Auxiliary functions for variant hashing. More...

#include <cstdint>
#include <cstddef>
#include <vector>
#include <array>
#include <unordered_map>
#include "gvarHash.hpp"
#include "similarityMatrix.hpp"

Include dependency graph for vashFunctions.hpp:

Functions
uint16_t	BayesicSpace::countSetBits (uint16_t inVal)
	Count set bits in a 16-bit word.

uint64_t	BayesicSpace::countSetBits (const std::vector< uint8_t > &inVec)
	Count set bits in a vector.

uint64_t	BayesicSpace::countSetBits (const std::vector< uint8_t > &inVec, const LocationWithLength &window)
	Count set bits in a range within a vector.

size_t	BayesicSpace::getAvailableRAM ()
	Get available RAM.

uint32_t	BayesicSpace::murMurHashMixer (const std::array< uint32_t, SIZE_OF_SIZET > &key, const uint32_t &seed)
	MurMurHash mixer module of an index value.

uint32_t	BayesicSpace::murMurHashFinalizer (const uint32_t &inputHash)
	MurMurHash finalizer.

uint32_t	BayesicSpace::murMurHash (const std::array< uint32_t, SIZE_OF_SIZET > &key, const uint32_t &seed)
	MurMurHash of an index value.

uint32_t	BayesicSpace::murMurHash (const std::vector< size_t > &key, const uint32_t &seed)
	MurMurHash of a vector of indexes.

uint32_t	BayesicSpace::murMurHash (const std::vector< uint32_t > &key, const uint32_t &seed)
	MurMurHash of a vector of 32-bit unsigned integers.

uint32_t	BayesicSpace::murMurHash (const std::vector< uint16_t > &key, const LocationWithLength &keyWindow, const uint32_t &seed)
	MurMurHash of a vector of indexes.

void	BayesicSpace::testBedMagicBytes (const std::array< char, N_BED_TEST_BYTES > &bytesToTest)
	Test .bed magic bytes.

std::vector< std::pair< size_t, size_t > >	BayesicSpace::makeThreadRanges (const CountAndSize &threadPoolSizes)
	Build thread ranges.

std::vector< size_t >	BayesicSpace::makeChunkSizes (const size_t &nElements, const size_t &nChunks)
	Build chunk sizes.

std::vector< std::pair< RowColIdx, RowColIdx > >	BayesicSpace::makeChunkRanges (const LocationWithLength &startAndChunkSize, const size_t nChunks)
	Build chunk ranges.

std::pair< HashGroupItPairCount, HashGroupItPairCount >	BayesicSpace::makeGroupRanges (const std::vector< HashGroup > &groupVector, const HashGroupItPairCount &startHGPC, const size_t &chunkSize)
	Delimit a chunk of indexes.

void	BayesicSpace::binarizeBedLocus (const LocationWithLength &bedLocusWindow, const std::vector< char > &bedLocus, const size_t &nIndividuals, const LocationWithLength &binLocusWindow, std::vector< uint8_t > &binLocus)
	Convert a locus from .bed to binary format.

void	BayesicSpace::binarizeMacLocus (const std::vector< int > &macLocus, const LocationWithLength &binLocusWindow, std::vector< uint8_t > &binLocus)
	Convert a locus from a vector of minor allele counts.

std::vector< std::string >	BayesicSpace::getLocusNames (const std::string &bimFileName)
	Extract locus names.

void	BayesicSpace::parseCL (int &argc, char **argv, std::unordered_map< std::string, std::string > &cli)
	Command line parser.

void	BayesicSpace::extractCLinfo (const std::unordered_map< std::string, std::string > &parsedCLI, std::unordered_map< std::string, int > &intVariables, std::unordered_map< std::string, float > &floatVariables, std::unordered_map< std::string, std::string > &stringVariables)
	Extract parameters from parsed command line interface flags.

Variables
constexpr size_t	BayesicSpace::SIZE_OF_SIZET {sizeof(size_t) / sizeof(uint32_t)}
	Number of 32-bit values in `size_t`

constexpr size_t	BayesicSpace::N_BED_TEST_BYTES {3}
	Number of test bytes in a .bed file.

Detailed Description

Auxiliary functions for variant hashing.

Author: Anthony J. Greenberg

Copyright: Copyright (c) 2023 Anthony J. Greenberg

Version: 0.5

Definitions of class-external functions needed by hashing classes.

Function Documentation

◆ binarizeBedLocus()

void BayesicSpace::binarizeBedLocus	(	const LocationWithLength &	bedLocusWindow,
		const std::vector< char > &	bedLocus,
		const size_t &	nIndividuals,
		const LocationWithLength &	binLocusWindow,
		std::vector< uint8_t > &	binLocus )

Convert a locus from .bed to binary format.

Convert the .bed two-bit format to one-bit binary.

00 (homozygous first .bim allele) becomes 0 if it is the major allele, 1 otherwise
11 (homozygous second .bim allele) becomes 1 if it is the minor allele, 0 otherwise
01 (missing genotype) becomes 0
10 (heterozygous) becomes 1 with probability 0.5

If the number of individuals is not divisible by eight, the last binary byte is padded with 0s.

Parameters

[in]	bedLocusWindow	.bed locus window
[in]	bedLocus	vector of .bed format bytes
[in]	nIndividuals	number of individuals
[in]	binLocusWindow	binary locus window
[out]	binLocus	vector of binary format bytes

◆ binarizeMacLocus()

void BayesicSpace::binarizeMacLocus	(	const std::vector< int > &	macLocus,
		const LocationWithLength &	binLocusWindow,
		std::vector< uint8_t > &	binLocus )

Convert a locus from a vector of minor allele counts.

Convert minor allele counts to one-bit binary. Input is a vector of minor allele counts (0, 1, or 2) or -9 for missing data. Heterozygotes are assigned the major or minor allele at random, missing genotypes are assigned the major allele. The counts are checked and re-coded if necessary so that set bits represent the minor allele. This function should run faster if the 0 is the major allele homozygote. While the above values are the norm, any negative number will be interpreted as missing, any odd number as 1, and any (non-0) even number as 2.

If the number of individuals is not divisible by eight, the last binary byte is padded with 0s.

Parameters

[in]	macLocus	vector of minor allele counts at a locus
[in]	binLocusWindow	window into the binary vector giving the locus index and length in bytes
[out]	binLocus	vector of binary format bytes

◆ countSetBits() [1/3]

uint64_t BayesicSpace::countSetBits ( const std::vector< uint8_t > & inVec )

Count set bits in a vector.

Counting the set bits in a vector of bytes using Karnigan's method.

Parameters

[in] inVec input vector

Returns: number of bits set

◆ countSetBits() [2/3]

uint64_t BayesicSpace::countSetBits	(	const std::vector< uint8_t > &	inVec,
		const LocationWithLength &	window )

Count set bits in a range within a vector.

Counting the set bits in a range within a vector of bytes using Karnigan's method.

Parameters

[in]	inVec	input vector
[in]	window	vector window in bytes

Returns: number of bits set

◆ countSetBits() [3/3]

uint16_t BayesicSpace::countSetBits ( uint16_t inVal )

Count set bits in a 16-bit word.

Counting the set bits using Karnigan's method. Passing by value to modify the copy and also because the address is much bigger than 16 bits.

Parameters

[in] inVal input value

Returns: number of bits set

◆ extractCLinfo()

void BayesicSpace::extractCLinfo	(	const std::unordered_map< std::string, std::string > &	parsedCLI,
		std::unordered_map< std::string, int > &	intVariables,
		std::unordered_map< std::string, float > &	floatVariables,
		std::unordered_map< std::string, std::string > &	stringVariables )

Extract parameters from parsed command line interface flags.

Extracts needed variable values, indexed by std::string encoded variable names.

Parameters

[in]	parsedCLI	flag values parsed from the command line
[out]	intVariables	indexed `int` variables for use by `main()`
[out]	floatVariables	indexed `float` variables for use by `main()`
[out]	stringVariables	indexed `std::string` variables for use by `main()`

◆ getAvailableRAM()

size_t BayesicSpace::getAvailableRAM ( )

Get available RAM.

Estimates available RAM. If procfs is mounted, uses information from there. Otherwise, sets available RAM to 2 GiB.

Returns: estimated available RAM in bytes

◆ getLocusNames()

std::vector< std::string > BayesicSpace::getLocusNames ( const std::string & bimFileName )

Extract locus names.

Extract locus names from a .bim file.

Parameters

[in] bimFileName .bim file name

Returns: vector of locus names

◆ makeChunkRanges()

std::vector< std::pair< RowColIdx, RowColIdx > > BayesicSpace::makeChunkRanges	(	const LocationWithLength &	startAndChunkSize,
		const size_t	nChunks )

Build chunk ranges.

Build ranges of row/column index pairs for each chunk for a given length of a vectorized similarity matrix.

Parameters

[in]	startAndChunkSize	full matrix index and chunk size
[in]	nChunks	number of chunks

Returns: vector of row/column pair ranges, an element per chunk

◆ makeChunkSizes()

std::vector< size_t > BayesicSpace::makeChunkSizes	(	const size_t &	nElements,
		const size_t &	nChunks )

Build chunk sizes.

Build a vector of chunk sizes. If the number of elements is not evenly divisible by the number of chunks, the remainder is spread among the first number of elements modulo number of chunks.

Parameters

[in]	nElements	number of elements
[in]	nChunks	number of chunks

Returns: vector of chunk sizes

◆ makeGroupRanges()

std::pair< HashGroupItPairCount, HashGroupItPairCount > BayesicSpace::makeGroupRanges	(	const std::vector< HashGroup > &	groupVector,
		const HashGroupItPairCount &	startHGPC,
		const size_t &	chunkSize )

Delimit a chunk of indexes.

Identifies the start and end hash table buckets (groups of loci) and indexes within them for chunked pair-wise LD estimation.

Parameters

[in]	groupVector	vector of hash table buckets
[in]	startHGPC	start group iterator with index into the group
[in]	chunkSize	size of the chunk to be processed

Returns: pair of hash group (bucket) iterators with indexes into the first and last group

◆ makeThreadRanges()

std::vector< std::pair< size_t, size_t > > BayesicSpace::makeThreadRanges ( const CountAndSize & threadPoolSizes )

Build thread ranges.

Build index ranges to use within each thread.

Parameters

[in] threadPoolSizes number of threads and number of loci per thread

Returns: vector of index ranges

◆ murMurHash() [1/4]

uint32_t BayesicSpace::murMurHash	(	const std::array< uint32_t, SIZE_OF_SIZET > &	key,
		const uint32_t &	seed )

MurMurHash of an index value.

Generates a 32-bit hash of an index value using the MurMurHash3 algorithm.

Parameters

[in]	key	the key to be hashed
[in]	seed	the seed

Returns: the hash value

◆ murMurHash() [2/4]

uint32_t BayesicSpace::murMurHash	(	const std::vector< size_t > &	key,
		const uint32_t &	seed )

MurMurHash of a vector of indexes.

Generates a 32-bit hash of an index value vector using the MurMurHash3 algorithm.

Parameters

[in]	key	the key vector to be hashed
[in]	seed	the seed

Returns: the hash value

◆ murMurHash() [3/4]

uint32_t BayesicSpace::murMurHash	(	const std::vector< uint16_t > &	key,
		const LocationWithLength &	keyWindow,
		const uint32_t &	seed )

MurMurHash of a vector of indexes.

Generates a 32-bit hash of a vector of uint16_t values using the MurMurHash3 algorithm.

Parameters

[in]	key	the vector to be hashed
[in]	keyWindow	the range of elements in the key to hash
[in]	seed	the hash seed

Returns: hash value

◆ murMurHash() [4/4]

uint32_t BayesicSpace::murMurHash	(	const std::vector< uint32_t > &	key,
		const uint32_t &	seed )

MurMurHash of a vector of 32-bit unsigned integers.

Generates a 32-bit hash of a vector of unsigned 32-bit integers using the MurMurHash3 algorithm.

Parameters

[in]	key	the key vector to be hashed
[in]	seed	the seed

Returns: the hash value

◆ murMurHashFinalizer()

uint32_t BayesicSpace::murMurHashFinalizer ( const uint32_t & inputHash )

MurMurHash finalizer.

MurMurHash3 finalizer for a hash value.

Parameters

[in] inputHash input unfinlized hash value

Returns: finalized hash value

◆ murMurHashMixer()

uint32_t BayesicSpace::murMurHashMixer	(	const std::array< uint32_t, SIZE_OF_SIZET > &	key,
		const uint32_t &	seed )

MurMurHash mixer module of an index value.

Generates a 32-bit an unfinalized hash of an index value using the MurMurHash3 algorithm.

Parameters

[in]	key	the key to be hashed
[in]	seed	the seed

Returns: the hash value

◆ parseCL()

void BayesicSpace::parseCL	(	int &	argc,
		char **	argv,
		std::unordered_map< std::string, std::string > &	cli )

Command line parser.

Maps flags to values. Flags assumed to be of the form --flag-name value.

Parameters

[in]	argc	size of the `argv` array
[in]	argv	command line input array
[out]	cli	map of tags to values

◆ testBedMagicBytes()

void BayesicSpace::testBedMagicBytes ( const std::array< char, N_BED_TEST_BYTES > & bytesToTest )

Test .bed magic bytes.

Throws if one of the input bytes does not match the three magic values in plink .bed files.

Parameters

[in] bytesToTest the byte set to test

Functions

Variables

Detailed Description

Function Documentation

◆ binarizeBedLocus()

◆ binarizeMacLocus()

◆ countSetBits() [1/3]

◆ countSetBits() [2/3]

◆ countSetBits() [3/3]

◆ extractCLinfo()

◆ getAvailableRAM()

◆ getLocusNames()

◆ makeChunkRanges()

◆ makeChunkSizes()

◆ makeGroupRanges()

◆ makeThreadRanges()

◆ murMurHash() [1/4]

◆ murMurHash() [2/4]

◆ murMurHash() [3/4]

◆ murMurHash() [4/4]

◆ murMurHashFinalizer()

◆ murMurHashMixer()

◆ parseCL()

◆ testBedMagicBytes()