inpro.irmrsc.parser
Class CandidateAnalysis

java.lang.Object
  inpro.irmrsc.parser.CandidateAnalysis

All Implemented Interfaces:

java.lang.Comparable<CandidateAnalysis>

public class CandidateAnalysis
extends java.lang.Object
implements java.lang.Comparable<CandidateAnalysis>

A candidate analysis entertained by the SITDBSParser, most importantly consisting of a stack, a derivation history and the probability of the derivation.

The main parsing operations modifying a candidate analysis are implemented here: The classic top down parser actions predict (expand(Production)) and match (match(Symbol)), as well as the robust operations for insertions (insert(Symbol)), deletions (deletion(Symbol)) and repairs (repair(Symbol)). The probability of a derivation can be externally degraded, as in the reference-pruning-setting.

Author:

Andreas Peldszus

Field Summary

private CandidateAnalysis	mAntecedent the former analysis of which this is an extension
private java.util.List<java.lang.String>	mDerivation the sequence of all rules used in this derivation
private double	mFigureOfMerit the product of the derivations probability and the lookahead probability -- not used at the moment
private java.util.List<java.lang.String>	mLastDerive the sequence of rules used in the last incremental step of the derivation
private double	mProbability the product of all rule probabilities used in the derivation and all external mali
private java.util.Deque<Symbol>	mStack the sequence of nonterminal symbols that need to be accounted for

Constructor Summary

CandidateAnalysis(CandidateAnalysis ca)
copy constructor

CandidateAnalysis(java.util.Deque<Symbol> mStack)
constructor to initiate a chain of derivations

CandidateAnalysis(java.util.List<java.lang.String> mDerivation, java.util.List<java.lang.String> mLastDerive, CandidateAnalysis mAntecedent, java.util.Deque<Symbol> mStack, double mProbability, double mFigureOfMerit)

Method Summary

int	compareTo(CandidateAnalysis y)
void	consumeFiller(java.lang.String fillername) simply adds a filler to this derivation without changing anything else
void	degradeProbability(double f)
CandidateAnalysis	deletion(Symbol deletedToken) The robust 'deletion' parsing action: An expected but non-existing token is considered to be existing and simply matched.
CandidateAnalysis	expand(Production p) The 'predict' parsing action: The top stack element is popped and the symbols on the righthandside of the predicte rule are pushed onto the stack.
CandidateAnalysis	getAntecedent()
double	getFigureOfMerit()
java.util.List<java.lang.String>	getLastDerive()
int	getNumberOfDeletions()
int	getNumberOfInsertions()
int	getNumberOfMatches()
int	getNumberOfRepairs()
double	getProbability()
java.util.Deque<Symbol>	getStack()
Symbol	getTopSymbol()
boolean	hasRobustOperationsLately()
CandidateAnalysis	insert(Symbol insertedToken) The robust 'insertion' parsing action: A unexpected token is added to the derivation without altering the stack.
boolean	isCompletable(java.util.Set<Symbol> epsilonproductions)
boolean	isComplete()
private CandidateAnalysis	match_intern(java.lang.String deriveIdentifier)
CandidateAnalysis	match(Symbol nextToken) The 'match' parsing action: An expected token is matched.
void	newIncrementalStep(CandidateAnalysis oldCA) links this analysis to an antecedent analysis
java.lang.String	printDerivation()
CandidateAnalysis	repair(Symbol requiredToken) The robust 'repair' parsing action: An unexpected token is replaced by the expected token and simply matched.
java.lang.String	toFinalString()
java.lang.String	toFullString()
java.lang.String	toString()

Methods inherited from class java.lang.Object

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

Field Detail

mDerivation

private java.util.List<java.lang.String> mDerivation

the sequence of all rules used in this derivation

mLastDerive

private java.util.List<java.lang.String> mLastDerive

the sequence of rules used in the last incremental step of the derivation

mAntecedent

private CandidateAnalysis mAntecedent

the former analysis of which this is an extension

mStack

private java.util.Deque<Symbol> mStack

the sequence of nonterminal symbols that need to be accounted for

mProbability

private double mProbability

the product of all rule probabilities used in the derivation and all external mali

mFigureOfMerit

private double mFigureOfMerit

the product of the derivations probability and the lookahead probability -- not used at the moment

Constructor Detail

CandidateAnalysis

public CandidateAnalysis(java.util.List<java.lang.String> mDerivation,
                         java.util.List<java.lang.String> mLastDerive,
                         CandidateAnalysis mAntecedent,
                         java.util.Deque<Symbol> mStack,
                         double mProbability,
                         double mFigureOfMerit)

CandidateAnalysis

public CandidateAnalysis(java.util.Deque<Symbol> mStack)

constructor to initiate a chain of derivations

CandidateAnalysis

public CandidateAnalysis(CandidateAnalysis ca)

copy constructor

Method Detail

expand

public CandidateAnalysis expand(Production p)

The 'predict' parsing action: The top stack element is popped and the symbols on the righthandside of the predicte rule are pushed onto the stack. The derivations probability is multiplied with the probability of the predicted rule.

Parameters:

p - the grammar rule that is predicted

Returns:

the resulting new candidate analysis

repair

public CandidateAnalysis repair(Symbol requiredToken)

The robust 'repair' parsing action: An unexpected token is replaced by the expected token and simply matched. The stack's top element is popped.

Parameters:

requiredToken - the token that was expected next in the derivation

Returns:

the resulting new candidate analysis

deletion

public CandidateAnalysis deletion(Symbol deletedToken)

The robust 'deletion' parsing action: An expected but non-existing token is considered to be existing and simply matched. The stack's top element is popped.

Parameters:

deletedToken - the token that was expected but not found

Returns:

the resulting new candidate analysis

match

public CandidateAnalysis match(Symbol nextToken)

The 'match' parsing action: An expected token is matched. The stack's top element is popped.

Parameters:

nextToken - the expected token

Returns:

the resulting new candidate analysis

match_intern

private CandidateAnalysis match_intern(java.lang.String deriveIdentifier)

insert

public CandidateAnalysis insert(Symbol insertedToken)

The robust 'insertion' parsing action: A unexpected token is added to the derivation without altering the stack.

Parameters:

insertedToken - the unexpected token to be inserted

Returns:

the resulting new candidate analysis

consumeFiller

public void consumeFiller(java.lang.String fillername)

simply adds a filler to this derivation without changing anything else

newIncrementalStep

public void newIncrementalStep(CandidateAnalysis oldCA)

links this analysis to an antecedent analysis

getAntecedent

public CandidateAnalysis getAntecedent()

isComplete

public boolean isComplete()

isCompletable

public boolean isCompletable(java.util.Set<Symbol> epsilonproductions)

getTopSymbol

public Symbol getTopSymbol()

getStack

public java.util.Deque<Symbol> getStack()

getProbability

public double getProbability()

degradeProbability

public void degradeProbability(double f)

getFigureOfMerit

public double getFigureOfMerit()

getLastDerive

public java.util.List<java.lang.String> getLastDerive()

hasRobustOperationsLately

public boolean hasRobustOperationsLately()

getNumberOfMatches

public int getNumberOfMatches()

getNumberOfRepairs

public int getNumberOfRepairs()

getNumberOfInsertions

public int getNumberOfInsertions()

getNumberOfDeletions

public int getNumberOfDeletions()

compareTo

public int compareTo(CandidateAnalysis y)

Specified by:

compareTo in interface java.lang.Comparable<CandidateAnalysis>

toFinalString

public java.lang.String toFinalString()

toFullString

public java.lang.String toFullString()

toString

public java.lang.String toString()

Overrides:

toString in class java.lang.Object

printDerivation

public java.lang.String printDerivation()