Blob handling for arrays. More...

Macros
#define	BLOBARRAY_PUTGET_SPEC(TP)
	Specializations for the standard types (including complex and string). More...

Functions
void	dp3::blob::convertArrayHeader (common::DataFormat, char *header, bool useBlobHeader)
	Convert the array header data. More...

template<typename T >
BlobIStream &	dp3::blob::getBlobArray (BlobIStream &bs, T *&arr, std::vector< uint64_t > &shape, bool fortranOrder)

template<typename T >
void	dp3::blob::getBlobArrayData (BlobIStream &bs, T *data, uint64_t nr)
	Helper function to get an array of data. More...

uint64_t	dp3::blob::getBlobArrayShape (BlobIStream &bs, uint64_t *shape, unsigned int ndim, bool swapAxes, unsigned int nalign)

unsigned int	dp3::blob::getBlobArrayStart (BlobIStream &bs, bool &fortranOrder, uint16_t &ndim)

template<typename T >
uint64_t	dp3::blob::getSpaceBlobArray (BlobIStream &bs, bool useBlobHeader, std::vector< uint64_t > &shape, bool fortranOrder)

template<typename T >
void	dp3::blob::putBlobArrayData (BlobOStream &bs, const T *data, uint64_t nr)

uint64_t	dp3::blob::putBlobArrayHeader (BlobOStream &bs, bool useBlobHeader, const std::string &headerName, const uint64_t *shape, uint16_t ndim, bool fortranOrder, unsigned int alignment)

General function to write a data array
Usually it is used by the other functions, but it can be used on its own to write, say, a C-style array. A 1-dim C-style array can be written with putBlobVector.
template<typename T >
BlobOStream &	dp3::blob::putBlobArray (BlobOStream &bs, const T data, const uint64_t shape, uint16_t ndim, bool fortranOrder)

template<typename T >
BlobOStream &	dp3::blob::putBlobVector (BlobOStream &bs, const T *vec, uint64_t size)
	Put a C-style vector of values as an array. More...

Reserve space for an array with the given shape
The axes ordering (Fortran or C-style) has to be given. It returns the offset of the array in the blob. It is useful for allocating a static blob in a dynamic way. It is only possible if the underlying buffer is seekable. It is meant for use with the BlobOBufString buffer. The function getPointer in that class can be used to turn the position into a pointer. The data will be aligned on the given alignment which defaults to sizeof(T) bytes with a maximum of 8.
template<typename T >
uint64_t	dp3::blob::setSpaceBlobArray1 (BlobOStream &bs, bool useBlobHeader, uint64_t size0, unsigned int alignment)
	Reserve space for a 1-dim array of the given size. More...

template<typename T >
uint64_t	dp3::blob::setSpaceBlobArray2 (BlobOStream &bs, bool useBlobHeader, uint64_t size0, uint64_t size1, bool fortranOrder, unsigned int alignment=0)

template<typename T >
uint64_t	dp3::blob::setSpaceBlobArray3 (BlobOStream &bs, bool useBlobHeader, uint64_t size0, uint64_t size1, uint64_t size2, bool fortranOrder, unsigned int alignment=0)

template<typename T >
uint64_t	dp3::blob::setSpaceBlobArray4 (BlobOStream &bs, bool useBlobHeader, uint64_t size0, uint64_t size1, uint64_t size2, uint64_t size3, bool fortranOrder, unsigned int alignment=0)

template<typename T >
uint64_t	dp3::blob::setSpaceBlobArray (BlobOStream &bs, bool useBlobHeader, const std::vector< uint64_t > &shape, bool fortranOrder, unsigned int alignment=0)

template<typename T >
uint64_t	dp3::blob::setSpaceBlobArray (BlobOStream &bs, bool useBlobHeader, const uint64_t *shape, uint16_t ndim, bool fortranOrder, unsigned int alignment=0)

Write a vector of objects
BlobOStream &	dp3::blob::operator<< (BlobOStream &, const std::vector< bool > &)

template<typename T >
BlobOStream &	dp3::blob::operator<< (BlobOStream &bs, const std::vector< T > &vec)
	Put a vector object as an array. More...

Read back a vector of objects
The dimensionality found in the stream has to be 1. The vector is resized as needed.
BlobIStream &	dp3::blob::operator>> (BlobIStream &, std::vector< bool > &)

template<typename T >
BlobIStream &	dp3::blob::operator>> (BlobIStream &, std::vector< T > &)

Detailed Description

Blob handling for arrays.

Define functions to write N-dimensional arrays into a blob and to read them back from a blob. The arrays can be:

A plain N-dimensional C-array.
A blitz array in Fortran order (minor axis first) or in C order.
An AIPS++ array (which is in Fortran order).

Special functions exist to read or write a vector (1-dimensional array). Because all array types are written in a standard way, it is possible to write, for example, a blitz array and read it back as an AIPS++ array. If the axes ordering is different, the axes are reversed.

The write functions follow the same standard as the static array header defined in BlobArrayHeader.h, so it is possible to read a static array back in a dynamic way.

Macro Definition Documentation

◆ BLOBARRAY_PUTGET_SPEC

#define BLOBARRAY_PUTGET_SPEC ( TP )

Value:

  template <>                                                                  \
  inline void putBlobArrayData(BlobOStream& bs, const TP* data, uint64_t nr) { \
    bs.put(data, nr);                                                          \
  }                                                                            \
  template <>                                                                  \
  inline void getBlobArrayData(BlobIStream& bs, TP* data, uint64_t nr) {       \
    bs.get(data, nr);                                                          \
  }

Specializations for the standard types (including complex and string).

Function Documentation

◆ convertArrayHeader()

void dp3::blob::convertArrayHeader	(	common::DataFormat	,
		char *	header,
		bool	useBlobHeader
	)

Convert the array header data.

◆ getBlobArray()

template<typename T >

BlobIStream& dp3::blob::getBlobArray	(	BlobIStream &	bs,
		T *&	arr,
		std::vector< uint64_t > &	shape,
		bool	fortranOrder
	)

Read back as a C-array with the axes in Fortran or C-style order. It allocates the required storage and puts the pointer to it in arr. The shape is stored in the vector. The user is responsible for deleting the data.

◆ getBlobArrayData()

template<typename T >

void dp3::blob::getBlobArrayData	(	BlobIStream &	bs,
		T *	data,
		uint64_t	nr
	)

Helper function to get an array of data.

◆ getBlobArrayShape()

uint64_t dp3::blob::getBlobArrayShape	(	BlobIStream &	bs,
		uint64_t *	shape,
		unsigned int	ndim,
		bool	swapAxes,
		unsigned int	nalign
	)

Get the shape of an array from the blob. This is a helper function for the functions reading an array. It returns the number of elements in the array.

◆ getBlobArrayStart()

unsigned int dp3::blob::getBlobArrayStart	(	BlobIStream &	bs,
		bool &	fortranOrder,
		uint16_t &	ndim
	)

inline

Get the ordering and dimensionality. This is a helper function for the functions reading an array. It returns the number of alignment bytes used.

◆ getSpaceBlobArray()

template<typename T >

uint64_t dp3::blob::getSpaceBlobArray	(	BlobIStream &	bs,
		bool	useBlobHeader,
		std::vector< uint64_t > &	shape,
		bool	fortranOrder
	)

Find an array in the blob with the axes in Fortran or C-style order. The shape is stored in the vector. It returns the offset of the array in the buffer and treats the array as being read (thus skips over the data). It is only possible if the underlying buffer is seekable. It is meant for use with the BlobIBufString buffer. The function getPointer in that class can be used to turn the position into a pointer. The data are assumed to be aligned on the given alignment which defaults to sizeof(T) bytes with a maximum of 8.

◆ operator<<() [1/2]

BlobOStream& dp3::blob::operator<<	(	BlobOStream &	,
		const std::vector< bool > &
	)

◆ operator<<() [2/2]

template<typename T >

BlobOStream & dp3::blob::operator<<	(	BlobOStream &	bs,
		const std::vector< T > &	vec
	)

inline

Put a vector object as an array.

◆ operator>>() [1/2]

BlobIStream& dp3::blob::operator>>	(	BlobIStream &	,
		std::vector< bool > &
	)

◆ operator>>() [2/2]

template<typename T >

BlobIStream& dp3::blob::operator>>	(	BlobIStream &	,
		std::vector< T > &
	)

◆ putBlobArray()

template<typename T >

BlobOStream& dp3::blob::putBlobArray	(	BlobOStream &	bs,
		const T *	data,
		const uint64_t *	shape,
		uint16_t	ndim,
		bool	fortranOrder
	)

◆ putBlobArrayData()

template<typename T >

void dp3::blob::putBlobArrayData	(	BlobOStream &	bs,
		const T *	data,
		uint64_t	nr
	)

Helper function to put an array of data. It is specialized for the standard types (including complex and string).

◆ putBlobArrayHeader()

uint64_t dp3::blob::putBlobArrayHeader	(	BlobOStream &	bs,
		bool	useBlobHeader,
		const std::string &	headerName,
		const uint64_t *	shape,
		uint16_t	ndim,
		bool	fortranOrder,
		unsigned int	alignment
	)

Put a blob array header. It returns the number of elements in the array. This is a helper function for the functions writing an array. After writing the shape it aligns the stream on the given alignment.

◆ putBlobVector()

template<typename T >

BlobOStream & dp3::blob::putBlobVector	(	BlobOStream &	bs,
		const T *	data,
		uint64_t	size
	)

inline

Put a C-style vector of values as an array.

◆ setSpaceBlobArray() [1/2]

template<typename T >

uint64_t dp3::blob::setSpaceBlobArray	(	BlobOStream &	bs,
		bool	useBlobHeader,
		const std::vector< uint64_t > &	shape,
		bool	fortranOrder,
		unsigned int	alignment = `0`
	)

◆ setSpaceBlobArray() [2/2]

template<typename T >

uint64_t dp3::blob::setSpaceBlobArray	(	BlobOStream &	bs,
		bool	useBlobHeader,
		const uint64_t *	shape,
		uint16_t	ndim,
		bool	fortranOrder,
		unsigned int	alignment = `0`
	)

◆ setSpaceBlobArray1()

template<typename T >

uint64_t dp3::blob::setSpaceBlobArray1	(	BlobOStream &	bs,
		bool	useBlobHeader,
		uint64_t	size0,
		unsigned int	alignment = `0`
	)

inline

Reserve space for a 1-dim array of the given size.

◆ setSpaceBlobArray2()

template<typename T >

uint64_t dp3::blob::setSpaceBlobArray2	(	BlobOStream &	bs,
		bool	useBlobHeader,
		uint64_t	size0,
		uint64_t	size1,
		bool	fortranOrder,
		unsigned int	alignment = `0`
	)

◆ setSpaceBlobArray3()

template<typename T >

uint64_t dp3::blob::setSpaceBlobArray3	(	BlobOStream &	bs,
		bool	useBlobHeader,
		uint64_t	size0,
		uint64_t	size1,
		uint64_t	size2,
		bool	fortranOrder,
		unsigned int	alignment = `0`
	)

◆ setSpaceBlobArray4()

template<typename T >

uint64_t dp3::blob::setSpaceBlobArray4	(	BlobOStream &	bs,
		bool	useBlobHeader,
		uint64_t	size0,
		uint64_t	size1,
		uint64_t	size2,
		uint64_t	size3,
		bool	fortranOrder,
		unsigned int	alignment = `0`
	)

Macros

Functions

General function to write a data array

Reserve space for an array with the given shape

Write a vector of objects

Read back a vector of objects

Detailed Description

Macro Definition Documentation

◆ BLOBARRAY_PUTGET_SPEC

Function Documentation

◆ convertArrayHeader()

◆ getBlobArray()

◆ getBlobArrayData()

◆ getBlobArrayShape()

◆ getBlobArrayStart()

◆ getSpaceBlobArray()

◆ operator<<() [1/2]

◆ operator<<() [2/2]

◆ operator>>() [1/2]

◆ operator>>() [2/2]

◆ putBlobArray()

◆ putBlobArrayData()

◆ putBlobArrayHeader()

◆ putBlobVector()

◆ setSpaceBlobArray() [1/2]

◆ setSpaceBlobArray() [2/2]

◆ setSpaceBlobArray1()

◆ setSpaceBlobArray2()

◆ setSpaceBlobArray3()

◆ setSpaceBlobArray4()