edu.stanford.nlp.ie.util

Class RelationTriple

java.lang.Object

edu.stanford.nlp.ie.util.RelationTriple

All Implemented Interfaces:

java.lang.Comparable<RelationTriple>, java.lang.Iterable<CoreLabel>

Direct Known Subclasses:

RelationTriple.AsKBEntry, RelationTriple.WithTree

public class RelationTriple
extends java.lang.Object
implements java.lang.Comparable<RelationTriple>, java.lang.Iterable<CoreLabel>

A (subject, relation, object) triple; e.g., as used in the KBP challenges or in OpenIE systems.

Author:

Gabor Angeli

Nested Class Summary

Nested Classes

Modifier and Type	Class and Description
static class	RelationTriple.AsKBEntry A RelationTriple, optimized for tasks such as KBP where we care about named entities, and care about things like provenance and coref.
static class	RelationTriple.WithTree A RelationTriple, but with the tree saved as well.

Field Summary

Fields

Modifier and Type	Field and Description
double	confidence An optional score (confidence) for this triple
java.util.List<CoreLabel>	object The object (third argument) of this triple
java.util.List<CoreLabel>	relation The relation (second argument) of this triple
java.util.List<CoreLabel>	subject The subject (first argument) of this triple
static java.util.Set<java.lang.String>	VALID_ADVERB_ARCS A set of valid arcs denoting an entity we are interested in
static java.util.Set<java.lang.String>	VALID_OBJECT_ARCS A set of valid arcs denoting an object entity we are interested in
static java.util.Set<java.lang.String>	VALID_SUBJECT_ARCS A set of valid arcs denoting a subject entity we are interested in

Constructor Summary

Constructors

Constructor and Description
RelationTriple(java.util.List<CoreLabel> subject, java.util.List<CoreLabel> relation, java.util.List<CoreLabel> object)
RelationTriple(java.util.List<CoreLabel> subject, java.util.List<CoreLabel> relation, java.util.List<CoreLabel> object, double confidence) Create a new triple with known values for the subject, relation, and object.

Method Summary

Modifier and Type	Method and Description
java.util.List<CoreLabel>	allTokens() Returns all the tokens in the extraction, in the order subject then relation then object.
java.util.Optional<SemanticGraph>	asDependencyTree() An optional method, returning the dependency tree this triple was extracted from
java.util.List<CoreLabel>	asSentence() Return the given relation triple as a flat sentence
int	compareTo(RelationTriple o)
java.lang.String	confidenceGloss() A textual representation of the confidence.
boolean	equals(java.lang.Object o)
static java.util.List<RelationTriple>	extract(SemanticGraph parse, java.util.List<CoreLabel> tokens) Extract the nominal patterns from this sentence.
protected Pair<java.lang.Integer,java.lang.Integer>	getSpan(java.util.List<CoreLabel> tokens, java.util.function.Function<CoreLabel,java.lang.Integer> toMin, java.util.function.Function<CoreLabel,java.lang.Integer> toMax)
int	hashCode()
java.util.Iterator<CoreLabel>	iterator()
java.lang.String	objectGloss() The object of this relation triple, as a String
CoreLabel	objectHead() The head of the object of this relation triple.
java.lang.String	objectLemmaGloss() The object of this relation triple, as a String of the object's lemmas.
Pair<java.lang.Integer,java.lang.Integer>	objectTokenSpan()
static java.util.Optional<RelationTriple>	optimizeForKB(RelationTriple input, java.util.Optional<CoreMap> sentence, java.util.Map<CoreLabel,java.util.List<CoreLabel>> canonicalMentions)
java.lang.String	relationGloss() The relation of this relation triple, as a String
java.lang.String	relationLemmaGloss() The relation of this relation triple, as a String of the relation's lemmas.
static java.util.Optional<RelationTriple>	segment(SemanticGraph parse, java.util.Optional<java.lang.Double> confidence) Segment the given parse tree, forcing all nodes to be consumed.
static java.util.Optional<RelationTriple>	segment(SemanticGraph parse, java.util.Optional<java.lang.Double> confidence, boolean consumeAll) Try to segment this sentence as a relation triple.
java.lang.String	subjectGloss() The subject of this relation triple, as a String
CoreLabel	subjectHead() The head of the subject of this relation triple.
java.lang.String	subjectLemmaGloss() The subject of this relation triple, as a String of the subject's lemmas.
Pair<java.lang.Integer,java.lang.Integer>	subjectTokenSpan()
java.lang.String	toString() Print a human-readable description of this relation triple, as a tab-separated line

Methods inherited from class java.lang.Object

clone, finalize, getClass, notify, notifyAll, wait, wait, wait

Methods inherited from interface java.lang.Iterable

forEach, spliterator

Field Detail

subject

public final java.util.List<CoreLabel> subject

The subject (first argument) of this triple

relation

public final java.util.List<CoreLabel> relation

The relation (second argument) of this triple

object

public final java.util.List<CoreLabel> object

The object (third argument) of this triple

confidence

public final double confidence

An optional score (confidence) for this triple

VALID_SUBJECT_ARCS

public static final java.util.Set<java.lang.String> VALID_SUBJECT_ARCS

A set of valid arcs denoting a subject entity we are interested in

VALID_OBJECT_ARCS

public static final java.util.Set<java.lang.String> VALID_OBJECT_ARCS

A set of valid arcs denoting an object entity we are interested in

VALID_ADVERB_ARCS

public static final java.util.Set<java.lang.String> VALID_ADVERB_ARCS

A set of valid arcs denoting an entity we are interested in

Constructor Detail

RelationTriple

public RelationTriple(java.util.List<CoreLabel> subject,
                      java.util.List<CoreLabel> relation,
                      java.util.List<CoreLabel> object,
                      double confidence)

Create a new triple with known values for the subject, relation, and object. For example, "(cats, play with, yarn)"

Parameters:

subject - The subject of this triple; e.g., "cats".

relation - The relation of this triple; e.g., "play with".

object - The object of this triple; e.g., "yarn".

RelationTriple

public RelationTriple(java.util.List<CoreLabel> subject,
                      java.util.List<CoreLabel> relation,
                      java.util.List<CoreLabel> object)

See Also:

RelationTriple(java.util.List, java.util.List, java.util.List, double)

Method Detail

allTokens

public java.util.List<CoreLabel> allTokens()

Returns all the tokens in the extraction, in the order subject then relation then object.

subjectGloss

public java.lang.String subjectGloss()

The subject of this relation triple, as a String

subjectHead

public CoreLabel subjectHead()

The head of the subject of this relation triple.

subjectLemmaGloss

public java.lang.String subjectLemmaGloss()

The subject of this relation triple, as a String of the subject's lemmas. This method will additionally strip out punctuation as well.

objectGloss

public java.lang.String objectGloss()

The object of this relation triple, as a String

objectHead

public CoreLabel objectHead()

The head of the object of this relation triple.

objectLemmaGloss

public java.lang.String objectLemmaGloss()

The object of this relation triple, as a String of the object's lemmas. This method will additionally strip out punctuation as well.

relationGloss

public java.lang.String relationGloss()

The relation of this relation triple, as a String

relationLemmaGloss

public java.lang.String relationLemmaGloss()

The relation of this relation triple, as a String of the relation's lemmas. This method will additionally strip out punctuation as well, and lower-cases the relation.

confidenceGloss

public java.lang.String confidenceGloss()

A textual representation of the confidence.

getSpan

protected Pair<java.lang.Integer,java.lang.Integer> getSpan(java.util.List<CoreLabel> tokens,
                                                            java.util.function.Function<CoreLabel,java.lang.Integer> toMin,
                                                            java.util.function.Function<CoreLabel,java.lang.Integer> toMax)

subjectTokenSpan

public Pair<java.lang.Integer,java.lang.Integer> subjectTokenSpan()

objectTokenSpan

public Pair<java.lang.Integer,java.lang.Integer> objectTokenSpan()

asDependencyTree

public java.util.Optional<SemanticGraph> asDependencyTree()

An optional method, returning the dependency tree this triple was extracted from

asSentence

public java.util.List<CoreLabel> asSentence()

Return the given relation triple as a flat sentence

equals

public boolean equals(java.lang.Object o)

Overrides:

equals in class java.lang.Object

hashCode

public int hashCode()

Overrides:

hashCode in class java.lang.Object

toString

public java.lang.String toString()

Print a human-readable description of this relation triple, as a tab-separated line

Overrides:

toString in class java.lang.Object

compareTo

public int compareTo(RelationTriple o)

Specified by:

compareTo in interface java.lang.Comparable<RelationTriple>

iterator

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

Specified by:

iterator in interface java.lang.Iterable<CoreLabel>

extract

public static java.util.List<RelationTriple> extract(SemanticGraph parse,
                                                     java.util.List<CoreLabel> tokens)

Extract the nominal patterns from this sentence.

Parameters:

parse - The parse tree of the sentence to annotate.

tokens - The tokens of the sentence to annotate.

Returns:

A list of RelationTriples. Note that these do not have an associated tree with them.

See Also:

NOUN_TOKEN_PATTERNS, NOUN_DEPENDENCY_PATTERNS

segment

public static java.util.Optional<RelationTriple> segment(SemanticGraph parse,
                                                         java.util.Optional<java.lang.Double> confidence,
                                                         boolean consumeAll)

Try to segment this sentence as a relation triple. This sentence must already match one of a few strict patterns for a valid OpenIE extraction. If it does not, then no relation triple is created. That is, this is not a relation extractor; it is just a utility to segment what is already a (subject, relation, object) triple into these three parts.

Parameters:

parse - The sentence to process, as a dependency tree.

confidence - An optional confidence to pass on to the relation triple.

consumeAll - if true, force the entire parse to be consumed by the pattern.

Returns:

A relation triple, if this sentence matches one of the patterns of a valid relation triple.

segment

public static java.util.Optional<RelationTriple> segment(SemanticGraph parse,
                                                         java.util.Optional<java.lang.Double> confidence)

Segment the given parse tree, forcing all nodes to be consumed.

See Also:

segment(edu.stanford.nlp.semgraph.SemanticGraph, Optional)

optimizeForKB

public static java.util.Optional<RelationTriple> optimizeForKB(RelationTriple input,
                                                               java.util.Optional<CoreMap> sentence,
                                                               java.util.Map<CoreLabel,java.util.List<CoreLabel>> canonicalMentions)