kipe.fenicsx.sampling module

Spatial samplers on meshes: fenicsx.mesh.

MeshSampler interpolates a field of the solver onto a measurement mesh, which need not match the solver’s mesh, optionally projected onto a direction or the surface normal. Its entries are in the order of the measurement mesh file: one value per vertex for P1 (by the vertex’s row in the file), one per cell for DG0 (by the cell’s row); for a vector field without projection, the components of each vertex or cell one after another. dolfinx keeps these rows however it distributes and renumbers the mesh, so the data do not depend on the number of ranks (see kipe.sampling).

type kipe.fenicsx.sampling.Element = Literal['P1', 'DG0']
type kipe.fenicsx.sampling.Projection = list[float] | Literal['surface_normal'] | None
class kipe.fenicsx.sampling.MeshSampler(field: str, source: dolfinx.fem.FunctionSpace, mesh: str, element: Element, projection: Projection = None, padding: float = 0.0, mesh_name: str = 'mesh', comm: mpi4py.MPI.Comm = mpi4py.MPI.COMM_WORLD)[source]

Bases: SpatialSampler

Interpolates a field onto a measurement mesh, optionally projected.

Parameters:

  • field – name of the observed field

  • source – the field’s function space, on the solver’s mesh

  • mesh – path to the measurement mesh, an XDMF file

  • element – finite element on the measurement mesh

  • projection – a direction (one entry per component of the field) or "surface_normal" (the measurement mesh must be a surface); all components if None

  • padding – distance by which a measurement point may lie outside the solver’s mesh

  • mesh_name – name of the mesh in the XDMF file

  • comm – communicator of the run

Raises:

StudyFileError – if the file has no mesh named mesh_name, the measurement mesh has curved cells (P1 needs a linear geometry), the projection does not fit the field or the mesh, or points of the measurement mesh lie outside the solver’s mesh, beyond the padding

sample(fields: Mapping[str, NDArray_f64]) → NDArray_f64[source]

Interpolate the field onto the measurement mesh, and project it.

Parameters:

fields – the observed state field, this rank’s owned degrees of freedom

Returns:

the values of this rank’s owned degrees of freedom on the measurement mesh (a new array), at the positions global_indices

kipe.fenicsx.sampling.build_mesh_sampler(options: MeshSamplingOptions, fields: list[str], solver: ForwardSolver, comm: mpi4py.MPI.Comm = mpi4py.MPI.COMM_WORLD) → MeshSampler[source]

Construct a MeshSampler from the spatial_sampling options of a measurement.

Parameters:

  • options – spatial_sampling options of the measurement

  • fields – names of the fields the measurement observes; exactly one

  • solver – the forward solver, which provides the field’s function space

  • comm – communicator of the run

Returns:

the sampler

Raises:

StudyFileError – if the measurement does not observe exactly one distributed field whose function space the solver provides

kipe.fenicsx.sampling.observed_function_space(kind: str, fields: list[str], solver: ForwardSolver) → tuple[str, dolfinx.fem.FunctionSpace][source]

Return the field a FEniCSx sampler observes, and its function space.

Parameters:

  • kind – the sampler’s type, for error messages, e.g., "fenicsx.mesh"

  • fields – names of the fields the measurement observes; exactly one

  • solver – the forward solver, which provides the field’s function space

Returns:

  • the field’s name

  • its function space, on the solver’s mesh

Raises:

StudyFileError – if the measurement does not observe exactly one distributed field whose function space the solver provides