System documentation of the GNU Image-Finding Tool

CQInvertedFile Class Reference

The Query manager for Queries on inverted Files. More...

#include <CQInvertedFile.h>

Inheritance diagram for CQInvertedFile:

List of all members.

Public Member Functions
void	finishInit ()
	finish the initialisation phase make the weighting function know who its normalizers are
	CQInvertedFile (CAccessorAdminCollection &inAccessorAdminCollection, CAlgorithm &inAlgorithm)
	New constructor, taking as parameter the algorithm structure, which contains all the algorithm configuration.
	~CQInvertedFile ()
	destructor
virtual bool	setAlgorithm (CAlgorithm &inAlgorithm)
	set the Algorithm.
virtual CIDRelevanceLevelPairList *	fastQuery (const CXMLElement &inQuery, int inNumberOfInterestingImages, double inDifferenceToBest)
	a query which returns ID/RelevanceLevel pairs instead of instead of URL/RelevanceLevel pairs; this is faster for merging tasks (to explain the name) queries for URLs are answered by query (in herited)
virtual CIDRelevanceLevelPairList *	fastQueryByFeature (const CWeightingFunctionPointerList &inQuery, int inNumberOfInterestingImages, double inDifferenceToBest)
	Assuming that a correct CWeightingFunctionPointerHash has been built by fastQuery (or another function), this function will do the rest.
double	DIDToScore (TID inDID, const CWeightingFunctionPointerHash &inQuery) const
	Returns the score for one image in the query.
double	URLToScore (const string &inURL, const CWeightingFunctionPointerHash &inQuery) const
	Returns the score for one image in the query.
double	FeatureListToScore (TID inDID, const CDocumentFrequencyList &inFeatureList, const CWeightingFunctionPointerHash &inQuery) const
	Turns the featurelist of a document into a score.
void	buildNormalizedQueryHash (const CRelevanceLevel &inQuery, CWeightingFunctionPointerHash &outQueryFeatureWeighters) const
void	activateBlockingFeatures ()
	activateBlockingFeatures - sets the variable to block groups of features
void	releaseBlockingFeatures ()
	releaseBlockingFeatures - this turns off the blocking of features
bool	featuresBlocked () const
	featuresBlocked - returns true if the features are blocked
void	blockFeatureGroup (const int featureNumber)
	blockFeatureGroup - this blocks one special group of features
void	unblockFeatureGroup (const int featureNumber)
	unblockFeatureGroup - this releases the blocking of one feature group
bool	isBlocked (const int featureNumber) const
	isBlocked - returns true if the feature is blocked and false if not
void	releaseAllPrunings ()
	sets back all the variables for the pruning
void	useFeaturePruning (double percentage)
	activates the feature pruning with evaluating a certain percentage of the features
void	releaseFeaturePruning ()
	set the variables back to its normal status
void	useTimePruning (double inTimeCutoffPoint)
	activates the time pruning with the give cuOffPoint
void	releaseTimePruning ()
	releases the timePruning
void	useScoreBoardPruning (double inCutAfterFraction, double inReduceToFactor)
	creates one entry in the scoreboardPruning
void	releaseScoreBoardPruning ()
	releases all the settings for the ScorboardPruning
void	useEvaluateAfterPruning ()
	activates the evaluation after the scoreboardPruning
void	releaseEvaluateAfterPruning ()
	releases the evaluation after the scoreboard pruning
CWeighter &	getWeighter ()
	For FerDeLance queries.
Public Attributes
bool	mBlockingOn
bool	mBlockingArray [MAXIMUMNUMBEROFEATUREGROUPS]
bool	mPruningUsed
	Shows if any pruning is used.
bool	mScoreBoardPruningUsed
	true if the scoreboardis Pruned
parameterPruningType	mScoreBoardPruningArray [MAX_SCOREBOARD_PRUNINGS]
	Henning Müller
int	mNumberofUsedScoreBoardPrunings
	Henning Müller
bool	mFeaturePruningUsed
double	mPercentageofFeatures
	Henning Müller
bool	mTimePruningUsed
double	mStoppingTime
	Henning Müller
bool	mEvaluateAfterPruning
Static Public Attributes
const int	MAXIMUMNUMBEROFEATUREGROUPS = 50
const int	MAX_SCOREBOARD_PRUNINGS = 10
	These are the variables and functions used for the pruning process.
Protected Member Functions
double	keepScore (CScoreBoard &inoutScoreBoard, const CWeightingFunctionPointerList &inFeatures, bool lPositive) const
	calculates the score for all the images based on a list of features
double	keepScorePruning (CScoreBoard &inoutScoreBoard, const CWeightingFunctionPointerList &inFeatures, bool inPositive, int inDesiredNumberOfDocuments) const
	calculates the results for the images using some sort of pruning
void	buildQueryHash (CRelevanceLevelList &inQuery, CWeightingFunctionPointerHash &outQueryFeatureWeighters) const
	creates a list of all the features which are in one or more of the query images should be replaced by a version which operates on IDs not URLs
void	buildNormalizedQueryList (double inPositiveRelevanceSum, double inNegativeRelevanceSum, CWeightingFunctionPointerHash &inQFW, CWeightingFunctionPointerList &outQFW) const
	Building a list of normalized WF from a hash of Weighting functions.
void	buildNormalizedQueryHash (double inPositiveRelevanceSum, double inNegativeRelevanceSum, CWeightingFunctionPointerHash &inQFW) const
	Building a list of normalized WF from a hash of Weighting functions.
void	init ()
	Initializer, used by both construcors.
Protected Attributes
CAlgorithm *	mDeb
int	mModulo
	You can choose to use only every lModulo-th feature.
int	mModuloClass
	Every lModulo-th feature, starting at lModuloClass.
CWeighterFactory	mWeighterFactory
	This is a factory for weighters.
CSelfDestroyPointer< CWeighter >	mWeighter
	The weighter currently used.
CSelfDestroyPointer< CWeightingFunctionPointerHash >	mQueryFeatureWeighters
	object of this type to have a list of the features with the weights

---

Detailed Description

The Query manager for Queries on inverted Files.

---

Member Function Documentation

void CQInvertedFile::buildQueryHash (  CRelevanceLevelList &  inQuery, CWeightingFunctionPointerHash &  outQueryFeatureWeighters )  const [protected]

creates a list of all the features which are in one or more of the query images should be replaced by a version which operates on IDs not URLs Author: Wolfgang Müller + Pruning: Henning Müller/reprogrammed WM 09-10-00

double CQInvertedFile::keepScore (  CScoreBoard &  inoutScoreBoard, const CWeightingFunctionPointerList &  inFeatures, bool  lPositive )  const [protected]

calculates the score for all the images based on a list of features Author: Wolfgang Müller

double CQInvertedFile::keepScorePruning (  CScoreBoard &  inoutScoreBoard, const CWeightingFunctionPointerList &  inFeatures, bool  inPositive, int  inDesiredNumberOfDocuments )  const [protected]

calculates the results for the images using some sort of pruning Author: Henning Müller

void CQInvertedFile::releaseAllPrunings (   )

sets back all the variables for the pruning Henning Müller

void CQInvertedFile::releaseEvaluateAfterPruning (   )

releases the evaluation after the scoreboard pruning Henning Müller

void CQInvertedFile::releaseFeaturePruning (   )

set the variables back to its normal status Henning Müller

void CQInvertedFile::releaseScoreBoardPruning (   )

releases all the settings for the ScorboardPruning Henning Müller

void CQInvertedFile::releaseTimePruning (   )

releases the timePruning Henning Müller

virtual bool CQInvertedFile::setAlgorithm (  CAlgorithm &  inAlgorithm  )  [virtual]

set the Algorithm. same scheme as in setCollection Reimplemented from CQuery.

void CQInvertedFile::useEvaluateAfterPruning (   )

activates the evaluation after the scoreboardPruning Henning Müller

void CQInvertedFile::useFeaturePruning (  double  percentage  )

activates the feature pruning with evaluating a certain percentage of the features Henning Müller

void CQInvertedFile::useScoreBoardPruning (  double  inCutAfterFraction, double  inReduceToFactor )

creates one entry in the scoreboardPruning Henning Müller

void CQInvertedFile::useTimePruning (  double  inTimeCutoffPoint  )

activates the time pruning with the give cuOffPoint Henning Müller

---

Member Data Documentation

const int CQInvertedFile::MAX_SCOREBOARD_PRUNINGS = 10 [static]

These are the variables and functions used for the pruning process. the maximum number of sets which can be used to prune the ScoreBoard

CWeighterFactory CQInvertedFile::mWeighterFactory [protected]

This is a factory for weighters. With this we can find the weighters

---

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

• CQInvertedFile.h

---

Need for discussion? Want to contribute? Contact