Namespaces
	detail

	experimental

	internal

Classes
class	Delaunay
	A class for incremental generation of a 2D or 3D Delaunay triangulation. More...

class	DiscreteShape
	Post-processed klee::Shape, with the geometry discretized and transformed according to the Shape's operators. More...

class	DistributedClosestPoint
	Encapsulated the Distributed closest point query for a collection of query points over an "object mesh". More...

class	GWNMomentData

class	NURBSCurveGWNQuery
	Preprocesses NURBSCurve geometry for GWN evaluation, and performs the calculation on the DOFs of an input MFEM mesh. More...

class	PolylineGWNQuery
	Preprocesses a linear mesh for GWN evaluation, and performs the calculation on the DOFs of an input MFEM mesh. More...

class	NURBSPatchGWNQuery
	Preprocesses NURBSPatch geometry for GWN evaluation, and performs the calculation on the DOFs of an input MFEM mesh. More...

class	TriangleGWNQuery
	Preprocesses a triangle mesh for GWN evaluation, and performs the calculation on the DOFs of an input MFEM mesh. More...

struct	gwn_input_traits

struct	gwn_input_traits< axom::quest::PolylineGWNQuery< ExecSpace, ORDER > >

struct	gwn_input_traits< axom::quest::NURBSCurveGWNQuery< ExecSpace, ORDER > >

struct	gwn_input_traits< axom::quest::TriangleGWNQuery< ExecSpace, ORDER > >

struct	gwn_input_traits< axom::quest::NURBSPatchGWNQuery< ExecSpace, ORDER > >

struct	FieldStats

struct	IntegralStats

class	InOutOctree

class	TempArrayView
	Provides a view over an MFEM grid function. MFEM grid functions are assumed to live in host memory. This class performs data movement needed to access the grid function data within a GPU device lambda. This view is limited in scope, though could be expanded in the future. More...

class	IntersectionShaper

class	LinearizeCurves
	This class linearizes a vector of NURBSCurve objects into line segments stored in a mint mesh. More...

class	PointInCellTraits
	A traits class for the mesh associated with a PointInCell query. More...

class	PointInCell
	A class to accelerate Point-In-Cell queries on a computational mesh. More...

class	SamplingShaper
	Concrete class for sample based shaping. More...

class	ScatteredInterpolation
	A class to perform scattered data interpolation at arbitrary points over an input point set. More...

class	Shaper

class	SignedDistance

Enumerations
enum	SearchStatus { NEIGHBOR_NOT_FOUND = -1 }

enum class	GWNInputType { Curve , Polyline , Surface , Triangulation }

enum class	WatertightStatus : signed char { WATERTIGHT = 0 , NOT_WATERTIGHT , CHECK_FAILED }

enum class	SignedDistExec { CPU = 0 , OpenMP = 1 , GPU = 2 }

Functions
template<typename T , int NDIMS, int ORD, typename LeafGeometry , typename TraverserType >
double	fast_approximate_winding_number (const primal::Point< T, NDIMS > &query, const TraverserType &traverser, const ArrayView< LeafGeometry > &leaf_objects, const ArrayView< GWNMomentData< T, NDIMS, ORD >> &internal_moments, const primal::WindingTolerances &wt, double beta=2.0)
	Evaluate a hierarchical approximation of the GWN for the shape in `leaf_objects`, encompassed by `traverser`'s BVH. More...

template<typename T >
axom::Array< primal::NURBSCurve< T, 2 > >	subdivide_curves (const axom::ArrayView< const primal::NURBSCurve< T, 2 >> &input_curves_view, double bbox_threshold, int max_curves=1e6, int npasses=10)
	Subdivides an array of NURBS curves to limit their maximum AABB diagonal. More...

template<typename T >
axom::Array< primal::NURBSPatch< T, 3 > >	subdivide_patches (const axom::ArrayView< const primal::NURBSPatch< T, 3 >> &input_patches_view, double bbox_threshold, int max_patches=1e5, int npasses=10)
	Subdivides an array of trimmed NURBS surfaces to limit their maximum AABB diagonal. More...

void	setup_gwn_mesh (mfem::DataCollection &dc, mfem::Mesh *query_mesh, int queryOrder)

template<int NDIMS>
void	generate_gwn_query_mesh (mfem::DataCollection &dc, const axom::primal::BoundingBox< double, NDIMS > &shape_bbox, const std::vector< double > &boxMins, const std::vector< double > &boxMaxs, const std::vector< int > &boxResolution, int queryOrder)

FieldStats	compute_field_stats (const mfem::GridFunction &gf)

IntegralStats	compute_integrals (const mfem::GridFunction &gf)

Nearest Neighbor query
void	all_nearest_neighbors (const double x, const double y, const double z, const int region, int n, double limit, int neighbor, double sqdistance)
	Given a list of point locations and regions, for each point, find the closest point in a different region within a given search radius. More...

Visualize octahedra as tet mesh
int	mesh_from_discretized_polyline (const axom::ArrayView< OctType > &octs, int octcount, int segcount, mint::Mesh *&mesh)
	Produces a mesh of tets from an array of Octahedra. More...

Mesh test and repair
template<typename ExecSpace , typename FloatType >
void	findTriMeshIntersectionsBVH (mint::UnstructuredMesh< mint::SINGLE_SHAPE > *surface_mesh, std::vector< std::pair< int, int >> &intersections, std::vector< int > &degenerateIndices, double intersectionThreshold=1E-8)
	Find self-intersections and degenerate triangles in a surface mesh utilizing a Bounding Volume Hierarchy. More...

template<typename ExecSpace , typename FloatType >
void	findTriMeshIntersectionsImplicitGrid (mint::UnstructuredMesh< mint::SINGLE_SHAPE > *surface_mesh, std::vector< std::pair< int, int >> &intersections, std::vector< int > &degenerateIndices, int spatialIndexResolution=0, double intersectionThreshold=1E-8)
	Find self-intersections and degenerate triangles in a surface mesh utilizing an implicit grid spatial index. More...

template<typename ExecSpace , typename FloatType >
void	findTriMeshIntersectionsUniformGrid (mint::UnstructuredMesh< mint::SINGLE_SHAPE > *surface_mesh, std::vector< std::pair< int, int >> &intersections, std::vector< int > &degenerateIndices, int spatialIndexResolution=0, double intersectionThreshold=1E-8)
	Find self-intersections and degenerate triangles in a surface mesh utilizing an uniform grid spatial index. More...

void	findTriMeshIntersections (mint::UnstructuredMesh< mint::SINGLE_SHAPE > *surface_mesh, std::vector< std::pair< int, int >> &intersections, std::vector< int > &degenerateIndices, int spatialIndexResolution=0, double intersectionThreshold=1E-8)
	Find self-intersections and degenerate triangles in a surface mesh utilizing a Uniform Grid. More...

WatertightStatus	isSurfaceMeshWatertight (mint::UnstructuredMesh< mint::SINGLE_SHAPE > *surface_mesh)
	Check a surface mesh for holes using its face relation. More...

void	weldTriMeshVertices (mint::UnstructuredMesh< mint::SINGLE_SHAPE > **surface_mesh, double eps)
	Mesh repair function to weld vertices that are closer than eps. More...

InOut query – initialization and finalization functions
int	inout_init (const std::string &file, MPI_Comm comm=MPI_COMM_SELF)
	Initializes the quest inout query from a mesh file. More...

int	inout_init (std::shared_ptr< mint::Mesh > &mesh, MPI_Comm comm=MPI_COMM_SELF)
	Initialize the inout query using a pre-loaded mesh. More...

int	inout_finalize ()
	Finalizes the inout query. More...

bool	inout_initialized ()
	Predicate to test whether the inout query has been initialized. More...

InOut query – properties and querying functions
Note These must be called after inout_init()
bool	inout_evaluate (double x, double y, double z=0.)
	Tests if the point (x, y, z) is inside the contained volume. More...

int	inout_evaluate (const double x, const double y, const double z, int npoints, int res)
	Tests an array of points for containment. More...

int	inout_mesh_min_bounds (double *coords)
	Returns the lower coordinates of the mesh's bounding box. More...

int	inout_mesh_max_bounds (double *coords)
	Returns the upper coordinates of the mesh's bounding box. More...

int	inout_mesh_center_of_mass (double *coords)
	Returns the center of mass of the mesh. More...

int	inout_get_dimension ()
	Gets the spatial dimension of the query. More...

InOut query – setup options and parameters
Note These must be called before inout_init()
int	inout_set_dimension (int dim)
	Sets the spatial dimension of the query. More...

int	inout_set_verbose (bool verbosity)
	Enables/disables verbose logging output. More...

int	inout_set_vertex_weld_threshold (double thresh)
	Sets the cutoff distance for welding vertices during initialization. More...

int	inout_set_segments_per_knot_span (int segmentsPerKnotSpan)
	Sets the number of samples for each knot span (2D only) More...

Signed Distance Query Initialization Methods
int	signed_distance_init (const std::string &file, MPI_Comm comm=MPI_COMM_SELF)
	Initializes the Signed Distance Query with a surface given in an STL formatted file. More...

int	signed_distance_init (const mint::Mesh *m, MPI_Comm comm=MPI_COMM_SELF)
	Initializes the Signed Distance Query with the given surface mesh. More...

bool	signed_distance_initialized ()
	Checks if the Signed Distance Query has been initialized. More...

Signed Distance Query Options
void	signed_distance_set_dimension (int dim)
	Sets the dimension for the Signed Distance Query. More...

void	signed_distance_set_closed_surface (bool status)
	Indicates whether the input to the signed distance consists of a water-tight surface mesh, or not. More...

void	signed_distance_set_compute_signs (bool computeSign)
	Sets whether the distance query should compute or ignore the sign. More...

void	signed_distance_set_allocator (int allocatorID)
	Sets the allocator to use for creating internal signed distance query data structures. More...

void	signed_distance_set_verbose (bool status)
	Enables/Disables verbose output for the Signed Distance Query. More...

void	signed_distance_use_shared_memory (bool status)
	Enable/Disable the use of on-node shared memory (via Umpire's shared-memory resource) for storing the surface mesh. By default this option is disabled. More...

void	signed_distance_set_shared_memory_size (std::size_t min_segment_size)
	Set the minimum size (in bytes) of the backing shared-memory segment used to store the surface mesh when signed_distance_use_shared_memory(true) is enabled. More...

void	signed_distance_set_execution_space (SignedDistExec execSpace)
	Set the execution space in which to run signed distance queries. By default this option is set to SIGNED_DIST_EVAL_CPU. More...

Signed Distance Query Evaluation Methods
double	signed_distance_evaluate (double x, double y, double z=0.0)
	Evaluates the signed distance function at the given point. More...

double	signed_distance_evaluate (double x, double y, double z, double &cp_x, double &cp_y, double &cp_z, double &n_x, double &n_y, double &n_z)
	Evaluates the signed distance function at the given 3D point. More...

void	signed_distance_evaluate (const double x, const double y, const double z, int npoints, double phi)
	Evaluates the signed distance function at the given set of points. More...

void	signed_distance_get_mesh_bounds (double lo, double hi)
	Computes the bounds of the specified input mesh supplied to the Signed Distance Query. More...

Signed Distance Query Finalization Methods
void	signed_distance_finalize ()
	Finalizes the SignedDistance query. More...

Variables
template<typename GWNQueryType >
constexpr GWNInputType	gwn_input_type_v = gwn_input_traits<GWNQueryType>::value

constexpr int	QUEST_INOUT_SUCCESS = 0

constexpr int	QUEST_INOUT_FAILED = -1

Discretize primitive shapes to linear shapes
using	SphereType = primal::Sphere< double, 3 >

using	OctType = primal::Octahedron< double, 3 >

using	Point2D = primal::Point< double, 2 >

bool	discretize (const SphereType &s, int levels, axom::Array< OctType > &out, int &octcount)
	Given a primitive sphere and a refinement level, allocate and return a list of Octahedra approximating the shape. More...

template<typename ExecSpace >
bool	discretize (const axom::ArrayView< Point2D > &polyline, int len, int levels, axom::Array< OctType > &out, int &octcount)
	Given a 2D polyline revolved around the positive X-axis, allocate and return a list of Octahedra approximating the shape. More...

Typedef Documentation

◆ SphereType

using axom::quest::SphereType = typedef primal::Sphere<double, 3>

◆ OctType

using axom::quest::OctType = typedef primal::Octahedron<double, 3>

◆ Point2D

using axom::quest::Point2D = typedef primal::Point<double, 2>

Enumeration Type Documentation

◆ SearchStatus

enum axom::quest::SearchStatus

Enumerator
NEIGHBOR_NOT_FOUND

◆ GWNInputType

enum axom::quest::GWNInputType

strong

Enumerator
Curve
Polyline
Surface
Triangulation

◆ WatertightStatus

enum axom::quest::WatertightStatus : signed char

strong

Enumeration indicating mesh watertightness

Enumerator
WATERTIGHT	Each edge in a surface mesh is incident in two cells.
NOT_WATERTIGHT	Each edge is incident in one or two cells.
CHECK_FAILED	Calculation failed (possibly a non-manifold mesh)

◆ SignedDistExec

enum axom::quest::SignedDistExec

strong

Enumerator
CPU
OpenMP
GPU

Function Documentation

◆ all_nearest_neighbors()

void axom::quest::all_nearest_neighbors	(	const double *	x,
		const double *	y,
		const double *	z,
		const int *	region,
		int	n,
		double	limit,
		int *	neighbor,
		double *	sqdistance
	)

Given a list of point locations and regions, for each point, find the closest point in a different region within a given search radius.

Parameters

[in]	x	X-coordinates of input points
[in]	y	Y-coordinates of input points
[in]	z	Z-coordinates of input points
[in]	region	Region of each point
[in]	n	Number of points
[in]	limit	Max distance for all-nearest-neighbors query
[out]	neighbor	Index of nearest neighbor not in the same class (or NEIGHBOR_NOT_FOUND)
[out]	sqdistance	Squared distance to nearest neighbor

Precondition: x, y, z, and region have n entries; neighbor is allocated with room for n entries

This method inserts all points p at (x[i], y[i], z[i]) into a UniformGrid index. Then for each point p, it gets the UniformGrid bins that overlap the box (p - (limit, limit, limit), p + (limit, limit, limit). The method compares p to each point in this list of bins and returns the index of the closest point.

We expect the use of the UniformGrid will result in a substantial time savings over a brute-force all-to-all algorithm, but the query's run time is dependent on the point distribution.

◆ discretize() [1/2]

bool axom::quest::discretize	(	const SphereType &	s,
		int	levels,
		axom::Array< OctType > &	out,
		int &	octcount
	)

Given a primitive sphere and a refinement level, allocate and return a list of Octahedra approximating the shape.

Parameters

[in]	s	The sphere to approximate
[in]	levels	The number of refinements to perform, in addition to a central level-zero octahedron
[out]	out	The newly-initialized Array of octahedra representing s
[out]	octcount	The number of elements in out

Returns: false for invalid input or error in computation; true otherwise

This routine generates O(4^level) octahedra. That's exponential growth. Use appropriate caution.

This routine initializes an Array, out.

◆ discretize() [2/2]

template<typename ExecSpace >

bool axom::quest::discretize	(	const axom::ArrayView< Point2D > &	polyline,
		int	len,
		int	levels,
		axom::Array< OctType > &	out,
		int &	octcount
	)

Given a 2D polyline revolved around the positive X-axis, allocate and return a list of Octahedra approximating the shape.

Parameters

[in]	polyline	The polyline to revolve around the X-axis. Data should be in a host-accessible memory space.
[in]	len	The number of points in polyline
[in]	levels	The number of refinements to perform, in addition to a central level-zero octahedron in each segment
[out]	out	The newly-initialized Array of octahedra representing the revolved polyline
[out]	octcount	The number of elements in out

Returns: false for invalid input or error in computation; true otherwise

This routine generates n*O(2^level) octahedra, where n is the number of segments in polyline (one less than the length). That's exponential growth. Use appropriate caution.

This routine initializes an Array, out.

◆ mesh_from_discretized_polyline()

int axom::quest::mesh_from_discretized_polyline	(	const axom::ArrayView< OctType > &	octs,
		int	octcount,
		int	segcount,
		mint::Mesh *&	mesh
	)

Produces a mesh of tets from an array of Octahedra.

Parameters

[in]	octs	Array of Octahedron objects
[in]	octcount	Length of octs
[in]	segcount	Number of segments in the polyline originating the octs
[out]	mesh	Pointer to a mesh object where the mesh will be generated.

Returns: status error code, zero on success

This function creates a new Mint mesh out of an array of Octahedron objects. The mesh will be valid for any oct array, but the fields it includes make sense for the output of discretizing a revolved polyline:

octahedron_volume: octahedron volume calculated by summing tetrahedron volumes.
oct_as_polyhedron_volume: octahedron volume calculated by changing the oct into a Polyhedron and reporting the Polyhedron's volume. Any discrepancy between this field and octahedron_volume is a bug.
segment_index: the zero-based index of the segment to which the parent octahedron belongs.
octahedron_index: the zero-based index of the parent octahedron within its segment.
level_of_refinement: the level of refinement to which parent octahedron belongs.

Note: Ownership of the mesh object is passed to the caller. Consequently, the caller is responsible for properly deallocating the mesh object that the return mesh pointer points to.

◆ fast_approximate_winding_number()

template<typename T , int NDIMS, int ORD, typename LeafGeometry , typename TraverserType >

double axom::quest::fast_approximate_winding_number	(	const primal::Point< T, NDIMS > &	query,
		const TraverserType &	traverser,
		const ArrayView< LeafGeometry > &	leaf_objects,
		const ArrayView< GWNMomentData< T, NDIMS, ORD >> &	internal_moments,
		const primal::WindingTolerances &	wt,
		double	beta = `2.0`
	)

Evaluate a hierarchical approximation of the GWN for the shape in leaf_objects, encompassed by traverser's BVH.

Traverse the BVH at the input query. For any node which is considered "far-away", approximate the GWN for that node For any node which is a leaf (i.e. close to the query), compute the GWN directly

Template Parameters

T	Numeric type of geometric primitives
NDIMS	The spatial dimensions of the shape (2 or 3)
ORD	The number of terms in the Taylor expansion
LeafGeometry	The type of the objects which could require direct evaluation
TraverserType	The derived type of the traverser for the BVH tree

Parameters

[in]	query	The query point at which to evaluate the GWN
[in]	traverser	A traverser for the BVH tree
[in]	leaf_objects	A view of all individual geometric objects in the shape, indexed by position in the BVH tree
[in]	internal_moments	A view of GWNMomentData objects for each internal node of the BVH tree, each representing a cluster of leaf objects
[in]	wt	A structure of possible tolerances for GWN evaluation to permit flexible evaluation for many different type of leaf objects
[in]	beta	An "accuracy parameter" which scales at what distance from an AABB a query point is considered to be "far-away".

The default parameter beta = 2.0 is suggested by the work "Fast Winding Numbers for Soups and Clouds" by Barill et al. (2018)

Returns: The approximated GWN at the query point

References axom::primal::WindingTolerances::disk_size, axom::primal::WindingTolerances::edge_tol, axom::primal::WindingTolerances::EPS, axom::primal::WindingTolerances::ls_tol, axom::primal::WindingTolerances::quad_tol, axom::primal::squared_distance(), and axom::primal::winding_number().

◆ subdivide_curves()

template<typename T >

axom::Array<primal::NURBSCurve<T, 2> > axom::quest::subdivide_curves	(	const axom::ArrayView< const primal::NURBSCurve< T, 2 >> &	input_curves_view,
		double	bbox_threshold,
		int	max_curves = `1e6`,
		int	npasses = `10`
	)

Subdivides an array of NURBS curves to limit their maximum AABB diagonal.

Parameters

[in]	input_curves_view	A const view to the const NURBS input curves
[in]	bbox_threshold	The maximum AABB diagonal of each subdivided curve, as a percent of an AABB of the entire input shape
[in]	max_curves	The maximum number of curves after which further curves will not be subdivided
[in]	npasses	The maximum number of passes through the input curves before an early return

Starts with a collection of Bezier-extracted components, then iterates through the curves at most npasses times, and subdivides any which have an AABB diagonal greater than the threshold. The AABB of the entire shape is based on the union of AABB for the subdivisions, which is tighter than that of the original

Returns: The array of subdivided NURBS curves.

References axom::Array< T, DIM, SPACE, StoragePolicy >::back(), axom::Array< T, DIM, SPACE, StoragePolicy >::clear(), axom::Array< T, DIM, SPACE, StoragePolicy >::emplace_back(), axom::Array< T, DIM, SPACE, StoragePolicy >::push_back(), axom::Array< T, DIM, SPACE, StoragePolicy >::size(), SLIC_WARNING, SLIC_WARNING_IF, and axom::Array< T, DIM, SPACE, StoragePolicy >::swap().

◆ subdivide_patches()

template<typename T >

axom::Array<primal::NURBSPatch<T, 3> > axom::quest::subdivide_patches	(	const axom::ArrayView< const primal::NURBSPatch< T, 3 >> &	input_patches_view,
		double	bbox_threshold,
		int	max_patches = `1e5`,
		int	npasses = `10`
	)

Subdivides an array of trimmed NURBS surfaces to limit their maximum AABB diagonal.

Parameters

[in]	input_patches_view	A const view to the const NURBS input surfaces
[in]	bbox_threshold	The maximum AABB diagonal of each subdivided surface, as a percent of an AABB of the entire input shape
[in]	max_patches	The maximum number of surfaces before additional surfaces will not be subdivided
[in]	npasses	The maximum number of passes through the input surfaces before an early return

Iterates through the curves at most npasses times, and subdivides any which have an AABB diagonal greater than the threshold. The AABB of the entire shape is based on the union of AABB for the subdivisions, which is tighter than that of the original Subdivision is performed on the single axis which is determined to be "longer" in physical space, as determined by NURBSPatch::nearBisectOnLongerAxis

Returns: The array of subdivided NURBS surfaces.

References axom::Array< T, DIM, SPACE, StoragePolicy >::back(), axom::Array< T, DIM, SPACE, StoragePolicy >::clear(), axom::Array< T, DIM, SPACE, StoragePolicy >::emplace_back(), axom::Array< T, DIM, SPACE, StoragePolicy >::push_back(), axom::Array< T, DIM, SPACE, StoragePolicy >::reserve(), axom::Array< T, DIM, SPACE, StoragePolicy >::size(), SLIC_WARNING, and axom::Array< T, DIM, SPACE, StoragePolicy >::swap().

◆ setup_gwn_mesh()

void axom::quest::setup_gwn_mesh	(	mfem::DataCollection &	dc,
		mfem::Mesh *	query_mesh,
		int	queryOrder
	)

Helper function to set up the mesh and associated winding and inout fields. Uses an mfem::DataCollection to hold everything together.

References AXOM_ANNOTATE_SCOPE.

◆ generate_gwn_query_mesh()

template<int NDIMS>

void axom::quest::generate_gwn_query_mesh	(	mfem::DataCollection &	dc,
		const axom::primal::BoundingBox< double, NDIMS > &	shape_bbox,
		const std::vector< double > &	boxMins,
		const std::vector< double > &	boxMaxs,
		const std::vector< int > &	boxResolution,
		int	queryOrder
	)

Query grid setup; has some dimension-specific types; if user did not provide a bounding box, use shape bounding box scaled by 10%

References AXOM_ANNOTATE_SCOPE, axom::primal::BoundingBox< T, NDIMS >::getMax(), axom::primal::BoundingBox< T, NDIMS >::getMin(), setup_gwn_mesh(), and SLIC_INFO.

◆ compute_field_stats()

FieldStats axom::quest::compute_field_stats ( const mfem::GridFunction & gf )

References axom::quest::FieldStats::dof_l2, and axom::utilities::max().

◆ compute_integrals()

IntegralStats axom::quest::compute_integrals ( const mfem::GridFunction & gf )

◆ findTriMeshIntersectionsBVH()

template<typename ExecSpace , typename FloatType >

void axom::quest::findTriMeshIntersectionsBVH	(	mint::UnstructuredMesh< mint::SINGLE_SHAPE > *	surface_mesh,
		std::vector< std::pair< int, int >> &	intersections,
		std::vector< int > &	degenerateIndices,
		double	intersectionThreshold = `1E-8`
	)

Find self-intersections and degenerate triangles in a surface mesh utilizing a Bounding Volume Hierarchy.

Parameters

[in]	surface_mesh	A triangle mesh in three dimensions
[out]	intersection	Pairs of indices of intersecting mesh triangles
[out]	degenerateIndices	indices of degenerate mesh triangles
[in]	intersectionThreshold	Tolerance threshold for triangle intersection tests (default: 1E-8) After running this function over a surface mesh, intersection will be filled with pairs of indices of intersecting triangles and degenerateIndices will be filled with the indices of the degenerate triangles in the mesh. Triangles that share vertex pairs (adjacent triangles in a watertight surface mesh) are not reported as intersecting. Degenerate triangles are not reported as intersecting other triangles.

References AXOM_ANNOTATE_SCOPE, axom::Array< T, DIM, SPACE, StoragePolicy >::begin(), axom::Array< T, DIM, SPACE, StoragePolicy >::end(), axom::Array< T, DIM, SPACE, StoragePolicy >::resize(), axom::Array< T, DIM, SPACE, StoragePolicy >::size(), and SLIC_INFO.

◆ findTriMeshIntersectionsImplicitGrid()

template<typename ExecSpace , typename FloatType >

void axom::quest::findTriMeshIntersectionsImplicitGrid	(	mint::UnstructuredMesh< mint::SINGLE_SHAPE > *	surface_mesh,
		std::vector< std::pair< int, int >> &	intersections,
		std::vector< int > &	degenerateIndices,
		int	spatialIndexResolution = `0`,
		double	intersectionThreshold = `1E-8`
	)

Find self-intersections and degenerate triangles in a surface mesh utilizing an implicit grid spatial index.

Parameters

[in]	surface_mesh	A triangle mesh in three dimensions
[out]	intersection	Pairs of indices of intersecting mesh triangles
[out]	degenerateIndices	indices of degenerate mesh triangles
[in]	spatialIndexResolution	The grid resolution for the index structure (default: 0)
[in]	intersectionThreshold	Tolerance threshold for triangle intersection tests (default: 1E-8) After running this function over a surface mesh, intersection will be filled with pairs of indices of intersecting triangles and degenerateIndices will be filled with the indices of the degenerate triangles in the mesh. Triangles that share vertex pairs (adjacent triangles in a watertight surface mesh) are not reported as intersecting. Degenerate triangles are not reported as intersecting other triangles.

This function uses a quest::ImplicitGrid spatial index. Input spatialIndexResolution specifies the bin size for the UniformGrid. The default value of 0 causes this routine to calculate a heuristic bin size based on the cube root of the number of cells in the mesh.

References AXOM_ANNOTATE_SCOPE, axom::Array< T, DIM, SPACE, StoragePolicy >::begin(), axom::Array< T, DIM, SPACE, StoragePolicy >::end(), axom::Array< T, DIM, SPACE, StoragePolicy >::resize(), axom::Array< T, DIM, SPACE, StoragePolicy >::size(), and SLIC_INFO.

◆ findTriMeshIntersectionsUniformGrid()

template<typename ExecSpace , typename FloatType >

void axom::quest::findTriMeshIntersectionsUniformGrid	(	mint::UnstructuredMesh< mint::SINGLE_SHAPE > *	surface_mesh,
		std::vector< std::pair< int, int >> &	intersections,
		std::vector< int > &	degenerateIndices,
		int	spatialIndexResolution = `0`,
		double	intersectionThreshold = `1E-8`
	)

Find self-intersections and degenerate triangles in a surface mesh utilizing an uniform grid spatial index.

Parameters

[in]	surface_mesh	A triangle mesh in three dimensions
[out]	intersection	Pairs of indices of intersecting mesh triangles
[out]	degenerateIndices	indices of degenerate mesh triangles
[in]	spatialIndexResolution	The grid resolution for the index structure (default: 0)
[in]	intersectionThreshold	Tolerance threshold for triangle intersection tests (default: 1E-8) After running this function over a surface mesh, intersection will be filled with pairs of indices of intersecting triangles and degenerateIndices will be filled with the indices of the degenerate triangles in the mesh. Triangles that share vertex pairs (adjacent triangles in a watertight surface mesh) are not reported as intersecting. Degenerate triangles are not reported as intersecting other triangles.

This function uses a quest::UniformGrid spatial index. Input spatialIndexResolution specifies the bin size for the UniformGrid. The default value of 0 causes this routine to calculate a heuristic bin size based on the cube root of the number of cells in the mesh.

References AXOM_ANNOTATE_SCOPE, axom::Array< T, DIM, SPACE, StoragePolicy >::begin(), axom::Array< T, DIM, SPACE, StoragePolicy >::end(), axom::Array< T, DIM, SPACE, StoragePolicy >::resize(), axom::Array< T, DIM, SPACE, StoragePolicy >::size(), and SLIC_INFO.

◆ findTriMeshIntersections()

void axom::quest::findTriMeshIntersections	(	mint::UnstructuredMesh< mint::SINGLE_SHAPE > *	surface_mesh,
		std::vector< std::pair< int, int >> &	intersections,
		std::vector< int > &	degenerateIndices,
		int	spatialIndexResolution = `0`,
		double	intersectionThreshold = `1E-8`
	)

Find self-intersections and degenerate triangles in a surface mesh utilizing a Uniform Grid.

Parameters

[in]	surface_mesh	A triangle mesh in three dimensions
[out]	intersection	Pairs of indices of intersecting mesh triangles
[out]	degenerateIndices	indices of degenerate mesh triangles
[in]	spatialIndexResolution	The grid resolution for the index structure (default: 0)
[in]	intersectionThreshold	Tolerance threshold for triangle intersection tests (default: 1E-8)

After running this function over a surface mesh, intersection will be filled with pairs of indices of intersecting triangles and degenerateIndices will be filled with the indices of the degenerate triangles in the mesh. Triangles that share vertex pairs (adjacent triangles in a watertight surface mesh) are not reported as intersecting. Degenerate triangles are not reported as intersecting other triangles.

This function uses a quest::UniformGrid spatial index. Input spatialIndexResolution specifies the bin size for the UniformGrid. The default value of 0 causes this routine to calculate a heuristic bin size based on the cube root of the number of cells in the mesh.

◆ isSurfaceMeshWatertight()

WatertightStatus axom::quest::isSurfaceMeshWatertight ( mint::UnstructuredMesh< mint::SINGLE_SHAPE > * surface_mesh )

Check a surface mesh for holes using its face relation.

Parameters

[in] surface_mesh A surface mesh in three dimensions

Returns: status If the mesh is watertight, is not watertight, or if an error occurred (possibly due to non-manifold mesh).

Note: This method marks the cells on the boundary by creating a new cell-centered field variable, called "boundary", on the given input mesh.; This function computes the mesh's cell-face and face-vertex relations. For large meshes, this can take a long time. The relations are used to check for holes, and remain cached with the mesh after this function finishes.

◆ weldTriMeshVertices()

void axom::quest::weldTriMeshVertices	(	mint::UnstructuredMesh< mint::SINGLE_SHAPE > **	surface_mesh,
		double	eps
	)

Mesh repair function to weld vertices that are closer than eps.

Parameters

[in,out]	surface_mesh	A pointer to a pointer to a triangle mesh
[in]	eps	Distance threshold for welding vertices (using the max norm)

Precondition: eps must be greater than zero; surface_mesh is a pointer to a pointer to a non-null triangle mesh.

Postcondition: The triangles of surface_mesh are reindexed using the welded vertices and degenerate triangles are removed. The mesh can still contain vertices that are not referenced by any triangles.

This utility function repairs an input triangle mesh (embedded in three dimensional space) by 'welding' vertices that are closer than eps. The vertices are quantized to an integer lattice with spacing eps and vertices that fall into the same cell on this lattice are identified. All identified vertices are given the coordinates of the first such vertex and all incident triangles use the same index for this vertex.

The input mesh can be a "soup of triangles", where the vertices of adjacent triangles have distinct indices. After running this function, vertices that are closer than eps are welded, and their incident triangles use the new vertex indices. Thus, the output is an "indexed triangle mesh".

This function also removes degenerate triangles from the mesh. These are triangles without three distinct vertices after the welding.

Note: This function is destructive. It modifies the input triangle mesh in place.; The distance metric in this function uses the "max" norm (l_inf).

◆ inout_init() [1/2]

int axom::quest::inout_init	(	const std::string &	file,
		MPI_Comm	comm = `MPI_COMM_SELF`
	)

Initializes the quest inout query from a mesh file.

Parameters

[in]	file	Path to an STL file containing the surface mesh
[in]	comm	The MPI communicator (when running in parallel)

Returns: Return code is QUEST_INOUT_SUCCESS if successful and QUEST_INOUT_FAILED otherwise.

Precondition: inout_initialized() == false

Postcondition: inout_initialized() == true, when rc is QUEST_INOUT_SUCCESS

◆ inout_init() [2/2]

int axom::quest::inout_init	(	std::shared_ptr< mint::Mesh > &	mesh,
		MPI_Comm	comm = `MPI_COMM_SELF`
	)

Initialize the inout query using a pre-loaded mesh.

Parameters

[in,out]	mesh	Pointer to the input mesh. This pointer will be updated during this invocation
[in]	comm	The MPI communicator (when running in parallel)

Returns: Return code is QUEST_INOUT_SUCCESS if successful and QUEST_INOUT_FAILED otherwise.

Precondition: inout_initialized() == false

Postcondition: inout_initialized() == true, when rc is QUEST_INOUT_SUCCESS

Warning: The underlying data structure modifies the input mesh (e.g. by welding vertices) and updates the mesh pointer. It is the user's responsibility to update any other pointers to this same mesh.

◆ inout_finalize()

int axom::quest::inout_finalize ( )

Finalizes the inout query.

Postcondition: inout_initialized() == false

◆ inout_initialized()

bool axom::quest::inout_initialized ( )

Predicate to test whether the inout query has been initialized.

Returns: True if the inout query has been initialized, false otherwise.

◆ inout_evaluate() [1/2]

bool axom::quest::inout_evaluate	(	double	x,
		double	y,
		double	z = `0.`
	)

Tests if the point (x, y, z) is inside the contained volume.

Parameters

[in]	x	The x-coordinate of the query point
[in]	y	The y-coordinate of the query point
[in]	z	The z-coordinate of the query point

Returns: True if the point is within the contained volume or on the surface mesh, false otherwise.

Precondition: inout_initialized() == true

◆ inout_evaluate() [2/2]

int axom::quest::inout_evaluate	(	const double *	x,
		const double *	y,
		const double *	z,
		int	npoints,
		int *	res
	)

Tests an array of points for containment.

Upon successful completion, entries in array res will have the value 1 for points that are inside and value 0 otherwise.

Parameters

[in]	x	Array of x-coordinates for the query points
[in]	y	Array of y-coordinates for the query points
[in]	z	Array of z-coordinates for the query points
[in]	npoints	The number of points to test
[out]	res	An array of results. Each entry has value 1 if the corresponding point is inside or on the mesh and 0 otherwise.

Returns: Return code is QUEST_INOUT_SUCCESS when all preconditions are satisfied and QUEST_INOUT_FAILED otherwise.

Precondition: inout_initialized() == true; When npoints is greater than zero, arrays x, y, z and res are not nullptr and contain sufficient data/space for npoints points.

◆ inout_mesh_min_bounds()

int axom::quest::inout_mesh_min_bounds ( double * coords )

Returns the lower coordinates of the mesh's bounding box.

Parameters

[in] coords A buffer for the returned coordinates

Precondition: coords != nullptr and has sufficient storage for the coordinates; inout_initialized() == true

Returns: Return code is QUEST_INOUT_SUCCESS if successful and QUEST_INOUT_FAILED otherwise.

◆ inout_mesh_max_bounds()

int axom::quest::inout_mesh_max_bounds ( double * coords )

Returns the upper coordinates of the mesh's bounding box.

Parameters

[in] coords A buffer for the returned coordinates

Precondition: coords != nullptr and has sufficient storage for the coordinates; inout_initialized() == true

Returns: Return code is QUEST_INOUT_SUCCESS if successful and QUEST_INOUT_FAILED otherwise.

◆ inout_mesh_center_of_mass()

int axom::quest::inout_mesh_center_of_mass ( double * coords )

Returns the center of mass of the mesh.

The function computes a discrete center of mass defined by the average of the mesh coordinates rather than a continuous center of mass defined by the mesh faces.

Parameters

[in] coords A buffer for the returned coordinates

Precondition: coords != nullptr and has sufficient storage for the coordinates; inout_initialized() == true

Returns: Return code is QUEST_INOUT_SUCCESS if successful and QUEST_INOUT_FAILED otherwise.

◆ inout_get_dimension()

int axom::quest::inout_get_dimension ( )

Gets the spatial dimension of the query.

Returns: Returns the spatial dimension for the inout query

Note: This determines the number of coordinates for the input mesh; The default dimension is 3

See also: inout_set_dimension

◆ inout_set_dimension()

int axom::quest::inout_set_dimension ( int dim )

Sets the spatial dimension of the query.

Parameters

dim	The dimension for the inout query

Returns: Return code is QUEST_INOUT_SUCCESS if successful, QUEST_INOUT_FAILED otherwise

Precondition: dim == 2 || dim == 3; inout_initialized() == false

◆ inout_set_verbose()

int axom::quest::inout_set_verbose ( bool verbosity )

Enables/disables verbose logging output.

By default, the logging verbosity is set to false.

Parameters

verbosity True for more verbose, false for less verbose

Returns: Return code is QUEST_INOUT_SUCCESS if successful and QUEST_INOUT_FAILED otherwise.

Precondition: inout_initialized() == false

◆ inout_set_vertex_weld_threshold()

int axom::quest::inout_set_vertex_weld_threshold ( double thresh )

Sets the cutoff distance for welding vertices during initialization.

By default, the welding threshold is 1E-9.

The inout query requires the input surface to be watertight so this parameter should be set with care. A welding threshold that is too high could unnecessarily merge vertices and create topological defects, while a value that is too low risks leaving gaps in meshes with tolerances between vertices. The default value tends to work well in practice.

Parameters

thresh Cutoff distance for welding vertices

Returns: Return code is QUEST_INOUT_SUCCESS if successful and QUEST_INOUT_FAILED otherwise.

Precondition: inout_initialized() == false; thresh > 0

◆ inout_set_segments_per_knot_span()

int axom::quest::inout_set_segments_per_knot_span ( int segmentsPerKnotSpan )

Sets the number of samples for each knot span (2D only)

By default, the welding threshold is 25

Span intervals are the segments of each curve. This parameter controls the number of samples to use when linearizing each knot span of the contour splines

Parameters

segmentsPerKnotSpan The number of segments for each knot span

Returns: Return code is QUEST_INOUT_SUCCESS if successful and QUEST_INOUT_FAILED otherwise.

Precondition: inout_initialized() == false; segmentsPerKnotSpan >= 1

◆ signed_distance_init() [1/2]

int axom::quest::signed_distance_init	(	const std::string &	file,
		MPI_Comm	comm = `MPI_COMM_SELF`
	)

Initializes the Signed Distance Query with a surface given in an STL formatted file.

Parameters

[in]	file	the name of the file consisting of the surface mesh.
[in]	comm	the MPI communicator (applicable when MPI is available)

Returns: status zero on success, or a non-zero value if an error occurs

Note: The Signed Distance Query currently only supports reading in meshes defined in STL file format.

Precondition: file.empty() == false; comm != MPI_COMM_NULL (when MPI is available); signed_distance_initialized() == false

Postcondition: signed_distance_initialized() == true

◆ signed_distance_init() [2/2]

int axom::quest::signed_distance_init	(	const mint::Mesh *	m,
		MPI_Comm	comm = `MPI_COMM_SELF`
	)

Initializes the Signed Distance Query with the given surface mesh.

Parameters

[in]	m	pointer to the surface mesh object
[in]	comm	the MPI communicator (applicable whem MPI is available)

Returns: status zero on success, or a non-zero value if an error occurs.

Note: The Signed Distance Query currently only supports 3-D triangular surface meshes.

Precondition: m != nullptr; comm != MPI_COMM_NULL (when MPI is available); signed_distance_initialized() == false

Postcondition: signed_distance_initialized() == true

See also: mint::Mesh

◆ signed_distance_initialized()

bool axom::quest::signed_distance_initialized ( )

Checks if the Signed Distance Query has been initialized.

Returns: status true if initialized, else, false.

◆ signed_distance_set_dimension()

void axom::quest::signed_distance_set_dimension ( int dim )

Sets the dimension for the Signed Distance Query.

Parameters

[in] dim the dimension, e.g., 2 or 3

Warning: The Signed Distance function is currently supported in 3D

Note: Options must be set before initializing the Signed Distance Query.

◆ signed_distance_set_closed_surface()

void axom::quest::signed_distance_set_closed_surface ( bool status )

Indicates whether the input to the signed distance consists of a water-tight surface mesh, or not.

Parameters

[in] status flag indicating whether the input is water-tight

Note: By default the input type is assumed to be a water-tight surface mesh.; Options must be set before initializing the Signed Distance Query.; When the input is not a closed surface mesh, the assumption is that the surface mesh divides the computational mesh domain into two regions. Hence, the surface mesh has to span the entire domain of interest, e.g., the computational mesh at which the signed distance field is evaluated, along some plane.

Warning: The sign of the distance from a given query point is determined by a pseudo-normal which is computed at the closest point on the surface mesh. For a non-watertight mesh, the sign of the distance is not defined everywhere. Specifically, the sign is ambiguous for all points for which a normal projection onto the surface does not exist.

◆ signed_distance_set_compute_signs()

void axom::quest::signed_distance_set_compute_signs ( bool computeSign )

Sets whether the distance query should compute or ignore the sign.

Parameters

[in] computeSign predicate indicating if sign should be computed

Note: Options must be set before initializing the Signed Distance Query.

◆ signed_distance_set_allocator()

void axom::quest::signed_distance_set_allocator ( int allocatorID )

Sets the allocator to use for creating internal signed distance query data structures.

Parameters

[in] allocatorID the allocator ID to use

Note: Allocator should be compatible with the set execution space.

◆ signed_distance_set_verbose()

void axom::quest::signed_distance_set_verbose ( bool status )

Enables/Disables verbose output for the Signed Distance Query.

Parameters

[in] status flag indicating whether to enable/disable verbose output

Note: Options must be set before initializing the Signed Distance Query.; Currently, this is only applicable when the Signed Distance Query initializes the SLIC logging environment.

◆ signed_distance_use_shared_memory()

void axom::quest::signed_distance_use_shared_memory ( bool status )

Enable/Disable the use of on-node shared memory (via Umpire's shared-memory resource) for storing the surface mesh. By default this option is disabled.

Parameters

[in] status flag indicating whether to enable/disable shared memory.

Note: This option requires Axom to be built with MPI and Umpire shared memory support.

◆ signed_distance_set_shared_memory_size()

void axom::quest::signed_distance_set_shared_memory_size ( std::size_t min_segment_size )

Set the minimum size (in bytes) of the backing shared-memory segment used to store the surface mesh when signed_distance_use_shared_memory(true) is enabled.

Parameters

[in] min_segment_size Minimum desired shared-memory segment size in bytes (0 to use defaults). This value is treated as a minimum; the implementation may choose a larger value to accommodate the mesh buffer (plus some overhead).

Note: This must be called before initializing the Signed Distance Query.; The shared-memory segment size cannot be increased after creation. To ensure a larger segment is used, call this before the first shared-memory allocation that creates Axom's shared-memory allocator.; This option is only used when Axom is built with MPI and Umpire shared memory support; otherwise it has no effect.

◆ signed_distance_set_execution_space()

void axom::quest::signed_distance_set_execution_space ( SignedDistExec execSpace )

Set the execution space in which to run signed distance queries. By default this option is set to SIGNED_DIST_EVAL_CPU.

Parameters

[in] exec_space the execution space setting, one of the enum values in SignedDistExec.

Note: This function resets the default allocator based on the requested execution space. If a custom allocator ID is desired, set it in a call to axom::setDefaultAllocator after this call.

◆ signed_distance_evaluate() [1/3]

double axom::quest::signed_distance_evaluate	(	double	x,
		double	y,
		double	z = `0.0`
	)

Evaluates the signed distance function at the given point.

Parameters

[in]	x	the x-coordinate of the point in query
[in]	y	the y-coordinate of the point in query
[in]	z	the z-coordinate of the point in query

Returns: d the signed distance evaluated at the specified point.

◆ signed_distance_evaluate() [2/3]

double axom::quest::signed_distance_evaluate	(	double	x,
		double	y,
		double	z,
		double &	cp_x,
		double &	cp_y,
		double &	cp_z,
		double &	n_x,
		double &	n_y,
		double &	n_z
	)

Evaluates the signed distance function at the given 3D point.

Parameters

[in]	x	the x-coordinate of the point in query
[in]	y	the y-coordinate of the point in query
[in]	z	the z-coordinate of the point in query
[out]	cp_x	the x-coordinate of the computed closest point on surface to query point
[out]	cp_y	the y-coordinate of the computed closest point on surface to query point
[out]	cp_z	the z-coordinate of the computed closest point on surface to query point
[out]	n_x	the x-coordinate of the surface normal at the computed closest point
[out]	n_y	the y-coordinate of the surface normal at the computed closest point
[out]	n_z	the z-coordinate of the surface normal at the computed closest point

Returns: d the signed distance evaluated at the specified point. The closest surface point to the query point and the normal at that point are returned as OUT parameters

◆ signed_distance_evaluate() [3/3]

void axom::quest::signed_distance_evaluate	(	const double *	x,
		const double *	y,
		const double *	z,
		int	npoints,
		double *	phi
	)

Evaluates the signed distance function at the given set of points.

Parameters

[in]	x	array consisting of the x-coordinates for each query point
[in]	y	array consisting of the y-coordinates for each query point
[in]	z	array consisting of the z-coordinates for each query point
[in]	npoints	the number of query point
[out]	phi	output array storing the signed distance of each point

Precondition: x != nullptr; y != nullptr; z != nullptr; phi != nullptr

◆ signed_distance_get_mesh_bounds()

void axom::quest::signed_distance_get_mesh_bounds	(	double *	lo,
		double *	hi
	)

Computes the bounds of the specified input mesh supplied to the Signed Distance Query.

Parameters

[out]	lo	buffer to store the lower bound mesh coordinates.
[out]	hi	buffer to store the upper bound mesh coordinates.

Precondition: lo != nullptr; hi != nullptr; hi & lo must point to a buffer that is at least ndims long.; signed_distance_initialized() == true

◆ signed_distance_finalize()

void axom::quest::signed_distance_finalize ( )

Finalizes the SignedDistance query.

Postcondition: signed_distance_initialized() == false.

Variable Documentation

◆ gwn_input_type_v

template<typename GWNQueryType >

constexpr GWNInputType axom::quest::gwn_input_type_v = gwn_input_traits<GWNQueryType>::value

inlineconstexpr

◆ QUEST_INOUT_SUCCESS

constexpr int axom::quest::QUEST_INOUT_SUCCESS = 0

constexpr

◆ QUEST_INOUT_FAILED

constexpr int axom::quest::QUEST_INOUT_FAILED = -1

constexpr

Namespaces

Classes

Enumerations

Functions

Variables

Discretize primitive shapes to linear shapes

Typedef Documentation

◆ SphereType

◆ OctType

◆ Point2D

Enumeration Type Documentation

◆ SearchStatus

◆ GWNInputType

◆ WatertightStatus

◆ SignedDistExec

Function Documentation

◆ all_nearest_neighbors()

◆ discretize() [1/2]

◆ discretize() [2/2]

◆ mesh_from_discretized_polyline()

◆ fast_approximate_winding_number()

◆ subdivide_curves()

◆ subdivide_patches()

◆ setup_gwn_mesh()

◆ generate_gwn_query_mesh()

◆ compute_field_stats()

◆ compute_integrals()

◆ findTriMeshIntersectionsBVH()

◆ findTriMeshIntersectionsImplicitGrid()

◆ findTriMeshIntersectionsUniformGrid()

◆ findTriMeshIntersections()

◆ isSurfaceMeshWatertight()

◆ weldTriMeshVertices()

◆ inout_init() [1/2]

◆ inout_init() [2/2]

◆ inout_finalize()

◆ inout_initialized()

◆ inout_evaluate() [1/2]

◆ inout_evaluate() [2/2]

◆ inout_mesh_min_bounds()

◆ inout_mesh_max_bounds()

◆ inout_mesh_center_of_mass()

◆ inout_get_dimension()

◆ inout_set_dimension()

◆ inout_set_verbose()

◆ inout_set_vertex_weld_threshold()

◆ inout_set_segments_per_knot_span()

◆ signed_distance_init() [1/2]

◆ signed_distance_init() [2/2]

◆ signed_distance_initialized()

◆ signed_distance_set_dimension()

◆ signed_distance_set_closed_surface()

◆ signed_distance_set_compute_signs()

◆ signed_distance_set_allocator()

◆ signed_distance_set_verbose()

◆ signed_distance_use_shared_memory()

◆ signed_distance_set_shared_memory_size()

◆ signed_distance_set_execution_space()

◆ signed_distance_evaluate() [1/3]

◆ signed_distance_evaluate() [2/3]

◆ signed_distance_evaluate() [3/3]

◆ signed_distance_get_mesh_bounds()

◆ signed_distance_finalize()

Variable Documentation

◆ gwn_input_type_v

◆ QUEST_INOUT_SUCCESS

◆ QUEST_INOUT_FAILED