def unpicklable_misc_operator_with_arrays_and_products_factory(n):
if n == 0:
from pymor.operators.numpy import NumpyGenericOperator
op, _, U, V, sp, rp = numpy_matrix_operator_with_arrays_and_products_factory(100, 20, 4, 3, n)
mat = op.matrix
op2 = NumpyGenericOperator(mapping=lambda U: mat.dot(U.T).T, adjoint_mapping=lambda U: mat.T.dot(U.T).T,
dim_source=100, dim_range=20, linear=True)
return op2, _, U, V, sp, rp
else:
assert False
def create_robin_solop(localizer, bilifo, source_space, omega_star_space):
# Robin-Loesungsoperator
def solve(va):
g = localizer.to_space(NumpyVectorSpace.make_array(va), source_space,
omega_star_space)
solution = bilifo.apply_inverse(g)
return solution.data
cavesize = len(localizer.join_spaces(source_space))
rangesize = len(localizer.join_spaces(omega_star_space))
return NumpyGenericOperator(solve,
dim_source=cavesize,
dim_range=rangesize,
linear=True)
def create_dirichlet_solop(localizer, local_op, rhsop, source_space,
training_space):
# Dirichlet-Loesungsoperator
def solve(va):
solution = local_op.apply_inverse(
-rhsop.apply(NumpyVectorSpace.make_array(va)))
return solution.data
cavesize = len(localizer.join_spaces(source_space))
rangesize = len(localizer.join_spaces(training_space))
return NumpyGenericOperator(solve,
dim_source=cavesize,
dim_range=rangesize,
linear=True)
def create_dirichlet_transfer(localizer, local_op, rhsop, source_space,
training_space, range_space, pou):
# Dirichlet-Transferoperator
def transfer(va):
range_solution = localizer.to_space(
local_op.apply_inverse(
-rhsop.apply(NumpyVectorSpace.make_array(va))), training_space,
range_space)
return pou[range_space](range_solution).data
cavesize = len(localizer.join_spaces(source_space))
rangesize = len(localizer.join_spaces(range_space))
return NumpyGenericOperator(transfer,
dim_source=cavesize,
dim_range=rangesize,
linear=True)
def create_robin_transfer(localizer, bilifo, source_space, omega_star_space,
range_space, pou):
# Robin-Transferoperator
def transfer(va):
g = localizer.to_space(NumpyVectorSpace.make_array(va), source_space,
omega_star_space)
solution = bilifo.apply_inverse(g)
range_solution = localizer.to_space(solution, omega_star_space,
range_space)
return pou[range_space](range_solution).data
cavesize = len(localizer.join_spaces(source_space))
rangesize = len(localizer.join_spaces(range_space))
return NumpyGenericOperator(transfer,
dim_source=cavesize,
dim_range=rangesize,
linear=True)
def discretize_instationary_fv(analytical_problem, diameter=None, domain_discretizer=None, grid_type=None,
num_flux='lax_friedrichs', lxf_lambda=1., eo_gausspoints=5, eo_intervals=1,
grid=None, boundary_info=None, num_values=None, time_stepper=None, nt=None,
preassemble=True):
"""Discretizes an |InstationaryProblem| with a |StationaryProblem| as stationary part
using the finite volume method.
Parameters
----------
analytical_problem
The |InstationaryProblem| to discretize.
diameter
If not `None`, `diameter` is passed to the `domain_discretizer`.
domain_discretizer
Discretizer to be used for discretizing the analytical domain. This has
to be a function `domain_discretizer(domain_description, diameter, ...)`.
If further arguments should be passed to the discretizer, use
:func:`functools.partial`. If `None`, |discretize_domain_default| is used.
grid_type
If not `None`, this parameter is forwarded to `domain_discretizer` to specify
the type of the generated |Grid|.
num_flux
The numerical flux to use in the finite volume formulation. Allowed
values are `'lax_friedrichs'`, `'engquist_osher'`, `'simplified_engquist_osher'`
(see :mod:`pymor.discretizers.builtin.fv`).
lxf_lambda
The stabilization parameter for the Lax-Friedrichs numerical flux
(ignored, if different flux is chosen).
eo_gausspoints
Number of Gauss points for the Engquist-Osher numerical flux
(ignored, if different flux is chosen).
eo_intervals
Number of sub-intervals to use for integration when using Engquist-Osher
numerical flux (ignored, if different flux is chosen).
grid
Instead of using a domain discretizer, the |Grid| can also be passed directly
using this parameter.
boundary_info
A |BoundaryInfo| specifying the boundary types of the grid boundary entities.
Must be provided if `grid` is specified.
num_values
The number of returned vectors of the solution trajectory. If `None`, each
intermediate vector that is calculated is returned.
time_stepper
The :class:`time-stepper <pymor.algorithms.timestepping.TimeStepper>`
to be used by :class:`~pymor.models.basic.InstationaryModel.solve`.
nt
If `time_stepper` is not specified, the number of time steps for implicit
Euler time stepping.
preassemble
If `True`, preassemble all operators in the resulting |Model|.
Returns
-------
m
The |Model| that has been generated.
data
Dictionary with the following entries:
:grid: The generated |Grid|.
:boundary_info: The generated |BoundaryInfo|.
:unassembled_m: In case `preassemble` is `True`, the generated |Model|
before preassembling operators.
"""
assert isinstance(analytical_problem, InstationaryProblem)
assert isinstance(analytical_problem.stationary_part, StationaryProblem)
assert grid is None or boundary_info is not None
assert boundary_info is None or grid is not None
assert grid is None or domain_discretizer is None
assert (time_stepper is None) != (nt is None)
p = analytical_problem
m, data = discretize_stationary_fv(p.stationary_part, diameter=diameter, domain_discretizer=domain_discretizer,
grid_type=grid_type, num_flux=num_flux, lxf_lambda=lxf_lambda,
eo_gausspoints=eo_gausspoints, eo_intervals=eo_intervals, grid=grid,
boundary_info=boundary_info)
grid = data['grid']
if p.initial_data.parametric:
def initial_projection(U, mu):
I = p.initial_data.evaluate(grid.quadrature_points(0, order=2), mu).squeeze()
I = np.sum(I * grid.reference_element.quadrature(order=2)[1], axis=1) * (1. / grid.reference_element.volume)
I = m.solution_space.make_array(I)
return I.lincomb(U).to_numpy()
I = NumpyGenericOperator(initial_projection, dim_range=grid.size(0), linear=True, range_id=m.solution_space.id,
parameter_type=p.initial_data.parameter_type)
else:
I = p.initial_data.evaluate(grid.quadrature_points(0, order=2)).squeeze()
I = np.sum(I * grid.reference_element.quadrature(order=2)[1], axis=1) * (1. / grid.reference_element.volume)
I = m.solution_space.make_array(I)
if time_stepper is None:
if p.stationary_part.diffusion is None:
time_stepper = ExplicitEulerTimeStepper(nt=nt)
else:
time_stepper = ImplicitEulerTimeStepper(nt=nt)
rhs = None if isinstance(m.rhs, ZeroOperator) else m.rhs
m = InstationaryModel(operator=m.operator, rhs=rhs, mass=None, initial_data=I, T=p.T,
products=m.products, time_stepper=time_stepper,
parameter_space=p.parameter_space, visualizer=m.visualizer,
num_values=num_values, name=f'{p.name}_FV')
if preassemble:
data['unassembled_m'] = m
m = preassemble_(m)
return m, data
def discretize_nonlinear_instationary_advection_fv(analytical_problem,
diameter=None,
nt=100,
num_flux='lax_friedrichs',
lxf_lambda=1.,
eo_gausspoints=5,
eo_intervals=1,
num_values=None,
domain_discretizer=None,
grid=None,
boundary_info=None):
"""Discretizes an |InstationaryAdvectionProblem| using the finite volume method.
Simple explicit Euler time-stepping is used for time-discretization.
Parameters
----------
analytical_problem
The |InstationaryAdvectionProblem| to discretize.
diameter
If not `None`, `diameter` is passed to the `domain_discretizer`.
nt
The number of time-steps.
num_flux
The numerical flux to use in the finite volume formulation. Allowed
values are `'lax_friedrichs'`, `'engquist_osher'`, `'simplified_engquist_osher'`.
(See :mod:`pymor.operators.fv`.)
lxf_lambda
The stabilization parameter for the Lax-Friedrichs numerical flux.
(Ignored, if different flux is chosen.)
eo_gausspoints
Number of Gauss points for the Engquist-Osher numerical flux.
(Ignored, if different flux is chosen.)
eo_intervals
Number of sub-intervals to use for integration when using Engquist-Osher
numerical flux. (Ignored, if different flux is chosen.)
num_values
The number of returned vectors of the solution trajectory. If `None`, each
intermediate vector that is calculated is returned.
domain_discretizer
Discretizer to be used for discretizing the analytical domain. This has
to be a function `domain_discretizer(domain_description, diameter, ...)`.
If further arguments should be passed to the discretizer, use
:func:`functools.partial`. If `None`, |discretize_domain_default| is used.
grid
Instead of using a domain discretizer, the |Grid| can also be passed directly
using this parameter.
boundary_info
A |BoundaryInfo| specifying the boundary types of the grid boundary entities.
Must be provided if `grid` is specified.
Returns
-------
discretization
The |Discretization| that has been generated.
data
Dictionary with the following entries:
:grid: The generated |Grid|.
:boundary_info: The generated |BoundaryInfo|.
"""
assert isinstance(analytical_problem, InstationaryAdvectionProblem)
assert grid is None or boundary_info is not None
assert boundary_info is None or grid is not None
assert grid is None or domain_discretizer is None
assert num_flux in ('lax_friedrichs', 'engquist_osher',
'simplified_engquist_osher')
if grid is None:
domain_discretizer = domain_discretizer or discretize_domain_default
if diameter is None:
grid, boundary_info = domain_discretizer(analytical_problem.domain)
else:
grid, boundary_info = domain_discretizer(analytical_problem.domain,
diameter=diameter)
p = analytical_problem
if num_flux == 'lax_friedrichs':
L = nonlinear_advection_lax_friedrichs_operator(
grid,
boundary_info,
p.flux_function,
dirichlet_data=p.dirichlet_data,
lxf_lambda=lxf_lambda)
elif num_flux == 'engquist_osher':
L = nonlinear_advection_engquist_osher_operator(
grid,
boundary_info,
p.flux_function,
p.flux_function_derivative,
gausspoints=eo_gausspoints,
intervals=eo_intervals,
dirichlet_data=p.dirichlet_data)
else:
L = nonlinear_advection_simplified_engquist_osher_operator(
grid,
boundary_info,
p.flux_function,
p.flux_function_derivative,
dirichlet_data=p.dirichlet_data)
F = None if p.rhs is None else L2ProductFunctional(grid, p.rhs)
if p.initial_data.parametric:
def initial_projection(U, mu):
I = p.initial_data.evaluate(grid.quadrature_points(0, order=2),
mu).squeeze()
I = np.sum(I * grid.reference_element.quadrature(order=2)[1],
axis=1) * (1. / grid.reference_element.volume)
I = NumpyVectorArray(I, copy=False)
return I.lincomb(U).data
inject_sid(
initial_projection, __name__ +
'.discretize_nonlinear_instationary_advection_fv.initial_data',
p.initial_data, grid)
I = NumpyGenericOperator(initial_projection,
dim_range=grid.size(0),
linear=True,
parameter_type=p.initial_data.parameter_type)
else:
I = p.initial_data.evaluate(grid.quadrature_points(0,
order=2)).squeeze()
I = np.sum(I * grid.reference_element.quadrature(order=2)[1],
axis=1) * (1. / grid.reference_element.volume)
I = NumpyVectorArray(I, copy=False)
inject_sid(
I, __name__ +
'.discretize_nonlinear_instationary_advection_fv.initial_data',
p.initial_data, grid)
products = {'l2': L2Product(grid, boundary_info)}
if grid.dim == 2:
visualizer = PatchVisualizer(grid=grid,
bounding_box=grid.domain,
codim=0)
elif grid.dim == 1:
visualizer = Matplotlib1DVisualizer(grid, codim=0)
else:
visualizer = None
parameter_space = p.parameter_space if hasattr(p,
'parameter_space') else None
time_stepper = ExplicitEulerTimeStepper(nt=nt)
discretization = InstationaryDiscretization(
operator=L,
rhs=F,
initial_data=I,
T=p.T,
products=products,
time_stepper=time_stepper,
parameter_space=parameter_space,
visualizer=visualizer,
num_values=num_values,
name='{}_FV'.format(p.name))
return discretization, {'grid': grid, 'boundary_info': boundary_info}
def discretize_parabolic_fv(analytical_problem, diameter=None, domain_discretizer=None,
grid=None, boundary_info=None, num_values=None, time_stepper=None, nt=None):
"""Discretizes an |ParabolicProblem| using the finite volume method.
Parameters
----------
analytical_problem
The |ParabolicProblem| to discretize.
diameter
If not `None`, `diameter` is passed to the `domain_discretizer`.
domain_discretizer
Discretizer to be used for discretizing the analytical domain. This has
to be a function `domain_discretizer(domain_description, diameter, ...)`.
If further arguments should be passed to the discretizer, use
:func:`functools.partial`. If `None`, |discretize_domain_default| is used.
grid
Instead of using a domain discretizer, the |Grid| can also be passed directly
using this parameter.
boundary_info
A |BoundaryInfo| specifying the boundary types of the grid boundary entities.
Must be provided if `grid` is specified.
num_values
The number of returned vectors of the solution trajectory. If `None`, each
intermediate vector that is calculated is returned.
time_stepper
The time-stepper to be used by :class:`~pymor.discretizations.basic.InstationaryDiscretization.solve`. Has to
satisfy the :class:`~pymor.algorithms.timestepping.TimeStepperInterface`.
nt
The number of time-steps. If provided implicit euler time-stepping is used.
Returns
-------
discretization
The |Discretization| that has been generated.
data
Dictionary with the following entries:
:grid: The generated |Grid|.
:boundary_info: The generated |BoundaryInfo|.
"""
assert isinstance(analytical_problem, ParabolicProblem)
assert grid is None or boundary_info is not None
assert boundary_info is None or grid is not None
assert grid is None or domain_discretizer is None
assert time_stepper is None or nt is None
p = analytical_problem
d, data = discretize_elliptic_fv(p.elliptic_part(), diameter=diameter, domain_discretizer=domain_discretizer,
grid=grid, boundary_info=boundary_info)
if p.initial_data.parametric:
def initial_projection(U, mu):
I = p.initial_data.evaluate(grid.quadrature_points(0, order=2), mu).squeeze()
I = np.sum(I * grid.reference_element.quadrature(order=2)[1], axis=1) * (1. / grid.reference_element.volume)
I = NumpyVectorArray(I, copy=False)
return I.lincomb(U).data
I = NumpyGenericOperator(initial_projection, dim_range=grid.size(0), linear=True,
parameter_type=p.initial_data.parameter_type)
else:
I = p.initial_data.evaluate(grid.quadrature_points(0, order=2)).squeeze()
I = np.sum(I * grid.reference_element.quadrature(order=2)[1], axis=1) * (1. / grid.reference_element.volume)
I = NumpyVectorArray(I, copy=False)
if time_stepper is None:
time_stepper = ImplicitEulerTimeStepper(nt=nt)
discretization = InstationaryDiscretization(operator=d.operator, rhs=d.rhs, mass=None, initial_data=I, T=p.T,
products=d.products, time_stepper=time_stepper,
parameter_space=d.parameter_space, visualizer=d.visualizer,
num_values=num_values, name='{}_FV'.format(p.name))
return discretization, data
