compressor class More...

#include <number_compressor.hpp>

Collaboration diagram for NumberCompressor:

Public Types
enum	compressor_type_e { CT_NONE = 0, CT_BLOSC, CT_QLIN16, CT_QLOG16 }
	supported compressor types More...

Public Member Functions
	NumberCompressor ()
	Ctor.

void	free_result ()
	Clear self created results with memory deallocation.

void	clear_data ()
	Reset input data (compressed and uncompressed) memory without deallocation!

void	clear_err ()
	Reset recent error message.

int	get_err ()
	Get recent error message id.

	~NumberCompressor ()
	Dtor.

void	setAllocator (void (Allocator)(size_t), void(DeAllocator)(void ))
	Set memory management. More...

bool	setCompressor (const char *strCompressorType, int iCompressionLevel=-1)
	Converts compressor ID string to category enum. More...

const char *	getCompressorName ()
	Get compressor name.

bool	isLossy ()
	Returns true, if current compressor modifies value data.

bool	pack (void *rdata, size_t rdata_size, size_t rdata_element_size, bool isDoubleClass)
	Calls the qualified compressor (deflate) which always allocates sufficient memory (m_cdata) More...

bool	unpack (void cdata, size_t cdata_size, void rdata, size_t rdata_size, size_t rdata_element_size)
	Calls the qualified compressor (inflate) More...

Public Attributes
bool	m_result_is_const
	true, if result is const type

void *	m_result
	compressor output

size_t	m_result_size
	size of compressor output in bytes

void *	m_rdata
	uncompressed data

size_t	m_rdata_size
	size of uncompressed data in bytes

size_t	m_rdata_element_size
	size of one element in bytes

bool	m_rdata_is_double_type
	Flag type is mxDOUBLE_CLASS.

void *	m_cdata
	compressed data

size_t	m_cdata_size
	size of compressed data in bytes

Private Member Functions
bool	bloscCompress ()
	Allocates memory for compressed data and use it to store results (lossless data compression) More...

bool	bloscDecompress ()
	Uncompress compressed data `m_cdata` to `data` m_rdata. More...

bool	linlogQuantizerCompress (bool bDoLog)
	Lossy data compression by linear or logarithmic quantization (16 bit) More...

bool	linlogQuantizerDecompress (bool bDoLog)


	NumberCompressor (const NumberCompressor &)

NumberCompressor &	operator= (const NumberCompressor &)

Private Attributes
Err	m_err
	recent error

const char *	m_strCompressorType
	name of compressor to use

compressor_type_e	m_eCompressorType
	enum type of compressor to use

int	m_iCompressionLevel
	compression level (0 to 9)

void (	m_Allocator )(size_t szBytes)
	memory allocator

void(*	m_DeAllocator )(void *ptr)
	memory deallocator

Detailed Description

compressor class

Definition at line 53 of file number_compressor.hpp.

Member Enumeration Documentation

◆ compressor_type_e

enum NumberCompressor::compressor_type_e

supported compressor types

Enumerator
CT_NONE	no compression
CT_BLOSC	using BLOSC compressor (lossless)
CT_QLIN16	using linear quantization (lossy)
CT_QLOG16	using logarithmic quantization (lossy)

Definition at line 57 of file number_compressor.hpp.

Constructor & Destructor Documentation

◆ NumberCompressor()

NumberCompressor::NumberCompressor ( const NumberCompressor & )

private

inhibit copy constructor and assignment operator

Member Function Documentation

◆ bloscCompress()

bool NumberCompressor::bloscCompress ( )

inlineprivate

Allocates memory for compressed data and use it to store results (lossless data compression)

Returns: true on success

Definition at line 367 of file number_compressor.hpp.

References m_Allocator, and Err::set().

Referenced by pack().

Here is the caller graph for this function:

◆ bloscDecompress()

bool NumberCompressor::bloscDecompress ( )

inlineprivate

Uncompress compressed data m_cdata to data m_rdata.

m_rdata must point to writable storage space and m_rdata_size must specify the legal space.

Returns: true on success

Definition at line 404 of file number_compressor.hpp.

References Err::set().

Referenced by unpack().

Here is the caller graph for this function:

◆ linlogQuantizerCompress()

bool NumberCompressor::linlogQuantizerCompress ( bool bDoLog )

inlineprivate

Lossy data compression by linear or logarithmic quantization (16 bit)

Allocates m_cdata and use it to store compressed data from m_rdata. Only double types accepted! NaN, +Inf and -Inf are allowed.

Parameters

[in] bDoLog Using logarithmic (true) or linear (false) quantization.

Definition at line 439 of file number_compressor.hpp.

References Err::set().

Referenced by pack().

Here is the caller graph for this function:

◆ linlogQuantizerDecompress()

bool NumberCompressor::linlogQuantizerDecompress ( bool bDoLog )

inlineprivate

Parameters

[in] bDoLog Using logarithmic (true) or linear (false) quantization.

Returns: true on success

Uncompress compressed data m_cdata to data m_rdata. m_rdata must point to writable storage space and m_rdata_size must specify the legal space. (lossy data compression)

Definition at line 570 of file number_compressor.hpp.

References Err::set().

Referenced by unpack().

Here is the caller graph for this function:

◆ pack()

bool NumberCompressor::pack	(	void *	rdata,
		size_t	rdata_size,
		size_t	rdata_element_size,
		bool	isDoubleClass
	)

inline

Calls the qualified compressor (deflate) which always allocates sufficient memory (m_cdata)

Parameters

[in]	rdata	pointer to raw data (byte stream)
[in]	rdata_size	length of raw data in bytes
[in]	rdata_element_size	size of one element in bytes
[in]	isDoubleClass	true, if elements represent double types

Returns: true on success

deploy result (compressed data)

Definition at line 264 of file number_compressor.hpp.

References bloscCompress(), clear_data(), clear_err(), CT_BLOSC, CT_QLIN16, CT_QLOG16, free_result(), linlogQuantizerCompress(), m_cdata, and m_cdata_size.

◆ setAllocator()

void NumberCompressor::setAllocator	(	void ()(size_t)	Allocator,
		void()(void )	DeAllocator
	)

inline

Set memory management.

Parameters

[in]	Allocator	memory allocating functor
[in]	DeAllocator	memory deallocating functor

Definition at line 163 of file number_compressor.hpp.

References m_Allocator, and m_DeAllocator.

◆ setCompressor()

bool NumberCompressor::setCompressor	(	const char *	strCompressorType,
		int	iCompressionLevel = `-1`
	)

inline

Converts compressor ID string to category enum.

Parameters

[in]	strCompressorType	Compressor name as string
[in]	iCompressionLevel	Compression level (compressor dependent)

Definition at line 183 of file number_compressor.hpp.

References Err::clear(), COMPRESSOR_DEFAULT_ID, and CT_NONE.

Referenced by NumberCompressor().

Here is the caller graph for this function:

◆ unpack()

bool NumberCompressor::unpack	(	void *	cdata,
		size_t	cdata_size,
		void *	rdata,
		size_t	rdata_size,
		size_t	rdata_element_size
	)

inline

Calls the qualified compressor (inflate)

Parameters

[in]	cdata	pointer to compressed data
[in]	cdata_size	length of compressed data in bytes
[in,out]	rdata	pointer to memory for decompressed data
[out]	rdata_size	available space ar `rdata` in bytes
[in]	rdata_element_size	size of one element in decompressed vector

Returns: true on success

acquire compressed data

dispatch

deplay result (uncompressed/raw data)

Definition at line 316 of file number_compressor.hpp.

References bloscDecompress(), clear_data(), clear_err(), CT_BLOSC, CT_QLIN16, CT_QLOG16, free_result(), linlogQuantizerDecompress(), m_rdata, and m_rdata_size.

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

number_compressor.hpp

Public Types

Public Member Functions

Public Attributes

Private Member Functions

Private Attributes

Detailed Description

Member Enumeration Documentation

◆ compressor_type_e

Constructor & Destructor Documentation

◆ NumberCompressor()

Member Function Documentation

◆ bloscCompress()

◆ bloscDecompress()

◆ linlogQuantizerCompress()

◆ linlogQuantizerDecompress()

◆ pack()

◆ setAllocator()

◆ setCompressor()

◆ unpack()