Contains general bit-comparison and similarity operations. More...

#include "BitVects.h"
#include <string>

Functions
template<typename T >
double	SimilarityWrapper (const T &bv1, const T &bv2, double(*metric)(const T &, const T &), bool returnDistance=false)
template<typename T >
double	SimilarityWrapper (const T &bv1, const T &bv2, double a, double b, double(*metric)(const T &, const T &, double, double), bool returnDistance=false)
bool	AllProbeBitsMatch (const char probe, const char ref)
bool	AllProbeBitsMatch (const std::string &probe, const std::string &ref)
template<typename T1 >
bool	AllProbeBitsMatch (const T1 &probe, const std::string &pkl)
template<typename T1 >
bool	AllProbeBitsMatch (const T1 &probe, const T1 &ref)
template<typename T1 , typename T2 >
int	NumOnBitsInCommon (const T1 &bv1, const T2 &bv2)
	returns the number of on bits in common between two bit vectors
int	NumOnBitsInCommon (const ExplicitBitVect &bv1, const ExplicitBitVect &bv2)
template<typename T1 , typename T2 >
double	TanimotoSimilarity (const T1 &bv1, const T2 &bv2)
	returns the Tanimoto similarity between two bit vects
template<typename T1 , typename T2 >
double	CosineSimilarity (const T1 &bv1, const T2 &bv2)
	returns the Cosine similarity between two bit vects
template<typename T1 , typename T2 >
double	KulczynskiSimilarity (const T1 &bv1, const T2 &bv2)
	returns the Kulczynski similarity between two bit vects
template<typename T1 , typename T2 >
double	DiceSimilarity (const T1 &bv1, const T2 &bv2)
	returns the Dice similarity between two bit vects
template<typename T1 , typename T2 >
double	TverskySimilarity (const T1 &bv1, const T2 &bv2, double a, double b)
	returns the Tversky similarity between two bit vects
template<typename T1 , typename T2 >
double	SokalSimilarity (const T1 &bv1, const T2 &bv2)
	returns the Sokal similarity between two bit vects
template<typename T1 , typename T2 >
double	McConnaugheySimilarity (const T1 &bv1, const T2 &bv2)
	returns the McConnaughey similarity between two bit vects
template<typename T1 , typename T2 >
double	AsymmetricSimilarity (const T1 &bv1, const T2 &bv2)
	returns the Asymmetric similarity between two bit vects
template<typename T1 , typename T2 >
double	BraunBlanquetSimilarity (const T1 &bv1, const T2 &bv2)
	returns the Braun-Blanquet similarity between two bit vects
template<typename T1 , typename T2 >
double	RusselSimilarity (const T1 &bv1, const T2 &bv2)
	returns the Russel similarity between two bit vects
template<typename T1 , typename T2 >
double	OnBitSimilarity (const T1 &bv1, const T2 &bv2)
	returns the on bit similarity between two bit vects
template<typename T1 , typename T2 >
int	NumBitsInCommon (const T1 &bv1, const T2 &bv2)
	returns the number of common bits (on and off) between two bit vects
template<typename T1 , typename T2 >
double	AllBitSimilarity (const T1 &bv1, const T2 &bv2)
	returns the commong-bit similarity (on and off) between two bit vects
template<typename T1 , typename T2 >
IntVect	OnBitsInCommon (const T1 &bv1, const T2 &bv2)
	returns an IntVect with indices of all on bits in common between two bit vects
template<typename T1 , typename T2 >
IntVect	OffBitsInCommon (const T1 &bv1, const T2 &bv2)
	returns an IntVect with indices of all off bits in common between two bit vects
template<typename T1 , typename T2 >
DoubleVect	OnBitProjSimilarity (const T1 &bv1, const T2 &bv2)
	returns the on-bit projected similarities between two bit vects
template<typename T1 , typename T2 >
DoubleVect	OffBitProjSimilarity (const T1 &bv1, const T2 &bv2)
	returns the on-bit projected similarities between two bit vects
template<typename T1 >
T1 *	FoldFingerprint (const T1 &bv1, unsigned int factor=2)
	folds a bit vector `factor` times and returns the result
template<typename T1 >
std::string	BitVectToText (const T1 &bv1)
	returns a text representation of a bit vector (a string of 0s and 1s)
template<typename T1 >
std::string	BitVectToFPSText (const T1 &bv1)
	returns a hex representation of a bit vector compatible with Andrew Dalke's FPS format
template<typename T1 >
void	UpdateBitVectFromFPSText (T1 &bv1, const std::string &fps)
	updates a bit vector from Andrew Dalke's FPS format

Detailed Description

Contains general bit-comparison and similarity operations.

The notation used to document the similarity metrics is:

V1_n: number of bits in vector 1
V1_o: number of on bits in vector 1
(V1&V2)_o: number of on bits in the intersection of vectors 1 and 2

Definition in file BitOps.h.

Function Documentation

template<typename T1 , typename T2 >

double AllBitSimilarity	(	const T1 &	bv1,
		const T2 &	bv2
	)

returns the commong-bit similarity (on and off) between two bit vects

Returns:: [bv1_n - (bv1^bv2)_o] / bv1_n

bool AllProbeBitsMatch	(	const char *	probe,
		const char *	ref
	)

bool AllProbeBitsMatch	(	const std::string &	probe,
		const std::string &	ref
	)

template<typename T1 >

bool AllProbeBitsMatch	(	const T1 &	probe,
		const std::string &	pkl
	)

template<typename T1 >

bool AllProbeBitsMatch	(	const T1 &	probe,
		const T1 &	ref
	)

template<typename T1 , typename T2 >

double AsymmetricSimilarity	(	const T1 &	bv1,
		const T2 &	bv2
	)

returns the Asymmetric similarity between two bit vects

Returns:: (bv1&bv2)_o / min(bv1_o,bv2_o)

template<typename T1 >

std::string BitVectToFPSText ( const T1 & bv1 )

returns a hex representation of a bit vector compatible with Andrew Dalke's FPS format

Parameters:

bv1

the vector to use

Returns:: an std::string

template<typename T1 >

std::string BitVectToText ( const T1 & bv1 )

returns a text representation of a bit vector (a string of 0s and 1s)

Parameters:

bv1

the vector to use

Returns:: an std::string

template<typename T1 , typename T2 >

double BraunBlanquetSimilarity	(	const T1 &	bv1,
		const T2 &	bv2
	)

returns the Braun-Blanquet similarity between two bit vects

Returns:: (bv1&bv2)_o / max(bv1_o,bv2_o)

template<typename T1 , typename T2 >

double CosineSimilarity	(	const T1 &	bv1,
		const T2 &	bv2
	)

returns the Cosine similarity between two bit vects

Returns:: (bv1&bv2)_o / sqrt(bv1_o + bv2_o)

template<typename T1 , typename T2 >

double DiceSimilarity	(	const T1 &	bv1,
		const T2 &	bv2
	)

returns the Dice similarity between two bit vects

Returns:: 2*(bv1&bv2)_o / [bv1_o + bv2_o]

template<typename T1 >

T1* FoldFingerprint	(	const T1 &	bv1,
		unsigned int	factor = `2`
	)

folds a bit vector factor times and returns the result

Parameters:

	bv1	the vector to be folded
	factor	(optional) the number of times to fold it

Returns:: a pointer to the folded fingerprint, which is bv1_n/factor long.

Note: The caller is responsible for deleteing the result.

Referenced by SimilarityWrapper().

template<typename T1 , typename T2 >

double KulczynskiSimilarity	(	const T1 &	bv1,
		const T2 &	bv2
	)

returns the Kulczynski similarity between two bit vects

Returns:: (bv1&bv2)_o * [bv1_o + bv2_o] / [2 * bv1_o * bv2_o]

template<typename T1 , typename T2 >

double McConnaugheySimilarity	(	const T1 &	bv1,
		const T2 &	bv2
	)

returns the McConnaughey similarity between two bit vects

Returns:: [(bv1&bv2)_o * (bv1_o + bv2_o) - (bv1_o * bv2_o)] / (bv1_o * bv2_o)

template<typename T1 , typename T2 >

int NumBitsInCommon	(	const T1 &	bv1,
		const T2 &	bv2
	)

returns the number of common bits (on and off) between two bit vects

Returns:: bv1_n - (bv1^bv2)_o

int NumOnBitsInCommon	(	const ExplicitBitVect &	bv1,
		const ExplicitBitVect &	bv2
	)

template<typename T1 , typename T2 >

int NumOnBitsInCommon	(	const T1 &	bv1,
		const T2 &	bv2
	)

returns the number of on bits in common between two bit vectors

Returns:: (bv1&bv2)_o

template<typename T1 , typename T2 >

DoubleVect OffBitProjSimilarity	(	const T1 &	bv1,
		const T2 &	bv2
	)

returns the on-bit projected similarities between two bit vects

Returns:

two values, as a DoubleVect:

[bv1_n - (bv1|bv2)_o] / [bv1_n - bv1_o]
[bv2_n - (bv1|bv2)_o] / [bv2_n - bv2_o]

Note: bv1_n = bv2_n

template<typename T1 , typename T2 >

IntVect OffBitsInCommon	(	const T1 &	bv1,
		const T2 &	bv2
	)

returns an IntVect with indices of all off bits in common between two bit vects

template<typename T1 , typename T2 >

DoubleVect OnBitProjSimilarity	(	const T1 &	bv1,
		const T2 &	bv2
	)

returns the on-bit projected similarities between two bit vects

Returns:

two values, as a DoubleVect:

(bv1&bv2)_o / bv1_o
(bv1&bv2)_o / bv2_o

template<typename T1 , typename T2 >

double OnBitSimilarity	(	const T1 &	bv1,
		const T2 &	bv2
	)

returns the on bit similarity between two bit vects

Returns:: (bv1&bv2)_o / (bv1|bv2)_o

template<typename T1 , typename T2 >

IntVect OnBitsInCommon	(	const T1 &	bv1,
		const T2 &	bv2
	)

returns an IntVect with indices of all on bits in common between two bit vects

template<typename T1 , typename T2 >

double RusselSimilarity	(	const T1 &	bv1,
		const T2 &	bv2
	)

returns the Russel similarity between two bit vects

Returns:: (bv1&bv2)_o / bv1_o

Note: that this operation is non-commutative: RusselSimilarity(bv1,bv2) != RusselSimilarity(bv2,bv1)

template<typename T >

double SimilarityWrapper	(	const T &	bv1,
		const T &	bv2,
		double	a,
		double	b,
		double(*)(const T &, const T &, double, double)	metric,
		bool	returnDistance = `false`
	)

This is an overloaded member function, provided for convenience. It differs from the above function only in what argument(s) it accepts.

Definition at line 50 of file BitOps.h.

References FoldFingerprint().

template<typename T >

double SimilarityWrapper	(	const T &	bv1,
		const T &	bv2,
		double(*)(const T &, const T &)	metric,
		bool	returnDistance = `false`
	)

general purpose wrapper for calculating the similarity between two bvs that may be of unequal size (will automatically fold as appropriate)

Definition at line 30 of file BitOps.h.

References FoldFingerprint().

Referenced by RDDataManip::TanimotoDistanceMetric(), and RDDataManip::TanimotoSimilarityMetric().

template<typename T1 , typename T2 >

double SokalSimilarity	(	const T1 &	bv1,
		const T2 &	bv2
	)

returns the Sokal similarity between two bit vects

Returns:: (bv1&bv2)_o / [2*bv1_o + 2*bv2_o - 3*(bv1&bv2)_o]

template<typename T1 , typename T2 >

double TanimotoSimilarity	(	const T1 &	bv1,
		const T2 &	bv2
	)

returns the Tanimoto similarity between two bit vects

Returns:: (bv1&bv2)_o / [bv1_o + bv2_o - (bv1&bv2)_o]

Referenced by RDDataManip::TanimotoDistanceMetric(), and RDDataManip::TanimotoSimilarityMetric().

template<typename T1 , typename T2 >

double TverskySimilarity	(	const T1 &	bv1,
		const T2 &	bv2,
		double	a,
		double	b
	)

returns the Tversky similarity between two bit vects

Returns:: (bv1&bv2)_o / [a*bv1_o + b*bv2_o + (1 - a - b)*(bv1&bv2)_o]

Notes: # 0 <= a,b <= 1 # Tversky(a=1,b=1) = Tanimoto # Tversky(a=1/2,b=1/2) = Dice

template<typename T1 >

void UpdateBitVectFromFPSText	(	T1 &	bv1,
		const std::string &	fps
	)

updates a bit vector from Andrew Dalke's FPS format

Parameters:

	bv1	the vector to use
	fps	the FPS hex string

BitOps.h File Reference

Functions

Detailed Description

Function Documentation