Utility to compress test sets of inverted lists. More...

#include <iostream>
#include <chrono>
#include <thread>
#include <time.h>
#include <stdio.h>
#include <cstdlib>
#include <vector>
#include <map>
#include "bm.h"
#include "bmalgo.h"
#include "bmserial.h"
#include "bmsparsevec.h"
#include "bmsparsevec_compr.h"
#include "bmsparsevec_algo.h"
#include "bmsparsevec_serial.h"
#include "bmalgo_similarity.h"
#include "bmdbg.h"
#include "bmtimer.h"
#include "bmundef.h"

Include dependency graph for inv_list.cpp:

Go to the source code of this file.

Typedefs
typedef bm::sparse_vector< unsigned, bm::bvector<> >	sparse_vector_u32
typedef bm::rsc_sparse_vector< unsigned, sparse_vector_u32 >	rsc_sparse_vector_u32

Functions
static void	show_help ()
static int	parse_args (int argc, char *argv[])
template<class VT>
int	io_read_u32_coll (std::ifstream &fin, VT &vec)
	Read 32-bit vector size-prefix format (length:0, 1, 2, 3, ....).
template<typename VT>
int	validate_inp_vec (const VT &vec, typename VT::value_type &min_delta, typename VT::value_type &min_delta_cnt)
	Check if input vector is monotonously sorted (true inverted list) along the way in computes a minimal delta between values.
template<typename VT, typename BV>
int	compare_vect (const VT &vec, const BV &bv)
	Verification check if integer vector is equivalent to a bit-vector.
template<typename BV>
bool	is_super_sparse (const BV &bv)
	Debug utility to detect super sparse bit-vectors which probably get bad compression rate.
template<typename VT>
bool	write_as_bvector (std::ofstream &bv_file, const VT &vec, bm::serializer< bm::bvector<> > &bvs, bm::serializer< bm::bvector<> >::buffer &sbuf)
	convert vector into bit-vector and append to the file
template<typename VT>
void	write_as_svector (std::ofstream &sv_file, const VT &vec, unsigned min_delta, bm::sparse_vector_serial_layout< sparse_vector_u32 > &sv_lay)
	convert vector into delta coded bit-transposed vector and append to the file
template<typename VT>
void	write_as_rsc_svector (std::ofstream &sv_file, const VT &vec, unsigned min_delta, bm::sparse_vector_serial_layout< rsc_sparse_vector_u32 > &sv_lay)
	convert vector into delta coded bit-transposed vector and append to the file
static void	compress_inv_dump_file (const std::string &fname, const std::string &bv_out_fname, const std::string &sv_out_fname)
	read the input collection sequence, write using various compression schemes
static int	read_bvector (std::ifstream &bv_file, bm::bvector<> &bv, bm::serializer< bm::bvector<> >::buffer &sbuf)
	read and desrialize bit-bector from the dump file
static void	verify_inv_dump_file (const std::string &fname, const std::string &bv_in_fname)
	read the input collection sequence and dump file, verify correctness
static void	decode_test_dump_file (const std::string &bv_in_fname)
	read and decode the compressed dump file
int	main (int argc, char *argv[])

Variables
std::string	bv_in_file
std::string	bv_out_file
std::string	sv_in_file
std::string	sv_out_file
std::string	u32_in_file
std::string	u32_out_file
bool	is_diag = false
bool	is_timing = false
bool	is_verify = false
bool	is_silent = false
bool	is_decode = false
unsigned	c_level = bm::set_compression_default
bm::chrono_taker ::duration_map_type	timing_map

Detailed Description

Utility to compress test sets of inverted lists.

Definition in file inv_list.cpp.

Typedef Documentation

◆ rsc_sparse_vector_u32

typedef bm::rsc_sparse_vector<unsigned, sparse_vector_u32> rsc_sparse_vector_u32

Definition at line 239 of file inv_list.cpp.

◆ sparse_vector_u32

typedef bm::sparse_vector<unsigned, bm::bvector<> > sparse_vector_u32

Definition at line 238 of file inv_list.cpp.

Function Documentation

◆ compare_vect()

template<typename VT, typename BV>

int compare_vect	(	const VT &	vec,
		const BV &	bv )

Verification check if integer vector is equivalent to a bit-vector.

Examples: inv_list.cpp.

Definition at line 305 of file inv_list.cpp.

Referenced by verify_inv_dump_file().

◆ compress_inv_dump_file()

void compress_inv_dump_file	(	const std::string &	fname,
		const std::string &	bv_out_fname,
		const std::string &	sv_out_fname )

static

read the input collection sequence, write using various compression schemes

Examples: inv_list.cpp.

Definition at line 460 of file inv_list.cpp.

References bm::serializer< BV >::byte_order_serialization(), c_level, bm::serializer< BV >::gap_length_serialization(), io_read_u32_coll(), is_silent, bm::serializer< BV >::set_compression_level(), bm::sparse_vector_serial_layout< SV >::size(), timing_map, validate_inp_vec(), write_as_bvector(), and write_as_rsc_svector().

Referenced by main().

◆ decode_test_dump_file()

void decode_test_dump_file ( const std::string & bv_in_fname )

static

read and decode the compressed dump file

Examples: inv_list.cpp.

Definition at line 763 of file inv_list.cpp.

References is_silent, read_bvector(), and timing_map.

Referenced by main().

◆ io_read_u32_coll()

template<class VT>

int io_read_u32_coll	(	std::ifstream &	fin,
		VT &	vec )

Read 32-bit vector size-prefix format (length:0, 1, 2, 3, ....).

Examples: inv_list.cpp.

Definition at line 248 of file inv_list.cpp.

Referenced by compress_inv_dump_file(), and verify_inv_dump_file().

◆ is_super_sparse()

template<typename BV>

bool is_super_sparse ( const BV & bv )

Debug utility to detect super sparse bit-vectors which probably get bad compression rate.

Examples: inv_list.cpp.

Definition at line 325 of file inv_list.cpp.

◆ main()

int main	(	int	argc,
		char *	argv[] )

Definition at line 817 of file inv_list.cpp.

References bv_in_file, bv_out_file, compress_inv_dump_file(), decode_test_dump_file(), is_decode, is_timing, is_verify, parse_args(), bm::chrono_taker< TOut >::print_duration_map(), show_help(), sv_out_file, timing_map, u32_in_file, and verify_inv_dump_file().

◆ parse_args()

int parse_args	(	int	argc,
		char *	argv[] )

static

Definition at line 103 of file inv_list.cpp.

References bv_in_file, bv_out_file, c_level, is_decode, is_diag, is_silent, is_timing, is_verify, show_help(), sv_in_file, sv_out_file, and u32_in_file.

Referenced by main().

◆ read_bvector()

int read_bvector	(	std::ifstream &	bv_file,
		bm::bvector<> &	bv,
		bm::serializer< bm::bvector<> >::buffer &	sbuf )

static

read and desrialize bit-bector from the dump file

Examples: inv_list.cpp.

Definition at line 650 of file inv_list.cpp.

References bm::deserialize().

Referenced by decode_test_dump_file(), and verify_inv_dump_file().

◆ show_help()

void show_help ( )

static

Definition at line 62 of file inv_list.cpp.

Referenced by main(), and parse_args().

◆ validate_inp_vec()

template<typename VT>

int validate_inp_vec	(	const VT &	vec,
		typename VT::value_type &	min_delta,
		typename VT::value_type &	min_delta_cnt )

Check if input vector is monotonously sorted (true inverted list) along the way in computes a minimal delta between values.

Examples: inv_list.cpp.

Definition at line 273 of file inv_list.cpp.

Referenced by compress_inv_dump_file().

◆ verify_inv_dump_file()

void verify_inv_dump_file	(	const std::string &	fname,
		const std::string &	bv_in_fname )

static

read the input collection sequence and dump file, verify correctness

Examples: inv_list.cpp.

Definition at line 677 of file inv_list.cpp.

References compare_vect(), io_read_u32_coll(), is_silent, read_bvector(), and timing_map.

Referenced by main().

◆ write_as_bvector()

template<typename VT>

bool write_as_bvector	(	std::ofstream &	bv_file,
		const VT &	vec,
		bm::serializer< bm::bvector<> > &	bvs,
		bm::serializer< bm::bvector<> >::buffer &	sbuf )

convert vector into bit-vector and append to the file

Returns: true if vector was detected as very low cardinality

Examples: inv_list.cpp.

Definition at line 343 of file inv_list.cpp.

References BM_DECLARE_TEMP_BLOCK, bm::BM_SORTED, bm::bvector< Alloc >::optimize(), and bm::bvector< Alloc >::set().

Referenced by compress_inv_dump_file().

◆ write_as_rsc_svector()

template<typename VT>

void write_as_rsc_svector	(	std::ofstream &	sv_file,
		const VT &	vec,
		unsigned	min_delta,
		bm::sparse_vector_serial_layout< rsc_sparse_vector_u32 > &	sv_lay )