knncolle_hnsw::DistanceOptions< Dim_, InternalData_ > Struct Template Reference

Distance options for the HNSW index. More...

#include <distances.hpp>

Public Attributes
std::function< hnswlib::SpaceInterface< InternalData_ > *(Dim_)>	create
std::function< InternalData_(InternalData_)>	normalize

Detailed Description

template<typename Dim_, typename InternalData_>
struct knncolle_hnsw::DistanceOptions< Dim_, InternalData_ >

Distance options for the HNSW index.

Template Parameters

Dim_	Integer type for the number of dimensions.
InternalData_	Floating point type for the HNSW index.

Member Data Documentation

create

template<typename Dim_ , typename InternalData_ >

std::function<hnswlib::SpaceInterface<InternalData_>*(Dim_)> knncolle_hnsw::DistanceOptions< Dim_, InternalData_ >::create

Create a hnswlib::SpaceInterface object, given the number of dimensions. If not provided, this defaults to hnswlib::L2Space if InternalData_ = float, otherwise it defaults to SquaredEuclideanDistance.

normalize

template<typename Dim_ , typename InternalData_ >

std::function<InternalData_(InternalData_)> knncolle_hnsw::DistanceOptions< Dim_, InternalData_ >::normalize

Normalization function to convert distance measures from hnswlib::SpaceInterface::get_dist_func() into actual distances. If not provided and create is also provided, this defaults to a no-op. If not provided and create is not provided, this defaults to the square root function (i.e., to convert the L2 norm to a Euclidean distance).

---

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

• knncolle_hnsw/distances.hpp