data_util.cpp File Reference

#include "data_util.h"

Include dependency graph for data_util.cpp:

Functions
Distance	Eucldistance (const Point &x, const Point &y)
Distance	Eucldistance (const Point &x, const Point &y, const Dimensions start, const Dimensions finish)
void	GetSubpoints (const Points &points, const Dimensions start_dim, const Dimensions final_dim, Points *subpoints)
ClusterId	GetNearestClusterId (const Point &point, const Centroids &centroids, const Dimensions start_dim, const Dimensions final_dim)
void	GetResidual (const Point &point, const CoarseQuantization &coarse_quantizations, const vector< Centroids > &centroids, Point *residual)
void	GetResidual (const Point &point, const CoarseQuantization &coarse_quantizations, const vector< Centroids > &centroids, Coord *residual)
void	GetNearestClusterIdsForPointSubset (const Points &points, const Centroids &centroids, const PointId start_pid, const PointId final_pid, vector< ClusterId > *nearest)
void	GetNearestClusterIdsForSubpoints (const Points &points, const Centroids &centroids, const Dimensions start_dim, const Dimensions final_dim, int threads_count, vector< ClusterId > *nearest)
void	GetPointsCoarseQuaintizations (const Points &points, const vector< Centroids > &centroids, const int threads_count, vector< CoarseQuantization > *coarse_quantizations)

---

Function Documentation

Distance Eucldistance	(	const Point &	x,
		const Point &	y
	)

Function calculates squared euclidian distance between two points (points must have the same dimensionality)

Parameters:

x	first point
y	second point

Distance Eucldistance	(	const Point &	x,
		const Point &	y,
		Dimensions	start,
		Dimensions	finish
	)

Function calculates squared euclidian distance point with small dimensionality and subpoint of point with bigger dimensionality.

Parameters:

x	first point
y	second point
start	first dimension of subpoint
finish	dimension after the last dimension of subpoint

ClusterId GetNearestClusterId	(	const Point &	point,
		const Centroids &	centroids,
		const Dimensions	start_dim,
		const Dimensions	final_dim
	)

This function returns identifier of clusters which centroid is the nearest to subpoint limited by start_dim and final_dim

Parameters:

point	full point
Centroids	all centroids (function finds the nearest one)
start_dim	first dimension of subpoint
final_dim	dimension after the last dimension of subpoint

void GetNearestClusterIdsForPointSubset	(	const Points &	points,
		const Centroids &	centroids,
		const PointId	start_pid,
		const PointId	final_pid,
		vector< ClusterId > *	nearest
	)

This function finds nearest cluster identifiers for points from start_pid to final_pid. We need this function for multi-threading

Parameters:

points	all points
centroids	centroids of clusters
start_pid	first point function finds nearest cluster
final_pid	point after the last point function finds nearest cluster

void GetNearestClusterIdsForSubpoints	(	const Points &	points,
		const Centroids &	centroids,
		const Dimensions	start_dim,
		const Dimensions	final_dim,
		int	threads_count,
		vector< ClusterId > *	nearest
	)

This function finds cluster identifiers nearest to subpoints for a number of points. Subpoints are limited by start_dim and finish_dim

Parameters:

points	all points
centroids	centroids of clusters
start_dim	first dimesion of subpoint
final_dim	dimesion after the last dimension of subpoint
threads_count	number of threads
nearest	result

void GetPointsCoarseQuaintizations	(	const Points &	points,
		const vector< Centroids > &	centroids,
		const int	threads_count,
		vector< CoarseQuantization > *	coarse_quantizations
	)

This function calculates points coarse product quantizations

Parameters:

points	all points
centroids	centroids of clusters
threads_count	number of threads
coarse_quantizations	result quantizations

void GetResidual	(	const Point &	point,
		const CoarseQuantization &	coarse_quantizations,
		const vector< Centroids > &	centroids,
		Point *	residual
	)

This function calculates quantization residual.

Parameters:

point	initial point
coarse_quantizations	point coarse quantization
centroids	lists of centroids
residual	result residual

void GetResidual	(	const Point &	point,
		const CoarseQuantization &	coarse_quantizations,
		const vector< Centroids > &	centroids,
		Coord *	residual
	)

This function calculates quantization residual.

Parameters:

point	initial point
coarse_quantizations	point coarse quantization
centroids	lists of centroids
residual	pointer to start of residual

void GetSubpoints	(	const Points &	points,
		const Dimensions	start_dim,
		const Dimensions	final_dim,
		Points *	subpoints
	)

This function returns subpoints limited by start_dim and final_dim for every point in points

Parameters:

points	all points
start_dim	first dimension of subpoint
final_dim	dimension after the last dimension of subpoint
subpoints	result subpoints