mumosa::BlockedIndicesFactory< Index_, Block_ > Class Template Reference

Factory for creating nearest-neighbor search indices for each block. More...

#include <blocked.hpp>

Classes
struct	Buffers
	Temporary buffers for build(). More...

Public Member Functions
	BlockedIndicesFactory (const Index_ num_cells, const Block_ *block)
const std::vector< Index_ > &	sizes () const
template<typename Input_ >
Buffers< Input_ >	create_buffers () const
template<typename Input_ , typename Distance_ , class Matrix_ = knncolle::Matrix<Index_, Input_>>
void	build (const std::size_t num_dim, const Input_ *const data, const knncolle::Builder< Index_, Input_, Distance_, Matrix_ > &builder, std::vector< std::shared_ptr< const knncolle::Prebuilt< Index_, Input_, Distance_ > > > &output, Buffers< Input_ > &work) const
template<typename Input_ , typename Distance_ , class Matrix_ = knncolle::Matrix<Index_, Input_>>
std::vector< std::shared_ptr< const knncolle::Prebuilt< Index_, Input_, Distance_ > > >	build (const std::size_t num_dim, const Input_ *const data, const knncolle::Builder< Index_, Input_, Distance_, Matrix_ > &builder) const

Detailed Description

template<typename Index_, typename Block_>
class mumosa::BlockedIndicesFactory< Index_, Block_ >

Factory for creating nearest-neighbor search indices for each block.

Unlike build_blocked_indices(), this class handles the scenario where cells from the same block do not occupy contiguous columns, i.e., cells from different blocks are intermingled.

Template Parameters

Index_	Integer type of the number of cells.
Block_	Integer type of the block assignments.

Constructor & Destructor Documentation

BlockedIndicesFactory()

template<typename Index_ , typename Block_ >

mumosa::BlockedIndicesFactory< Index_, Block_ >::BlockedIndicesFactory ( const Index_ num_cells, const Block_ * block )

inline

Parameters

	num_cells	Number of cells.
[in]	block	Pointer to an array of length equal to num_cells, containing the block assignment for each cell. Each value should be a non-negative integer in \([0, B)\) where \(B\) is the number of blocks. The lifetime of the underlying array should be no less than the last call to build().

Member Function Documentation

sizes()

template<typename Index_ , typename Block_ >

const std::vector< Index_ > & mumosa::BlockedIndicesFactory< Index_, Block_ >::sizes ( ) const

inline

Returns

Vector of length equal to the number of blocks, containing the number of cells in each block. This can be used in create_workspace().

create_buffers()

template<typename Index_ , typename Block_ >

template<typename Input_ >

Buffers< Input_ > mumosa::BlockedIndicesFactory< Index_, Block_ >::create_buffers ( ) const

inline

Returns

A collection of buffers that can be re-used for multiple calls to build().

build() [1/2]

template<typename Index_ , typename Block_ >

template<typename Input_ , typename Distance_ , class Matrix_ = knncolle::Matrix<Index_, Input_>>

void mumosa::BlockedIndicesFactory< Index_, Block_ >::build ( const std::size_t num_dim, const Input_ *const data, const knncolle::Builder< Index_, Input_, Distance_, Matrix_ > & builder, std::vector< std::shared_ptr< const knncolle::Prebuilt< Index_, Input_, Distance_ > > > & output, Buffers< Input_ > & work ) const

inline

Template Parameters

Input_	Numeric type of the input data.
Distance_	Floating-point type of the distances.
Matrix_	Class of the input data matrix for the neighbor search. This should satisfy the knncolle::Matrix interface.

Parameters

	num_dim	Number of dimensions.
[in]	data	Pointer to an array of length equal to the product of num_dim and num_obs. This contains the embedding matrix for a modality, stored in column-major layout where each row is a dimension and each column is a cell. The block assignment for each cell should be the same as that in block.
	builder	Algorithm to use for the neighbor search.
[out]	output	Vector in which to store the nearest-neighbor search indices constructed by builder. On output, this will have length equal to the number of blocks, where a new search index is constructed for each non-empty block. An empty block will be represented by a null pointer.
	work	Temporary buffers, typically created with create_buffers().

build() [2/2]

template<typename Index_ , typename Block_ >

template<typename Input_ , typename Distance_ , class Matrix_ = knncolle::Matrix<Index_, Input_>>

std::vector< std::shared_ptr< const knncolle::Prebuilt< Index_, Input_, Distance_ > > > mumosa::BlockedIndicesFactory< Index_, Block_ >::build ( const std::size_t num_dim, const Input_ *const data, const knncolle::Builder< Index_, Input_, Distance_, Matrix_ > & builder ) const

inline

Overload of build() that handles some of the memory allocation.

Template Parameters

Input_	Numeric type of the input data.
Distance_	Floating-point type of the distances.
Matrix_	Class of the input data matrix for the neighbor search. This should satisfy the knncolle::Matrix interface.

Parameters

	num_dim	Number of dimensions.
[in]	data	Pointer to an array of length equal to the product of num_dim and num_obs. This contains the embedding matrix for a modality, stored in column-major layout where each row is a dimension and each column is a cell.
	builder	Algorithm to use for the neighbor search.

Returns

Vector in which to store the nearest-neighbor search indices constructed by builder. This has length equal to the number of blocks, where a new search index is constructed for each non-empty block. Empty blocks are represented by null pointers.

---

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

• mumosa/blocked.hpp