Class PairsTableCalculator

Object

PairsTableCalculator

All Implemented Interfaces:

FeatureCalculatorMulti<FeatureInputPairObjects>, FeatureTableCalculator<FeatureInputPairObjects>

public class PairsTableCalculator
extends Object
implements FeatureTableCalculator<FeatureInputPairObjects>

A feature-session to evaluate pairs of objects.

Successive pairs are evaluated in five different ways:

• the image in which the object exists (on {code listImage}) i.e. the energy-stack.
• the left-object in the pair (on listSingle)
• the right-object in the pair (on listSingle)
• the pair (on {code listPair})
• both objects merged together (on {code listSingle}}

Due to the predictable pattern, feature-calculations can be cached predictably and appropriately to avoid redundancies.

Two types of caching are applied to avoid redundancy:

• The internal calculation caches of first/second/merged are reused as the internal calculation sub-caches in pair.
• The entire results are cached (as a function of the input) for first/second, as the same inputs reappear multiple times.

Author:

Owen Feehan

Constructor Summary

Constructors

Constructor	Description
PairsTableCalculator(MergedPairsFeatures features)	Constructs a PairsTableCalculator with default inclusion and error suppression.
PairsTableCalculator(MergedPairsFeatures features, MergedPairsInclude include, boolean suppressErrors)

Method Summary

Modifier and Type	Method	Description
ResultsVector	calculate(FeatureInputPairObjects input)
ResultsVector	calculate(FeatureInputPairObjects input, FeatureList<FeatureInputPairObjects> featuresSubset)
ResultsVector	calculateMaybeSuppressErrors(FeatureInputPairObjects input, ErrorReporter errorReporter)	Calculates features for the input, optionally suppressing errors.
ResultsVector	calculateSuppressErrors(FeatureInputPairObjects input, ErrorReporter errorReporter)
FeatureNameList	createFeatureNames()	Creates a list of names for each feature (columns of the table).
FeatureTableCalculator<FeatureInputPairObjects>	duplicateForNewThread()	Makes a copy of the feature-store for a new thread.
int	sizeFeatures()
void	start(ImageInitialization initializtion, Optional<EnergyStack> energyStack, Logger logger)	Initializes a feature store that has the same structure as that previously created by createFeatures() from the same object.

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Methods inherited from interface org.anchoranalysis.feature.calculate.bound.FeatureCalculatorMulti

calculate

Constructor Details

PairsTableCalculator

public PairsTableCalculator(MergedPairsFeatures features)

Constructs a PairsTableCalculator with default inclusion and error suppression.

Parameters:

features - the merged pairs features to use

PairsTableCalculator

public PairsTableCalculator(MergedPairsFeatures features,
 MergedPairsInclude include,
 boolean suppressErrors)

Method Details

start

public void start(ImageInitialization initializtion,
 Optional<EnergyStack> energyStack,
 Logger logger)
           throws InitializeException

Description copied from interface: FeatureTableCalculator

Initializes a feature store that has the same structure as that previously created by createFeatures() from the same object.

Specified by:

start in interface FeatureTableCalculator<FeatureInputPairObjects>

Parameters:

initializtion - the image initialization context

energyStack - an optional energy stack for feature calculation

logger - a logger for reporting initialization progress or errors

Throws:

InitializeException - if initialization fails

calculate

public ResultsVector calculate(FeatureInputPairObjects input)
                        throws NamedFeatureCalculateException

Specified by:

calculate in interface FeatureCalculatorMulti<FeatureInputPairObjects>

Throws:

NamedFeatureCalculateException

calculate

public ResultsVector calculate(FeatureInputPairObjects input,
 FeatureList<FeatureInputPairObjects> featuresSubset)

Specified by:

calculate in interface FeatureCalculatorMulti<FeatureInputPairObjects>

calculateSuppressErrors

public ResultsVector calculateSuppressErrors(FeatureInputPairObjects input,
 ErrorReporter errorReporter)

Specified by:

calculateSuppressErrors in interface FeatureCalculatorMulti<FeatureInputPairObjects>

calculateMaybeSuppressErrors

public ResultsVector calculateMaybeSuppressErrors(FeatureInputPairObjects input,
 ErrorReporter errorReporter)
                                           throws NamedFeatureCalculateException

Calculates features for the input, optionally suppressing errors.

Parameters:

input - the input pair of objects

errorReporter - the error reporter to use

Returns:

the calculated feature results

Throws:

NamedFeatureCalculateException - if an error occurs during calculation and error suppression is disabled

createFeatureNames

public FeatureNameList createFeatureNames()

Description copied from interface: FeatureTableCalculator

Creates a list of names for each feature (columns of the table).

Specified by:

createFeatureNames in interface FeatureTableCalculator<FeatureInputPairObjects>

Returns:

the list of feature names

sizeFeatures

public int sizeFeatures()

Specified by:

sizeFeatures in interface FeatureCalculatorMulti<FeatureInputPairObjects>

duplicateForNewThread

public FeatureTableCalculator<FeatureInputPairObjects> duplicateForNewThread()

Description copied from interface: FeatureTableCalculator

Makes a copy of the feature-store for a new thread.

Deep-copies the features. Shallow-copies everything else.

Specified by:

duplicateForNewThread in interface FeatureTableCalculator<FeatureInputPairObjects>

Returns:

the copied feature-store