ddg.optimize.he.functional module

class ddg.optimize.he.functional.HalfEdgeFunctional(surface, attr_name, cell_type, attr_dimension=None, boundary_cells={}, interior_cell_index_attr_name='interior_cell_index')

Bases: ABC

Abstract base class for functionals on half-edge objects. When creating functionals inheriting from this class the __init__ function should be inherited as the defined attributes are being used in optimization.

Parameters:

surface

attr_name

attr_dimension

boundary_cells

interior_cells

Notes

The functional can additionally implement the following methods which are then used in the optimization process: - gradient - hessian

Attributes:

surfaceddg.datastructures.halfedge.Surface

Half-edge surface to work on

attr_namestr

Name of the attribute for minimization

attr_dimensionint (optional)

Dimension of the attribute. If not given it will be determined automatically.

cell_typestr

Either “verts”, “edges” or “faces” depending on where the minimization attribute is located

boundary_cellsset (default = set())

Boundary cells for the minimization, i.e. vertices that have and keep a fixed value and are excluded from minimization process

interior_cell_index_attr_namestr (default = “interior_cell_index”)

Attribute name for indices of interior cells. The interior cells are enumerated using this attribute.

Methods

evaluate(x)

Evaluation function of the functional

abstract evaluate(x) → float

Evaluation function of the functional

Parameters:

x: numpy.ndarray of shape (n, d)

Variables of the minimization process, where n is the number of interior cells, and d=attr_dimension. The n interior cells are ordered according to the attribute given by interior_cell_index_attr_name. The values of the attribute on the boundary cells need to be retrieved directly from the surface.

Returns:

float

Value of the functional

property interior_cells

Method to distinguish the interior cells of the surface from the specified boundary cells. Equal to the complement of boundary_cells attribute in the surface.