cluster
Class Phrase

java.lang.Object
  cluster.Phrase

All Implemented Interfaces:

java.lang.Iterable<Term>

public class Phrase
extends java.lang.Object
implements java.lang.Iterable<Term>

A key component of the unsupervised clustering. The phrases hold the primary phrase that labels the cluster and alternative phrases. In addition, the phrases hold the document that belong to the cluster. Finally, each of these Phrases acts like a tree node by pointing to a parent node (null if doesn't exist) and to all children.

Author:

davidc

Constructor Summary

Phrase()

Phrase(java.util.List<Term> combinedOverlap)
Constructor based on just the terms of the primary phrase

Phrase(Phrase t)
Copy constructor

Phrase(java.lang.String terms)

Method Summary

boolean	add(Term arg0)
void	addAlternativePhrase(java.util.List<Term> ts) Add an alternative phrase, which is sufficient to containing all phrases held by this Phrase
void	addChild(Phrase t)
void	addCoveredDoc(ClusterDoc d) cover is a Set anyways, so don't worry about checking before calling this method
void	addCoveredDocs(java.util.Set<ClusterDoc> cover2)
void	assignId(int i)
void	clear()
void	clearChildren()
void	clearCover()
static Phrase	CombinedFactory(Phrase a, Phrase b)
int	coverSize()
void	deleteSelf() Has no effect if this termset has no parent
Term	get(int arg0)
java.util.List<Term>	getCombinedOverlap(Phrase set2) Assuming this set is in front
java.lang.String	getCondensedString() Get all fixed forms of the primary phrase concatenated together
java.util.Set<ClusterDoc>	getCover() Returns those documents that belong to this phrase by themself If a paper contains a child phrase but not this phrase, then this method will not return the paper although getNestedCover() will
int	getCoverSize()
java.util.Set<ClusterDoc>	getCoverUnion(Phrase s)
int	getId()
java.lang.String	getLabel()
int	getMaxIndex()
java.util.Set<ClusterDoc>	getNestedCover() All documents that support/contain this term or child terms
java.lang.Integer	getNumChildren() Recursively computes number of child nodes
int	getNumTermsShared(Phrase t)
Phrase	getParent()
java.util.List<java.util.List<Term>>	getPhrases() Gets all list of terms that describe this cluster Containing any of the list of terms is equivalent to containing this Phrase (ie.
Term	getTerm(int j)
java.util.List<Term>	getTerms()
java.util.List<Phrase>	getTermSetChildren()
java.util.Iterator<Term>	iterator()
int	numAlternativePhrases()
boolean	overLapTerms(Phrase set)
void	removeChild(Phrase toMove) toMove's parent is set to null, so assign a new parent after calling this function
void	removeCoveredDoc(TestDoc d)
void	setParent(Phrase parent)
void	setTerms(java.util.List<Term> terms)
int	size()
java.lang.String	toString() Returns the primary phrase (the label to describe this cluster)

Methods inherited from class java.lang.Object

equals, getClass, hashCode, notify, notifyAll, wait, wait, wait

Constructor Detail

Phrase

public Phrase()

Phrase

public Phrase(Phrase t)

Copy constructor

Parameters:

t -

Phrase

public Phrase(java.util.List<Term> combinedOverlap)

Constructor based on just the terms of the primary phrase

Parameters:

combinedOverlap -

Phrase

public Phrase(java.lang.String terms)

Method Detail

getCondensedString

public java.lang.String getCondensedString()

Get all fixed forms of the primary phrase concatenated together

Returns:

getPhrases

public java.util.List<java.util.List<Term>> getPhrases()

Gets all list of terms that describe this cluster
Containing any of the list of terms is equivalent to containing this Phrase (ie. containing one should be considered as containing all)

Returns:

addAlternativePhrase

public void addAlternativePhrase(java.util.List<Term> ts)

Add an alternative phrase, which is sufficient to containing all phrases held by this Phrase

Parameters:

ts -

addChild

public void addChild(Phrase t)

getTermSetChildren

public java.util.List<Phrase> getTermSetChildren()

add

public boolean add(Term arg0)

get

public Term get(int arg0)

size

public int size()

iterator

public java.util.Iterator<Term> iterator()

Specified by:

iterator in interface java.lang.Iterable<Term>

toString

public java.lang.String toString()

Returns the primary phrase (the label to describe this cluster)

Overrides:

toString in class java.lang.Object

getMaxIndex

public int getMaxIndex()

addCoveredDoc

public void addCoveredDoc(ClusterDoc d)

cover is a Set anyways, so don't worry about checking before calling this method

Parameters:

d -

getNestedCover

public java.util.Set<ClusterDoc> getNestedCover()

All documents that support/contain this term or child terms

Returns:

getTerms

public java.util.List<Term> getTerms()

CombinedFactory

public static Phrase CombinedFactory(Phrase a,
                                     Phrase b)

setTerms

public void setTerms(java.util.List<Term> terms)

coverSize

public int coverSize()

getNumTermsShared

public int getNumTermsShared(Phrase t)

getCoverUnion

public java.util.Set<ClusterDoc> getCoverUnion(Phrase s)

overLapTerms

public boolean overLapTerms(Phrase set)

getCombinedOverlap

public java.util.List<Term> getCombinedOverlap(Phrase set2)

Assuming this set is in front

Parameters:

set2 -

Returns:

getTerm

public Term getTerm(int j)

addCoveredDocs

public void addCoveredDocs(java.util.Set<ClusterDoc> cover2)

clear

public void clear()

assignId

public void assignId(int i)

getId

public int getId()

getParent

public Phrase getParent()

setParent

public void setParent(Phrase parent)

clearChildren

public void clearChildren()

removeChild

public void removeChild(Phrase toMove)

toMove's parent is set to null, so assign a new parent after calling this function

Parameters:

toMove -

removeCoveredDoc

public void removeCoveredDoc(TestDoc d)

clearCover

public void clearCover()

getCover

public java.util.Set<ClusterDoc> getCover()

Returns those documents that belong to this phrase by themself
If a paper contains a child phrase but not this phrase, then this method will not return the paper although getNestedCover() will

Returns:

deleteSelf

public void deleteSelf()

Has no effect if this termset has no parent

numAlternativePhrases

public int numAlternativePhrases()

getCoverSize

public int getCoverSize()

getLabel

public java.lang.String getLabel()

getNumChildren

public java.lang.Integer getNumChildren()

Recursively computes number of child nodes

Returns: