex,
t_min=0.0, t_max=2.0,
train_generator=train_gen,
max_epochs=3)
train_gen = Generator1D(size=32, t_min=0.0, t_max=2.0, method='equally-spaced-noisy')
solution_ex, _ = solve(ode=exponential, condition=init_val_ex,
t_min=0.0, t_max=2.0,
train_generator=train_gen,
max_epochs=3)
train_gen = Generator1D(size=32, t_min=0.0, t_max=2.0, method='equally-spaced-noisy', noise_std=0.01)
solution_ex, _ = solve(ode=exponential, condition=init_val_ex,
t_min=0.0, t_max=2.0,
train_generator=train_gen,
max_epochs=3)
train_gen = Generator1D(size=32, t_min=np.log10(0.1), t_max=np.log10(2.0), method='log-spaced')
solution_ex, _ = solve(ode=exponential, condition=init_val_ex,
t_min=0.1, t_max=2.0,
train_generator=train_gen,
max_epochs=3)
train_gen = Generator1D(size=32, t_min=np.log10(0.1), t_max=np.log10(2.0), method='log-spaced-noisy')
solution_ex, _ = solve(ode=exponential, condition=init_val_ex,
t_min=0.1, t_max=2.0,
train_generator=train_gen,
max_epochs=3)
train_gen = Generator1D(size=32, t_min=np.log10(0.1), t_max=np.log10(2.0), method='log-spaced-noisy', noise_std=0.01)
solution_ex, _ = solve(ode=exponential, condition=init_val_ex,
t_min=0.1, t_max=2.0,
train_generator=train_gen,
max_epochs=3)
with raises(ValueError):
train_gen = Generator1D(size=32, t_min=0.0, t_max=2.0, method='magic')
print('ExampleGenerator test passed.')