knncolle::KmknnPrebuilt< Distance_, Dim_, Index_, Store_, Float_ > Class Template Reference

Index for a KMKNN search. More...

#include <Kmknn.hpp>

Inheritance diagram for knncolle::KmknnPrebuilt< Distance_, Dim_, Index_, Store_, Float_ >:

Collaboration diagram for knncolle::KmknnPrebuilt< Distance_, Dim_, Index_, Store_, Float_ >:

Public Member Functions
Index_	num_observations () const
Dim_	num_dimensions () const
	KmknnPrebuilt (Dim_ num_dim, Index_ num_obs, std::vector< Store_ > data, const KmknnOptions< Dim_, Index_, Store_ > &options)
std::unique_ptr< Searcher< Index_, Float_ > >	initialize () const

Friends
class	KmknnSearcher< Distance_, Dim_, Index_, Store_, Float_ >

Detailed Description

template<class Distance_, typename Dim_, typename Index_, typename Store_, typename Float_>
class knncolle::KmknnPrebuilt< Distance_, Dim_, Index_, Store_, Float_ >

Index for a KMKNN search.

Instances of this class are usually constructed using KmknnBuilder::build_raw().

Template Parameters

Distance_	A distance calculation class satisfying the MockDistance contract.
Dim_	Integer type for the number of dimensions. For the output of KmknnBuilder::build_raw(), this is set to Matrix_::dimension_type.
Index_	Integer type for the indices. For the output of KmknnBuilder::build_raw(), this is set to Matrix_::index_type.
Store_	Floating point type for the stored data. For the output of KmknnBuilder::build_raw(), this is set to Matrix_::data_type. This may be set to a lower-precision type than Float_ to save memory.
Float_	Floating point type for the query data and distances.

Constructor & Destructor Documentation

KmknnPrebuilt()

template<class Distance_ , typename Dim_ , typename Index_ , typename Store_ , typename Float_ >

knncolle::KmknnPrebuilt< Distance_, Dim_, Index_, Store_, Float_ >::KmknnPrebuilt ( Dim_  num_dim, Index_  num_obs, std::vector< Store_ >  data, const KmknnOptions< Dim_, Index_, Store_ > &  options  )

inline

Parameters

num_dim	Number of dimensions.
num_obs	Number of observations.
data	Vector of length equal to num_dim * num_obs, containing a column-major matrix where rows are dimensions and columns are observations.
options	Options for constructing the k-means index.

Member Function Documentation

initialize()

template<class Distance_ , typename Dim_ , typename Index_ , typename Store_ , typename Float_ >

std::unique_ptr< Searcher< Index_, Float_ > > knncolle::KmknnPrebuilt< Distance_, Dim_, Index_, Store_, Float_ >::initialize ( ) const

inlinevirtual

Creates a KmknnSearcher instance.

Implements knncolle::Prebuilt< Dim_, Index_, Float_ >.

num_dimensions()

template<class Distance_ , typename Dim_ , typename Index_ , typename Store_ , typename Float_ >

Dim_ knncolle::KmknnPrebuilt< Distance_, Dim_, Index_, Store_, Float_ >::num_dimensions ( ) const

inlinevirtual

Returns

Number of dimensions.

Implements knncolle::Prebuilt< Dim_, Index_, Float_ >.

num_observations()

template<class Distance_ , typename Dim_ , typename Index_ , typename Store_ , typename Float_ >

Index_ knncolle::KmknnPrebuilt< Distance_, Dim_, Index_, Store_, Float_ >::num_observations ( ) const

inlinevirtual

Returns

Number of observations in the dataset to be searched.

Implements knncolle::Prebuilt< Dim_, Index_, Float_ >.

---

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

• knncolle/Kmknn.hpp