nearest

Defined in header <ArborX_Predicates.hpp>

template<typename Geometry>
KOKKOS_INLINE_FUNCTION auto nearest(Geometry const &geometry, int k = 1)

Generate a nearest predicate to perform a k-nearest neighbors search with in a spatial index’ query().

Template Parameters:

Geometry – The geometry type, e.g., ArborX::Point, ArborX::Box, or ArborX::Sphere. See Geometries for the geometries provided by ArborX.

Parameters:
  • geometry – The geometry object from which distance is calculated.

  • k – The number of values to search for.

Returns:

A predicate satisfied by the k values stored in the index whose indexable is the closest to geometry. If the index stores fewer than k values, all of them are found. The type of the returned object is unspecified.

Example

auto nearest_five_to_origin_pred =
    ArborX::nearest(ArborX::Point{0.f, 0.f, 0.f}, 5);

Helper functions

The helper function creates a set of predicates from a set of geometries rather than from a single one. The geometries are provided through AccessTraits, for instance as a Kokkos::View. The result is meant to be passed directly as the predicates argument of query().

template<typename Primitives>
auto Experimental::make_nearest(Primitives const &primitives, int k)

Create a set of nearest predicates, one per geometry.

This is equivalent to creating the set of predicates nearest(AccessTraits<Primitives>::get(primitives, i), k).

Parameters:
  • primitives – A set of geometries.

  • k – The number of values to search for, for each geometry.

using Geometries = Kokkos::View<Geometry*>;
Geometries geometries("geometries", 5);
int k = 3;
auto predicates = ArborX::Experimental::make_nearest(geometries, k);

See also