result averaging. More...

#include <operations.h>

Inheritance diagram for cass::pp61:

Collaboration diagram for cass::pp61:

Public Member Functions
	pp61 (const name_t &name)
	constructor More...

virtual void	process (const CASSEvent &, result_t &)
	process event More...

virtual void	loadSettings (size_t)
	load the settings More...

Public Member Functions inherited from cass::AccumulatingProcessor
	AccumulatingProcessor (const name_t &name)
	constructor More...

virtual	~AccumulatingProcessor ()
	virtual destructor More...

virtual void	processEvent (const CASSEvent &evt)
	process the event More...

virtual const result_t &	result (const CASSEvent::id_t)
	retrieve a result. More...

virtual void	releaseEvent (const CASSEvent &)
	overwrite default behaviour to do nothing More...

virtual void	createHistList (result_t::shared_pointer result)
	create the list of results More...

Public Member Functions inherited from cass::Processor
	Processor (const name_t &name)
	constructor More...

virtual	~Processor ()
	virtual destructor More...

result_t::shared_pointer	resultCopy (const uint64_t eventid)
	retrieve histogram for id More...

virtual void	load ()
	load the general settings More...

virtual void	aboutToQuit ()
	function that will be called when the processor is about to be deleted More...

const names_t &	dependencies ()
	Define all processors keys a processor depends on. More...

void	clearDependencies ()
	clear the dependenies More...

void	clearHistograms ()
	clear the histograms More...

virtual void	processCommand (std::string command)
	process command in pp More...

const name_t	name () const
	retrieve the name of this processor More...

bool	hide () const
	retrieve the hide flag of this processor More...

const std::string &	comment () const
	retrieve the comment of this processor More...

Protected Member Functions
result_t::value_t	average (result_t::value_t val, result_t::value_t aveOld, result_t::value_t scale)
	function for normal averaging. More...

result_t::value_t	squareAverage (result_t::value_t val, result_t::value_t aveOld, result_t::value_t scale)
	function that will calculate the square average More...

result_t::value_t	cumulativeScale ()
	retrieve the cumulative scale for the current datum More...

result_t::value_t	movingScale ()
	retrieve the moving average scale after the initialization of the "first" point More...

result_t::value_t	movingInitializationScale ()
	retrieve the moving average scale during the initialization More...

Protected Member Functions inherited from cass::Processor
virtual void	createHistList (result_t::shared_pointer result)
	create result list. More...

void	setupGeneral ()
	general setup of the processor More...

bool	setupCondition (bool defaultConditionType=true)
	setup the condition. More...

shared_pointer	setupDependency (const std::string &depVarName, const name_t &name="")
	setup the dependecy. More...

Protected Attributes
result_t::value_t	_alpha
	alpha for the running average More...

std::tr1::function< result_t::value_t(result_t::value_t, result_t::value_t, result_t::value_t)>	_func
	function that will do the averagin More...

std::tr1::function< result_t::value_t()>	_scale
	function that will return the scale that should be used for the average More...

shared_pointer	_pHist
	processor containing result to average More...

Protected Attributes inherited from cass::AccumulatingProcessor
result_t::shared_pointer	_result
	the result that accumulates the events More...

size_t	_nbrEventsAccumulated
	the number of events the processor has accumulated More...

Protected Attributes inherited from cass::Processor
const name_t	_name
	the processors name More...

bool	_hide
	flag to tell whether this pp should be hidden in the dropdown list More...

std::string	_comment
	optional comment that one can add to a processor. More...

CachedList	_resultList
	the list of results More...

names_t	_dependencies
	the list of dependencies More...

shared_pointer	_condition
	pointer to the processor that will contain the condition More...

Additional Inherited Members
Public Types inherited from cass::Processor
typedef std::tr1::shared_ptr< Processor >	shared_pointer
	a shared pointer of this More...

typedef std::string	name_t
	define the name type More...

typedef std::list< name_t >	names_t
	define the list of names More...

typedef CachedList::item_type	result_t
	define the results More...

typedef CachedList::item_sp	resultsp_t
	define the shared pointer to the result More...

Detailed Description

result averaging.

Available PostProcessors:: "61": Average of a result

Running or cummulative average of a result. Could also be a squared average.

See also: Processor for a list of all commonly available cass.ini settings.

User definable Parameters:

Processor/%name%/{AveragingType}
The type of averaging that should be performed. Default is "Normal"

"Normal": a normal average will be used
"Square"; a square averaging will be performed

Processor/%name%/{NbrOfAverages}
how many images should be averaged. When value is 0 its a cummulative average. Default is 1.

Processor/%name%/{InputName}
processor name containing the result that we average.

Author: Lutz Foucar

Definition at line 1096 of file operations.h.

Constructor & Destructor Documentation

pp61::pp61 ( const name_t & name )

constructor

Definition at line 1477 of file operations.cpp.

References loadSettings().

Member Function Documentation

Processor::result_t::value_t pp61::average	(	result_t::value_t	val,
		result_t::value_t	aveOld,
		result_t::value_t	scale
	)

protected

function for normal averaging.

this operator is capable of performing a cumulative moving average and a Exponential moving average.

See also: http://en.wikipedia.org/wiki/Moving_average

the operator calculates the average using the function $ave_{new} = ave_{old} + \scale(val-ave_{old})$ where when $\scale$ is equal to N it is a cumulative moving average, otherwise it will be a exponential moving average.

Returns: the new average

Parameters

val	the current value to be added to the average
aveOld	the old average
scale	the scale with which the new value will be weighted

Definition at line 1536 of file operations.cpp.

Referenced by loadSettings().

result_t::value_t cass::pp61::cumulativeScale ( )

inlineprotected

retrieve the cumulative scale for the current datum

Definition at line 1145 of file operations.h.

References cass::AccumulatingProcessor::_nbrEventsAccumulated.

Referenced by loadSettings(), and movingInitializationScale().

void pp61::loadSettings ( size_t )

virtual

load the settings

: In case we're running a moving average (_alpha is non-zero), we need to initialize the the first value of the moving average. Here we initialize it with the mean of the first datapoints, meaning we'll check if the _alpha calculated for the cumulative averaging is still larger than the _alpha for the moving averge. If its smaller than the _alpha of the cumulative average it means that the current datum will not have as much weight. Therefore from this point onward, we should use the _alpha of the moving averge to start "forgetting" older values by putting comparatively more weight on the newer datums.

: initialize the result to be zero, otherwise the cumulatinve averaging won't produce correct results

Reimplemented from cass::Processor.

Definition at line 1483 of file operations.cpp.

References _alpha, cass::Processor::_condition, _func, _pHist, cass::AccumulatingProcessor::_result, _scale, cass::Log::add(), average(), QSettings::beginGroup(), cass::AccumulatingProcessor::createHistList(), cumulativeScale(), QString::fromStdString(), cass::Log::INFO, movingInitializationScale(), cass::Processor::name(), cass::Processor::setupCondition(), cass::Processor::setupDependency(), cass::Processor::setupGeneral(), squareAverage(), cass::toString(), and QSettings::value().

Referenced by pp61().

Processor::result_t::value_t pp61::movingInitializationScale ( )

protected

retrieve the moving average scale during the initialization

when the first value of the moving avererage has been initialized (see comment above), we can then procede by using the moving average's _alpha

Definition at line 1550 of file operations.cpp.

References _alpha, _scale, cumulativeScale(), and movingScale().

Referenced by loadSettings().

result_t::value_t cass::pp61::movingScale ( )

inlineprotected

retrieve the moving average scale after the initialization of the "first" point

Definition at line 1153 of file operations.h.

References _alpha.

Referenced by movingInitializationScale().

void pp61::process	(	const CASSEvent &	evt,
		result_t &	result
	)

virtual

process event

: the nbr of accumulated events needs to be increased BEFORE calculating the scale because the cumulative scale expect the scale to be $\frac{1}{n+1}$

Reimplemented from cass::Processor.

Definition at line 1564 of file operations.cpp.

References _func, cass::AccumulatingProcessor::_nbrEventsAccumulated, _pHist, _scale, cass::Result< T >::begin(), cass::Result< T >::datasize(), cass::CASSEvent::id(), and cass::Result< T >::lock.

Processor::result_t::value_t pp61::squareAverage	(	result_t::value_t	val,
		result_t::value_t	aveOld,
		result_t::value_t	scale
	)

protected

function that will calculate the square average

the operator calculates the square average using the function $ave_{new} = ave_{old} + \scale(val*val-ave_{old})$ where when $\scale$ is equal to N it is a cumulative moving average.