org.hypergraphdb.app.wordnet

Class WNGraph

java.lang.Object

org.hypergraphdb.app.wordnet.WNGraph

public class WNGraph
extends Object

A collection of utility methods for applying graph traversals and algorithms onto the HyperGraph representation of WordNet.

Author:

Borislav Iordanov

Constructor Summary

Constructors

Constructor and Description
WNGraph(org.hypergraphdb.HyperGraph graph)

Method Summary

Methods

Modifier and Type	Method and Description
org.hypergraphdb.HGHandle	findIsaRoot(org.hypergraphdb.HGHandle synset) Find the top-level ancestor of node in an IS-A hierarchy.
List<org.hypergraphdb.HGHandle>	findRelated(org.hypergraphdb.HGHandle synset, Class<? extends SemanticLink> type, boolean leftToRight)
org.hypergraphdb.HGHandle	findWord(String lemma) Return the HGHandle of a WordNet Word instance given its lemma.
List<org.hypergraphdb.HGHandle>	getAdjSenses(org.hypergraphdb.HGHandle word) Return all senses of word when used as an adjective.
List<org.hypergraphdb.HGHandle>	getAdverbSenses(org.hypergraphdb.HGHandle word) Return all senses of word when used as an adverb.
org.hypergraphdb.HyperGraph	getGraph() Return the underlying HyperGraph instance.
long	getIsaDepth(org.hypergraphdb.HGHandle synset, org.hypergraphdb.HGHandle root) Assuming synset is a descendant of root in an IS-A hierarchy, finding the distance between them.
org.hypergraphdb.HGHandle	getNounIsaRoot() Retrieve the root of the noun ISA hierarchy.
List<org.hypergraphdb.HGHandle>	getNounSenses(org.hypergraphdb.HGHandle word) Return all senses of word when used as a noun.
org.hypergraphdb.HGHandle	getSenseById(long id) Return a WordNet sense by its offset/id in the original WordNet files.
List<org.hypergraphdb.HGHandle>	getSenses(org.hypergraphdb.HGHandle word, Class<? extends SynsetLink> type) Retrieve all senses of a particular word if used as a specific part-of-speech.
List<org.hypergraphdb.HGHandle>	getSenses(org.hypergraphdb.HGHandle word, org.hypergraphdb.HGHandle senseType) Retrieve all senses of a particular word if used as a specific part-of-speech.
org.hypergraphdb.HGHandle	getSenseType(Pos pos) Retrieve the sense type (i.e.
<T> WNStat<T>	getStatistic(Class<? extends WNStat<T>> type) Return the instance of a global WordNet statistic by its type.
<T> T	getStatisticValue(Class<? extends WNStat<T>> type) Return the value of a global WordNet statistic by its type or null if that statistic hasn't been calculated yet.
org.hypergraphdb.HGHandle	getVerbIsaRoot() Return the root verb sense of the IS-A hierarchy.
List<org.hypergraphdb.HGHandle>	getVerbSenses(org.hypergraphdb.HGHandle word) Return all senses of word when used as a verb.
org.hypergraphdb.algorithms.HGALGenerator	isaRelatedGenerator(boolean returnChildren, boolean returnParents) Return a generator for the Isa semantic relationships.
org.hypergraphdb.algorithms.HGTraversal	relatedBreadthFirst(org.hypergraphdb.HGHandle start, Class<? extends SemanticLink> type, boolean leftToRight)
org.hypergraphdb.algorithms.HGTraversal	relatedDepthFirst(org.hypergraphdb.HGHandle start, Class<? extends SemanticLink> type, boolean leftToRight)
org.hypergraphdb.algorithms.HGALGenerator	relatedGenerator(Class<? extends SemanticLink> type) Get a generator for a symmetric semantic relationship (such as the Similar relationships between adjective senses.
org.hypergraphdb.algorithms.HGALGenerator	relatedGenerator(Class<? extends SemanticLink> type, boolean leftToRight) Get an atom adjency generator for a WordNet semantic relationship.

Methods inherited from class java.lang.Object

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

Constructor Detail

WNGraph

public WNGraph(org.hypergraphdb.HyperGraph graph)

Method Detail

getGraph

public org.hypergraphdb.HyperGraph getGraph()

Return the underlying HyperGraph instance.

findWord

public org.hypergraphdb.HGHandle findWord(String lemma)

Return the HGHandle of a WordNet Word instance given its lemma.

Parameters:

lemma -

Returns:

getStatisticValue

public <T> T getStatisticValue(Class<? extends WNStat<T>> type)

Return the value of a global WordNet statistic by its type or null if that statistic hasn't been calculated yet.

Type Parameters:

T -

Parameters:

type -

Returns:

getStatistic

public <T> WNStat<T> getStatistic(Class<? extends WNStat<T>> type)

Return the instance of a global WordNet statistic by its type. A new instance will be created and added to the HyperGraphDB if it doesn't already exist. Note that the statistic will NOT automatically calculated. If you have not previously triggered a calculation of the statistic and want to lazily calculate it as soon as it is needed, call its isCalculated to check and trigger the calculation on the spot.

Type Parameters:

T -

Parameters:

type -

Returns:

findIsaRoot

public org.hypergraphdb.HGHandle findIsaRoot(org.hypergraphdb.HGHandle synset)

Find the top-level ancestor of node in an IS-A hierarchy. Note that this doesn't perform any cycle detection - it could loop forever if the underlying graph structure is not a DAG.

Parameters:

synset - A sense participating in an IS-A hierarchy.

Returns:

The top-level ancestor of synset

getIsaDepth

public long getIsaDepth(org.hypergraphdb.HGHandle synset,
               org.hypergraphdb.HGHandle root)

Assuming synset is a descendant of root in an IS-A hierarchy, finding the distance between them.

In case the method finds out that synset is not a descendent of root, then -1 is returned.

Parameters:

synset - The descendant.

root - The root.

Returns:

The distance between synset and root in an IS-A DAG.

getNounIsaRoot

public org.hypergraphdb.HGHandle getNounIsaRoot()

Retrieve the root of the noun ISA hierarchy. Recent versions of WordNet have the word "entity" mapped in a single synset which is the top-level noun sense.

Returns:

getVerbIsaRoot

public org.hypergraphdb.HGHandle getVerbIsaRoot()

Return the root verb sense of the IS-A hierarchy. Unlike nouns, WordNet doesn't define a top-level verb sense (simply because there isn't a good enough candidate in the English vocabulary). However, for computational purposes we can connect all top-level nodes in the verb IS-A DAG to a single root.

This method will return the dummy verb sense that serves as the top-level node. If it hasn't been already create, it will create and connect it to all existing top-level nodes. Which, of course, is somewhat computational intensive.

isaRelatedGenerator

public org.hypergraphdb.algorithms.HGALGenerator isaRelatedGenerator(boolean returnChildren,
                                                            boolean returnParents)

Return a generator for the Isa semantic relationships. The generator will examine atoms that are related to the current one through Isa or one of its derived relationship types (InstanceOf, KindOf).

Parameters:

returnChildren - Whether to include more specific (i.e. hyponym) concepts than the current one being examined.

returnParents - Whether to include more general (i.e. hypernym) concepts than the current one being examined.

Returns:

The adjency generator.

relatedGenerator

public org.hypergraphdb.algorithms.HGALGenerator relatedGenerator(Class<? extends SemanticLink> type)

Get a generator for a symmetric semantic relationship (such as the Similar relationships between adjective senses.

relatedGenerator

public org.hypergraphdb.algorithms.HGALGenerator relatedGenerator(Class<? extends SemanticLink> type,
                                                         boolean leftToRight)

Get an atom adjency generator for a WordNet semantic relationship.

Parameters:

type - One of the SemanticLink sub-classes.

leftToRight - true if the second targets of the relationships should be traversed and false if the first targets should be traversed.

relatedDepthFirst

public org.hypergraphdb.algorithms.HGTraversal relatedDepthFirst(org.hypergraphdb.HGHandle start,
                                                        Class<? extends SemanticLink> type,
                                                        boolean leftToRight)

relatedBreadthFirst

public org.hypergraphdb.algorithms.HGTraversal relatedBreadthFirst(org.hypergraphdb.HGHandle start,
                                                          Class<? extends SemanticLink> type,
                                                          boolean leftToRight)

findRelated

public List<org.hypergraphdb.HGHandle> findRelated(org.hypergraphdb.HGHandle synset,
                                          Class<? extends SemanticLink> type,
                                          boolean leftToRight)

getSenseById

public org.hypergraphdb.HGHandle getSenseById(long id)

Return a WordNet sense by its offset/id in the original WordNet files. Note that this id is not stable across WordNet versions. However, in general, there is a way to convert ids between consecutive versions so it's relatively safe to use it as an identifier outside HyperGraphDB.

Parameters:

id -

Returns:

getSenseType

public org.hypergraphdb.HGHandle getSenseType(Pos pos)

Retrieve the sense type (i.e. the concrete SynsetLink type atom) for a given part of speech.

Parameters:

pos - The part of speech.

Returns:

the HGHandle of the type atom representing the sense type.

getSenses

public List<org.hypergraphdb.HGHandle> getSenses(org.hypergraphdb.HGHandle word,
                                        org.hypergraphdb.HGHandle senseType)

Retrieve all senses of a particular word if used as a specific part-of-speech. The part-of-speech is identified by the senseType parameter.

Alternatively, you can use one of the getTSenses methods where T is one of Noun, Verb, Adverb or Adj.

Parameters:

word - The word handle.

senseType - The sense type - this corresponds to one of the sub-classes of the SynsetLink class.

Returns:

getSenses

public List<org.hypergraphdb.HGHandle> getSenses(org.hypergraphdb.HGHandle word,
                                        Class<? extends SynsetLink> type)

Retrieve all senses of a particular word if used as a specific part-of-speech. The part-of-speech is identified by the senseType parameter.

Alternatively, you can use one of the getTSenses methods where T is one of Noun, Verb, Adverb or Adj.

Parameters:

word - The word handle.

senseType - The sense type - one of the sub-classes of the SynsetLink class.

Returns:

getNounSenses

public List<org.hypergraphdb.HGHandle> getNounSenses(org.hypergraphdb.HGHandle word)

Return all senses of word when used as a noun.

getVerbSenses

public List<org.hypergraphdb.HGHandle> getVerbSenses(org.hypergraphdb.HGHandle word)

Return all senses of word when used as a verb.

getAdverbSenses

public List<org.hypergraphdb.HGHandle> getAdverbSenses(org.hypergraphdb.HGHandle word)

Return all senses of word when used as an adverb.

getAdjSenses

public List<org.hypergraphdb.HGHandle> getAdjSenses(org.hypergraphdb.HGHandle word)

Return all senses of word when used as an adjective.