clue::Clusterer< Ndim > Class Template Reference

The Clusterer class is the interface for running the clustering algorithm. It provides methods to set up the clustering parameters, initializes the internal buffers and runs the clustering algorithm on host or device points. More...

#include <Clusterer.hpp>

Public Member Functions
	Clusterer (float dc, float rhoc, float dm, float seed_dc=-1.f, int pPBin=128)
	Constuct a Clusterer object.
	Clusterer (Queue &queue, float dc, float rhoc, float dm, float seed_dc=-1.f, int pPBin=128)
	Constuct a Clusterer object.
	Clusterer (Queue &queue, TilesDevice *tile_buffer, float dc, float rhoc, float dm, float seed_dc=-1.f, int pPBin=128)
	Constuct a Clusterer object.
template<typename KernelType>
void	make_clusters (PointsHost &h_points, const KernelType &kernel, Queue &queue, std::size_t block_size)
	Construct the clusters from host points.
template<typename KernelType>
void	make_clusters (PointsHost &h_points, const KernelType &kernel, std::size_t block_size)
	Construct the clusters from host points.
template<typename KernelType>
void	make_clusters (PointsHost &h_points, PointsDevice &dev_points, const KernelType &kernel, Queue &queue, std::size_t block_size)
	Construct the clusters from host and device points.
template<typename KernelType>
void	make_clusters (PointsHost &h_points, PointsDevice &dev_points, const KernelType &kernel, std::size_t block_size)
	Construct the clusters from host and device points.
template<typename KernelType>
void	make_clusters (PointsDevice &dev_points, const KernelType &kernel, Queue &queue, std::size_t block_size)
	Construct the clusters from device points.
void	setWrappedCoordinates (const std::array< uint8_t, Ndim > &wrappedCoordinates)
	Specify which coordinates are periodic.
void	setWrappedCoordinates (std::array< uint8_t, Ndim > &&wrappedCoordinates)
	Specify which coordinates are periodic.
template<typename... TArgs>
void	setWrappedCoordinates (TArgs... wrappedCoordinates)
	Specify which coordinates are periodic.
std::vector< std::vector< int > >	getClusters (const PointsHost &h_points)
	Get the clusters from the host points.

Detailed Description

template<uint8_t Ndim>
class clue::Clusterer< Ndim >

The Clusterer class is the interface for running the clustering algorithm. It provides methods to set up the clustering parameters, initializes the internal buffers and runs the clustering algorithm on host or device points.

Template Parameters

Ndim	The number of dimensions of the points to cluster

Constructor & Destructor Documentation

Clusterer() [1/3]

template<uint8_t Ndim>

clue::Clusterer< Ndim >::Clusterer	(	float	dc,
		float	rhoc,
		float	dm,
		float	seed_dc = -1.f,
		int	pPBin = 128 )

Constuct a Clusterer object.

Parameters

dc	Distance threshold for clustering
rhoc	Density threshold for clustering
dm	Minimum distance between clusters
seed_dc	Distance threshold for seed points, if -1.f, dc is used
pPBin	Number of points per bin, used to determine the tile size

Clusterer() [2/3]

template<uint8_t Ndim>

clue::Clusterer< Ndim >::Clusterer	(	Queue &	queue,
		float	dc,
		float	rhoc,
		float	dm,
		float	seed_dc = -1.f,
		int	pPBin = 128 )

Constuct a Clusterer object.

Parameters

queue	The queue to use for the device operations
dc	Distance threshold for clustering
rhoc	Density threshold for clustering
dm	Minimum distance between clusters
seed_dc	Distance threshold for seed points, if -1.f, dc is used
pPBin	Number of points per bin, used to determine the tile size

Clusterer() [3/3]

template<uint8_t Ndim>

clue::Clusterer< Ndim >::Clusterer	(	Queue &	queue,
		TilesDevice *	tile_buffer,
		float	dc,
		float	rhoc,
		float	dm,
		float	seed_dc = -1.f,
		int	pPBin = 128 )

Constuct a Clusterer object.

Parameters

queue	The queue to use for the device operations
tile_buffer	Buffer to pre-allocated tiles
dc	Distance threshold for clustering
rhoc	Density threshold for clustering
dm	Minimum distance between clusters
seed_dc	Distance threshold for seed points, if -1.f, dc is used
pPBin	Number of points per bin, used to determine the tile size

Member Function Documentation

getClusters()

template<uint8_t Ndim>

std::vector< std::vector< int > > clue::Clusterer< Ndim >::getClusters

(

const PointsHost &

h_points

)

Get the clusters from the host points.

Parameters

h_points

Host points to cluster

Returns

A vector of clusters, where each cluster is a vector of point indices

make_clusters() [1/5]

template<uint8_t Ndim>

template<typename KernelType>

void clue::Clusterer< Ndim >::make_clusters	(	PointsDevice &	dev_points,
		const KernelType &	kernel,
		Queue &	queue,
		std::size_t	block_size )

Construct the clusters from device points.

Parameters

dev_points	Device points to cluster
kernel	The convolutional kernel to use for clustering
queue	The queue to use for the device operations
block_size	The size of the blocks to use for clustering

make_clusters() [2/5]

template<uint8_t Ndim>

template<typename KernelType>

void clue::Clusterer< Ndim >::make_clusters	(	PointsHost &	h_points,
		const KernelType &	kernel,
		Queue &	queue,
		std::size_t	block_size )

Construct the clusters from host points.

Parameters

h_points	Host points to cluster
kernel	The convolutional kernel to use for clustering
queue	The queue to use for the device operations
block_size	The size of the blocks to use for clustering

make_clusters() [3/5]

template<uint8_t Ndim>

template<typename KernelType>

void clue::Clusterer< Ndim >::make_clusters	(	PointsHost &	h_points,
		const KernelType &	kernel,
		std::size_t	block_size )

Construct the clusters from host points.

Parameters

h_points	Host points to cluster
kernel	The convolutional kernel to use for clustering
block_size	The size of the blocks to use for clustering

make_clusters() [4/5]

template<uint8_t Ndim>

template<typename KernelType>

void clue::Clusterer< Ndim >::make_clusters	(	PointsHost &	h_points,
		PointsDevice &	dev_points,
		const KernelType &	kernel,
		Queue &	queue,
		std::size_t	block_size )

Construct the clusters from host and device points.

Parameters

h_points	Host points to cluster
dev_points	Device points to cluster
kernel	The convolutional kernel to use for clustering
queue	The queue to use for the device operations
block_size	The size of the blocks to use for clustering

make_clusters() [5/5]

template<uint8_t Ndim>

template<typename KernelType>

void clue::Clusterer< Ndim >::make_clusters	(	PointsHost &	h_points,
		PointsDevice &	dev_points,
		const KernelType &	kernel,
		std::size_t	block_size )

Construct the clusters from host and device points.

Parameters

h_points	Host points to cluster
dev_points	Device points to cluster
kernel	The convolutional kernel to use for clustering
block_size	The size of the blocks to use for clustering

setWrappedCoordinates() [1/3]

template<uint8_t Ndim>

void clue::Clusterer< Ndim >::setWrappedCoordinates

(

const std::array< uint8_t, Ndim > &

wrappedCoordinates

)

Specify which coordinates are periodic.

Parameters

wrappedCoordinates

Array of wrapped coordinates, where 1 means periodic and 0 means non-periodic

setWrappedCoordinates() [2/3]

template<uint8_t Ndim>

void clue::Clusterer< Ndim >::setWrappedCoordinates

(

std::array< uint8_t, Ndim > &&

wrappedCoordinates

)

Specify which coordinates are periodic.

Parameters

wrappedCoordinates

Array of wrapped coordinates, where 1 means periodic and 0 means non-periodic

setWrappedCoordinates() [3/3]

template<uint8_t Ndim>

template<typename... TArgs>

void clue::Clusterer< Ndim >::setWrappedCoordinates

(

TArgs...

wrappedCoordinates

)

Specify which coordinates are periodic.

Template Parameters

TArgs

Types of the wrapped coordinates, should be convertible to uint8_t

Parameters

wrappedCoordinates

Wrapped coordinates, where 1 means periodic and 0 means non-periodic

---

The documentation for this class was generated from the following file:

• include/CLUEstering/core/Clusterer.hpp