Data provider which allows stats framework to iterate through a masked lattice. More...

#include <MaskedLatticeStatsDataProvider.h>

Public Member Functions
	MaskedLatticeStatsDataProvider ()
	default constructor. More...

	MaskedLatticeStatsDataProvider (MaskedLattice< T > &lattice, uInt iteratorLimitBytes=4096 *4096)
	`iteratorLimitBytes` is related to the size of the lattice. More...

	~MaskedLatticeStatsDataProvider ()

void	operator++ ()
	increment the data provider to the next dataset, mask, range set, and weights. More...

uInt	estimatedSteps () const
	estimated number of steps to iterate through the the lattice More...

Bool	atEnd () const
	Are there any data sets left to provide? More...

void	finalize ()
	Take any actions necessary to finalize the provider. More...

uInt64	getCount ()
	get the count of elements in the current data set. More...

const T *	getData ()
	get the current data set More...

const Bool *	getMask ()
	Get the associated mask of the current dataset. More...

uInt	getNMaxThreads () const
	returns something reasonable based on the lattice size. More...

Bool	hasMask () const
	Does the current data set have an associated mask? More...

void	reset ()
	reset the provider to point to the first data set it manages. More...

void	setLattice (const MaskedLattice< T > &lattice, uInt iteratorLimitBytes=4096 *4096)
	set the lattice. More...

void	updateMaxPos (const std::pair< Int64, Int64 > &maxpos)
	see base class documentation. More...

void	updateMinPos (const std::pair< Int64, Int64 > &minpos)

Public Member Functions inherited from casacore::LatticeStatsDataProviderBase< T >
virtual	~LatticeStatsDataProviderBase ()

uInt	getMaskStride ()
	Get the stride for the current mask (only called if hasMask() returns True). More...

std::vector< std::pair< typename NumericTraits< T >::PrecisionType, typename NumericTraits< T >::PrecisionType > >	getRanges ()
	Get the associated range(s) of the current dataset. More...

uInt	getStride ()
	Get the stride for the current data set. More...

const T *	getWeights ()
	Returns NULL; lattices do not have associated weights. More...

Bool	hasRanges () const
	Does the current data set have associated range(s)? More...

Bool	hasWeights () const
	returns False; lattices do not have associated weights. More...

Bool	isInclude () const
	If the associated data set has ranges, are these include (return True) or exclude (return False) ranges? More...

void	minMaxPos (IPosition &minpos, IPosition &maxpos) const
	get the positions of the min and max More...

void	setProgressMeter (CountedPtr< LattStatsProgress > pm)

void	setRanges (const std::vector< std::pair< typename NumericTraits< T >::PrecisionType, typename NumericTraits< T >::PrecisionType > > &ranges, Bool isInclude)
	set the data ranges More...

Public Member Functions inherited from casacore::StatsDataProvider< NumericTraits< T >::PrecisionType, const T , const Bool >
virtual	~StatsDataProvider ()

virtual void	updateMaxPos (const LocationType &)
	In general, unless you are writing statistics algorithm code, you shouldn't need to call these methods. More...

virtual void	updateMinPos (const LocationType &)

Private Member Functions
void	_freeStorage ()

Private Attributes
CountedPtr< RO_MaskedLatticeIterator< T > >	_iter

Array< T >	_currentSlice

Array< Bool >	_currentMaskSlice

const T *	_currentPtr

const Bool *	_currentMaskPtr

Bool	_delData

Bool	_delMask

Bool	_atEnd

uInt	_nMaxThreads

Additional Inherited Members
Protected Member Functions inherited from casacore::LatticeStatsDataProviderBase< T >
	LatticeStatsDataProviderBase ()

void	_updateMaxPos (const IPosition &maxPos)

void	_updateMinPos (const IPosition &minPos)

void	_updateProgress ()

Protected Member Functions inherited from casacore::StatsDataProvider< NumericTraits< T >::PrecisionType, const T , const Bool >
	StatsDataProvider ()

Detailed Description

template<class T>
class casacore::MaskedLatticeStatsDataProvider< T >

Data provider which allows stats framework to iterate through a masked lattice.

Definition at line 41 of file MaskedLatticeStatsDataProvider.h.

Constructor & Destructor Documentation

◆ MaskedLatticeStatsDataProvider() [1/2]

template<class T >

casacore::MaskedLatticeStatsDataProvider< T >::MaskedLatticeStatsDataProvider ( )

default constructor.

Must set lattice after construction but before using the object

◆ MaskedLatticeStatsDataProvider() [2/2]

template<class T >

casacore::MaskedLatticeStatsDataProvider< T >::MaskedLatticeStatsDataProvider	(	MaskedLattice< T > &	lattice,
		uInt	iteratorLimitBytes = `4096 *4096`
	)

iteratorLimitBytes is related to the size of the lattice.

If the lattice is greater than this size, then a lattice iterator will be used to step through the lattice. If less, then all the data in the values in the lattice are retrieved in a single chunk. The advantage of the iterator is that less memory is used. The disadvantage is there is a significant performace cost, so if the lattice is small, it is better to get all its values in a single chunk and forgo the iterator. This is particularly true when looping for a large number of iterations and creating a MaskedLatticeStatsDataProvider each loop (in that case, you probably will want to create a single object before the loop and use setLattice() to update its lattice).

◆ ~MaskedLatticeStatsDataProvider()

template<class T >

casacore::MaskedLatticeStatsDataProvider< T >::~MaskedLatticeStatsDataProvider ( )

Member Function Documentation

◆ _freeStorage()

template<class T >

void casacore::MaskedLatticeStatsDataProvider< T >::_freeStorage ( )

private

◆ atEnd()

template<class T >

Bool casacore::MaskedLatticeStatsDataProvider< T >::atEnd ( ) const

virtual

Are there any data sets left to provide?

Implements casacore::StatsDataProvider< NumericTraits< T >::PrecisionType, const T *, const Bool * >.

◆ estimatedSteps()

template<class T >

uInt casacore::MaskedLatticeStatsDataProvider< T >::estimatedSteps ( ) const

virtual

estimated number of steps to iterate through the the lattice

Implements casacore::LatticeStatsDataProviderBase< T >.

◆ finalize()

template<class T >

void casacore::MaskedLatticeStatsDataProvider< T >::finalize ( )

virtual

Take any actions necessary to finalize the provider.

This will be called when atEnd() returns True.

Reimplemented from casacore::LatticeStatsDataProviderBase< T >.

◆ getCount()

template<class T >

uInt64 casacore::MaskedLatticeStatsDataProvider< T >::getCount ( )

virtual

get the count of elements in the current data set.

When implementing this method, be certain to take stride into account; ie for a data set with nominally 100 elements that is to have a stride of two, this method should return 50.

Implements casacore::StatsDataProvider< NumericTraits< T >::PrecisionType, const T *, const Bool * >.

◆ getData()

template<class T >

const T* casacore::MaskedLatticeStatsDataProvider< T >::getData ( )

virtual

get the current data set

Implements casacore::StatsDataProvider< NumericTraits< T >::PrecisionType, const T *, const Bool * >.

◆ getMask()

template<class T >

const Bool* casacore::MaskedLatticeStatsDataProvider< T >::getMask ( )

virtual

Get the associated mask of the current dataset.

Only called if hasMask() returns True;

Implements casacore::StatsDataProvider< NumericTraits< T >::PrecisionType, const T *, const Bool * >.

◆ getNMaxThreads()

template<class T >

uInt casacore::MaskedLatticeStatsDataProvider< T >::getNMaxThreads ( ) const

virtual

returns something reasonable based on the lattice size.

Reimplemented from casacore::StatsDataProvider< NumericTraits< T >::PrecisionType, const T *, const Bool * >.

◆ hasMask()

template<class T >

Bool casacore::MaskedLatticeStatsDataProvider< T >::hasMask ( ) const

virtual

Does the current data set have an associated mask?

Implements casacore::StatsDataProvider< NumericTraits< T >::PrecisionType, const T *, const Bool * >.

◆ operator++()

template<class T >

void casacore::MaskedLatticeStatsDataProvider< T >::operator++ ( )

virtual

increment the data provider to the next dataset, mask, range set, and weights.

Implements casacore::StatsDataProvider< NumericTraits< T >::PrecisionType, const T *, const Bool * >.

◆ reset()

template<class T >

void casacore::MaskedLatticeStatsDataProvider< T >::reset ( )

virtual

reset the provider to point to the first data set it manages.

Reimplemented from casacore::LatticeStatsDataProviderBase< T >.

◆ setLattice()

template<class T >

void casacore::MaskedLatticeStatsDataProvider< T >::setLattice	(	const MaskedLattice< T > &	lattice,
		uInt	iteratorLimitBytes = `4096 *4096`
	)

set the lattice.

Automatically resets the lattice iterator. iteratorLimitBytes is related to the size of the lattice. If the lattice is greater than this size, then a lattice iterator will be used to step through the lattice. If less, then all the data in the values in the lattice are retrieved in a single chunk. The advantage of the iterator is that less memory is used. The disadvantage is there is a significant performace cost, so if the lattice is small, it is better to get all its values in a single chunk and forgo the iterator. This is particularly true when looping for a large number of iterations and creating a MaskedLatticeStatsDataProvider each loop (in that case, you probably will want to create a single object before the loop and use setLattice() to update its lattice).