org.aksw.simba.bengal.verbalizer

Class Verbalizer

java.lang.Object

org.aksw.simba.bengal.verbalizer.Verbalizer

public class Verbalizer
extends Object

A verbalizer for triples without variables.

Author:

ngonga

Field Summary

Fields

Modifier and Type	Field and Description
private String	cacheDirectory
private static String	DEFAULT_CACHE_BASE_DIR
private static org.aksw.avatar.dataset.DatasetBasedGraphGenerator.Cooccurrence	DEFAULT_COOCCURRENCE_TYPE
private static org.aksw.avatar.clustering.hardening.HardeningFactory.HardeningType	DEFAULT_HARDENING_TYPE
private static int	DEFAULT_MAX_NUMBER_OF_SHOWN_VALUES_PER_PROPERTY
private static double	DEFAULT_THRESHOLD
(package private) org.dllearner.kb.sparql.SparqlEndpoint	endpoint
(package private) org.aksw.triple2nl.gender.TypeAwareGenderDetector	gender
org.aksw.avatar.dataset.DatasetBasedGraphGenerator	graphGenerator
(package private) Map<org.apache.jena.rdf.model.Resource,String>	labels
(package private) String	language
(package private) org.aksw.avatar.rules.NumericLiteralFilter	litFilter
private static org.apache.log4j.Logger	logger
(package private) int	maxShownValuesPerProperty
org.aksw.sparql2nl.naturallanguagegeneration.SimpleNLGwithPostprocessing	nlg
(package private) boolean	omitContentInBrackets
(package private) org.aksw.avatar.rules.ObjectMergeRule	or
(package private) org.aksw.avatar.rules.PredicateMergeRule	pr
private org.aksw.jena_sparql_api.core.QueryExecutionFactory	qef
protected simplenlg.realiser.english.Realiser	realiser
Map<org.apache.jena.rdf.model.Resource,Collection<org.apache.jena.graph.Triple>>	resource2Triples
(package private) org.aksw.avatar.rules.SubjectMergeRule	sr

Constructor Summary

Constructors

Constructor and Description
Verbalizer(org.aksw.jena_sparql_api.core.QueryExecutionFactory qef, String cacheDirectory)
Verbalizer(org.dllearner.kb.sparql.SparqlEndpoint endpoint)
Verbalizer(org.dllearner.kb.sparql.SparqlEndpoint endpoint, String cacheDirectory)

Method Summary

Modifier and Type	Method and Description
List<simplenlg.framework.NLGElement>	applyMergeRules(List<simplenlg.phrasespec.SPhraseSpec> triples, org.aksw.triple2nl.gender.Gender g) Generates a set of sentences by merging the sentences in the list as well as possible
List<simplenlg.framework.NLGElement>	generateSentencesFromClusters(List<Set<org.aksw.avatar.clustering.Node>> clusters, org.apache.jena.rdf.model.Resource resource, org.semanticweb.owlapi.model.OWLClass OWLClass)
List<simplenlg.framework.NLGElement>	generateSentencesFromClusters(List<Set<org.aksw.avatar.clustering.Node>> clusters, org.apache.jena.rdf.model.Resource resource, org.semanticweb.owlapi.model.OWLClass namedClass, boolean replaceSubjects) Takes the output of the clustering for a given class and a resource.
List<simplenlg.framework.NLGElement>	generateSentencesFromTriples(Set<org.apache.jena.graph.Triple> triples, boolean outgoing, org.aksw.triple2nl.gender.Gender g) Generates sentence for a given set of triples
simplenlg.phrasespec.SPhraseSpec	generateSimplePhraseFromTriple(org.apache.jena.graph.Triple triple) Generates a simple phrase for a triple
simplenlg.phrasespec.SPhraseSpec	generateSimplePhraseFromTriple(org.apache.jena.graph.Triple triple, boolean outgoing) Generates a simple phrase for a triple
List<simplenlg.phrasespec.NPPhraseSpec>	generateSubjects(org.apache.jena.rdf.model.Resource resource, org.semanticweb.owlapi.model.OWLClass resourceType, org.aksw.triple2nl.gender.Gender resourceGender) Returns a list of synonymous expressions as subject for the given resource.
org.aksw.triple2nl.gender.Gender	getGender(org.apache.jena.rdf.model.Resource resource) Returns the gender of the given resource.
private org.semanticweb.owlapi.model.OWLClass	getMostSpecificType(org.semanticweb.owlapi.model.OWLIndividual ind) Returns the most specific type of a given individual.
List<simplenlg.phrasespec.SPhraseSpec>	getPhraseSpecsFromTriples(Set<org.apache.jena.graph.Triple> triples, boolean outgoing) Generates sentence for a given set of triples
private Set<org.apache.jena.graph.Triple>	getSubsetToShow(Set<org.apache.jena.graph.Triple> triples)
Map<org.semanticweb.owlapi.model.OWLIndividual,String>	getSummaries(Set<org.semanticweb.owlapi.model.OWLIndividual> individuals, org.semanticweb.owlapi.model.OWLClass nc, String namespace, double threshold, org.aksw.avatar.dataset.DatasetBasedGraphGenerator.Cooccurrence cooccurrence, org.aksw.avatar.clustering.hardening.HardeningFactory.HardeningType hType) Returns a textual summary of the given entity.
String	getSummary(org.semanticweb.owlapi.model.OWLIndividual individual, org.semanticweb.owlapi.model.OWLClass nc, double threshold, org.aksw.avatar.dataset.DatasetBasedGraphGenerator.Cooccurrence cooccurrence, org.aksw.avatar.clustering.hardening.HardeningFactory.HardeningType hType) Returns a textual summary of the given entity.
Set<org.aksw.avatar.clustering.Node>	getSummaryProperties(org.semanticweb.owlapi.model.OWLClass cls, double threshold, String namespace, org.aksw.avatar.dataset.DatasetBasedGraphGenerator.Cooccurrence cooccurrence)
Collection<org.apache.jena.graph.Triple>	getSummaryTriples(org.semanticweb.owlapi.model.OWLIndividual individual) Returns the triples of the summary for the given individual.
Collection<org.apache.jena.graph.Triple>	getSummaryTriples(org.apache.jena.rdf.model.Resource resource) Returns the triples of the summary for the given resource.
Set<org.apache.jena.graph.Triple>	getTriples(org.apache.jena.rdf.model.Resource r, org.apache.jena.rdf.model.Property p, boolean outgoing) Gets all triples for resource r and property p.
static void	main(String[] args)
String	realize(List<simplenlg.framework.NLGElement> elts)
String	realize(List<Set<org.aksw.avatar.clustering.Node>> properties, org.apache.jena.rdf.model.Resource resource, org.semanticweb.owlapi.model.OWLClass nc) Generates the string representation of a verbalization
protected simplenlg.framework.NLGElement	replaceSubject(simplenlg.framework.NLGElement phrase, List<simplenlg.phrasespec.NPPhraseSpec> subjects, org.aksw.triple2nl.gender.Gender g) Replaces the subject of a coordinated phrase or simple phrase with a subject from a list of precomputed subjects
void	setMaxShownValuesPerProperty(int maxShownValuesPerProperty)
void	setOmitContentInBrackets(boolean omitContentInBrackets)
void	setPersonTypes(Set<String> personTypes)
void	setPropertiesBlacklist(Set<String> blacklist)
private List<org.apache.jena.graph.Triple>	sortByObjectPopularity(Set<org.apache.jena.graph.Triple> triples) Sorts the given triples by the popularity of the triple objects.
String	summarize(org.semanticweb.owlapi.model.OWLIndividual individual) Returns a textual summary of the given entity.
String	summarize(org.semanticweb.owlapi.model.OWLIndividual individual, org.semanticweb.owlapi.model.OWLClass nc) Returns a textual summary of the given entity.
List<simplenlg.framework.NLGElement>	verbalize(org.semanticweb.owlapi.model.OWLIndividual ind, org.semanticweb.owlapi.model.OWLClass nc, String namespace, double threshold, org.aksw.avatar.dataset.DatasetBasedGraphGenerator.Cooccurrence cooccurrence, org.aksw.avatar.clustering.hardening.HardeningFactory.HardeningType hType)
Map<org.semanticweb.owlapi.model.OWLIndividual,List<simplenlg.framework.NLGElement>>	verbalize(Set<org.semanticweb.owlapi.model.OWLIndividual> individuals, org.semanticweb.owlapi.model.OWLClass nc, String namespace, double threshold, org.aksw.avatar.dataset.DatasetBasedGraphGenerator.Cooccurrence cooccurrence, org.aksw.avatar.clustering.hardening.HardeningFactory.HardeningType hType)

Methods inherited from class java.lang.Object

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

Field Detail

logger

private static final org.apache.log4j.Logger logger

DEFAULT_CACHE_BASE_DIR

private static String DEFAULT_CACHE_BASE_DIR

DEFAULT_THRESHOLD

private static final double DEFAULT_THRESHOLD

See Also:

Constant Field Values

DEFAULT_MAX_NUMBER_OF_SHOWN_VALUES_PER_PROPERTY

private static final int DEFAULT_MAX_NUMBER_OF_SHOWN_VALUES_PER_PROPERTY

See Also:

Constant Field Values

DEFAULT_COOCCURRENCE_TYPE

private static final org.aksw.avatar.dataset.DatasetBasedGraphGenerator.Cooccurrence DEFAULT_COOCCURRENCE_TYPE

DEFAULT_HARDENING_TYPE

private static final org.aksw.avatar.clustering.hardening.HardeningFactory.HardeningType DEFAULT_HARDENING_TYPE

nlg

public org.aksw.sparql2nl.naturallanguagegeneration.SimpleNLGwithPostprocessing nlg

endpoint

org.dllearner.kb.sparql.SparqlEndpoint endpoint

language

String language

realiser

protected simplenlg.realiser.english.Realiser realiser

labels

Map<org.apache.jena.rdf.model.Resource,String> labels

litFilter

org.aksw.avatar.rules.NumericLiteralFilter litFilter

gender

org.aksw.triple2nl.gender.TypeAwareGenderDetector gender

resource2Triples

public Map<org.apache.jena.rdf.model.Resource,Collection<org.apache.jena.graph.Triple>> resource2Triples

qef

private org.aksw.jena_sparql_api.core.QueryExecutionFactory qef

cacheDirectory

private String cacheDirectory

pr

org.aksw.avatar.rules.PredicateMergeRule pr

or

org.aksw.avatar.rules.ObjectMergeRule or

sr

org.aksw.avatar.rules.SubjectMergeRule sr

graphGenerator

public org.aksw.avatar.dataset.DatasetBasedGraphGenerator graphGenerator

maxShownValuesPerProperty

int maxShownValuesPerProperty

omitContentInBrackets

boolean omitContentInBrackets

Constructor Detail

Verbalizer

public Verbalizer(org.aksw.jena_sparql_api.core.QueryExecutionFactory qef,
                  String cacheDirectory)

Verbalizer

public Verbalizer(org.dllearner.kb.sparql.SparqlEndpoint endpoint,
                  String cacheDirectory)

Verbalizer

public Verbalizer(org.dllearner.kb.sparql.SparqlEndpoint endpoint)

Method Detail

setPropertiesBlacklist

public void setPropertiesBlacklist(Set<String> blacklist)

Parameters:

blacklist - a blacklist of properties that are omitted when building the summary

setPersonTypes

public void setPersonTypes(Set<String> personTypes)

Parameters:

personTypes - the personTypes to set

setOmitContentInBrackets

public void setOmitContentInBrackets(boolean omitContentInBrackets)

Parameters:

omitContentInBrackets - the omitContentInBrackets to set

getTriples

public Set<org.apache.jena.graph.Triple> getTriples(org.apache.jena.rdf.model.Resource r,
                                                    org.apache.jena.rdf.model.Property p,
                                                    boolean outgoing)

Gets all triples for resource r and property p. If outgoing is true it returns all triples with <r,p,o>, else <s,p,r>

Parameters:

r - the resource

p - the property

outgoing - whether to get outgoing or ingoing triples

Returns:

A set of triples

getSummaryProperties

public Set<org.aksw.avatar.clustering.Node> getSummaryProperties(org.semanticweb.owlapi.model.OWLClass cls,
                                                                 double threshold,
                                                                 String namespace,
                                                                 org.aksw.avatar.dataset.DatasetBasedGraphGenerator.Cooccurrence cooccurrence)

realize

public String realize(List<Set<org.aksw.avatar.clustering.Node>> properties,
                      org.apache.jena.rdf.model.Resource resource,
                      org.semanticweb.owlapi.model.OWLClass nc)

Generates the string representation of a verbalization

Parameters:

properties - List of property clusters to be used for verbalization

resource - Resource to summarize

Returns:

Textual representation

realize

public String realize(List<simplenlg.framework.NLGElement> elts)

generateSentencesFromClusters

public List<simplenlg.framework.NLGElement> generateSentencesFromClusters(List<Set<org.aksw.avatar.clustering.Node>> clusters,
                                                                          org.apache.jena.rdf.model.Resource resource,
                                                                          org.semanticweb.owlapi.model.OWLClass OWLClass)

generateSentencesFromClusters

public List<simplenlg.framework.NLGElement> generateSentencesFromClusters(List<Set<org.aksw.avatar.clustering.Node>> clusters,
                                                                          org.apache.jena.rdf.model.Resource resource,
                                                                          org.semanticweb.owlapi.model.OWLClass namedClass,
                                                                          boolean replaceSubjects)

Takes the output of the clustering for a given class and a resource. Returns the verbalization for the resource

Parameters:

clusters - Output of the clustering

resource - Resource to summarize

Returns:

List of NLGElement

getSubsetToShow

private Set<org.apache.jena.graph.Triple> getSubsetToShow(Set<org.apache.jena.graph.Triple> triples)

sortByObjectPopularity

private List<org.apache.jena.graph.Triple> sortByObjectPopularity(Set<org.apache.jena.graph.Triple> triples)

Sorts the given triples by the popularity of the triple objects.

Parameters:

triples - the triples

Returns:

a list of sorted triples

getSummaryTriples

public Collection<org.apache.jena.graph.Triple> getSummaryTriples(org.apache.jena.rdf.model.Resource resource)

Returns the triples of the summary for the given resource.

Parameters:

resource - the resource of the summary

Returns:

a set of triples

getSummaryTriples

public Collection<org.apache.jena.graph.Triple> getSummaryTriples(org.semanticweb.owlapi.model.OWLIndividual individual)

Returns the triples of the summary for the given individual.

Parameters:

individual - the individual of the summary

Returns:

a set of triples

generateSentencesFromTriples

public List<simplenlg.framework.NLGElement> generateSentencesFromTriples(Set<org.apache.jena.graph.Triple> triples,
                                                                         boolean outgoing,
                                                                         org.aksw.triple2nl.gender.Gender g)

Generates sentence for a given set of triples

Parameters:

triples - A set of triples

Returns:

A set of sentences representing these triples

getPhraseSpecsFromTriples

public List<simplenlg.phrasespec.SPhraseSpec> getPhraseSpecsFromTriples(Set<org.apache.jena.graph.Triple> triples,
                                                                        boolean outgoing)

Generates sentence for a given set of triples

Parameters:

triples - A set of triples

Returns:

A set of sentences representing these triples

applyMergeRules

public List<simplenlg.framework.NLGElement> applyMergeRules(List<simplenlg.phrasespec.SPhraseSpec> triples,
                                                            org.aksw.triple2nl.gender.Gender g)

Generates a set of sentences by merging the sentences in the list as well as possible

Parameters:

triples - List of triples

Returns:

List of sentences

generateSimplePhraseFromTriple

public simplenlg.phrasespec.SPhraseSpec generateSimplePhraseFromTriple(org.apache.jena.graph.Triple triple)

Generates a simple phrase for a triple

Parameters:

triple - A triple

Returns:

A simple phrase

generateSimplePhraseFromTriple

public simplenlg.phrasespec.SPhraseSpec generateSimplePhraseFromTriple(org.apache.jena.graph.Triple triple,
                                                                       boolean outgoing)

Generates a simple phrase for a triple

Parameters:

triple - A triple

Returns:

A simple phrase

verbalize

public List<simplenlg.framework.NLGElement> verbalize(org.semanticweb.owlapi.model.OWLIndividual ind,
                                                      org.semanticweb.owlapi.model.OWLClass nc,
                                                      String namespace,
                                                      double threshold,
                                                      org.aksw.avatar.dataset.DatasetBasedGraphGenerator.Cooccurrence cooccurrence,
                                                      org.aksw.avatar.clustering.hardening.HardeningFactory.HardeningType hType)

verbalize

public Map<org.semanticweb.owlapi.model.OWLIndividual,List<simplenlg.framework.NLGElement>> verbalize(Set<org.semanticweb.owlapi.model.OWLIndividual> individuals,
                                                                                                      org.semanticweb.owlapi.model.OWLClass nc,
                                                                                                      String namespace,
                                                                                                      double threshold,
                                                                                                      org.aksw.avatar.dataset.DatasetBasedGraphGenerator.Cooccurrence cooccurrence,
                                                                                                      org.aksw.avatar.clustering.hardening.HardeningFactory.HardeningType hType)

summarize

public String summarize(org.semanticweb.owlapi.model.OWLIndividual individual)

Returns a textual summary of the given entity.

Returns:

summarize

public String summarize(org.semanticweb.owlapi.model.OWLIndividual individual,
                        org.semanticweb.owlapi.model.OWLClass nc)

Returns a textual summary of the given entity.

Returns:

getSummary

public String getSummary(org.semanticweb.owlapi.model.OWLIndividual individual,
                         org.semanticweb.owlapi.model.OWLClass nc,
                         double threshold,
                         org.aksw.avatar.dataset.DatasetBasedGraphGenerator.Cooccurrence cooccurrence,
                         org.aksw.avatar.clustering.hardening.HardeningFactory.HardeningType hType)

Returns a textual summary of the given entity.

Returns:

getSummaries

public Map<org.semanticweb.owlapi.model.OWLIndividual,String> getSummaries(Set<org.semanticweb.owlapi.model.OWLIndividual> individuals,
                                                                           org.semanticweb.owlapi.model.OWLClass nc,
                                                                           String namespace,
                                                                           double threshold,
                                                                           org.aksw.avatar.dataset.DatasetBasedGraphGenerator.Cooccurrence cooccurrence,
                                                                           org.aksw.avatar.clustering.hardening.HardeningFactory.HardeningType hType)

Returns a textual summary of the given entity.

Returns:

getMostSpecificType

private org.semanticweb.owlapi.model.OWLClass getMostSpecificType(org.semanticweb.owlapi.model.OWLIndividual ind)

Returns the most specific type of a given individual.

Parameters:

ind -

Returns:

generateSubjects

public List<simplenlg.phrasespec.NPPhraseSpec> generateSubjects(org.apache.jena.rdf.model.Resource resource,
                                                                org.semanticweb.owlapi.model.OWLClass resourceType,
                                                                org.aksw.triple2nl.gender.Gender resourceGender)

Returns a list of synonymous expressions as subject for the given resource.

Parameters:

resource - the resource

resourceType - the type of the resource

resourceGender - the gender of the resource

Returns:

list of synonymous expressions

getGender

public org.aksw.triple2nl.gender.Gender getGender(org.apache.jena.rdf.model.Resource resource)

Returns the gender of the given resource.

Parameters:

resource -

Returns:

the gender

setMaxShownValuesPerProperty

public void setMaxShownValuesPerProperty(int maxShownValuesPerProperty)

Parameters:

maxShownValuesPerProperty - the max. number of shown values per properties for multi-value properties

replaceSubject

protected simplenlg.framework.NLGElement replaceSubject(simplenlg.framework.NLGElement phrase,
                                                        List<simplenlg.phrasespec.NPPhraseSpec> subjects,
                                                        org.aksw.triple2nl.gender.Gender g)

Replaces the subject of a coordinated phrase or simple phrase with a subject from a list of precomputed subjects

Parameters:

phrase -

subjects -

Returns:

Phrase with replaced subject

main

public static void main(String[] args)
                 throws IOException

Throws:

IOException