opennlp.tools.coref
Interface Linker

All Known Implementing Classes:

AbstractLinker, DefaultLinker, TreebankLinker

public interface Linker

A linker provides an interface for finding mentions, getMentionFinder, and creating entities out of those mentions, getEntities. This interface also allows for the training of a resolver with the method setEntitites which is used to give the resolver mentions whose entityId fields indicate which mentions refer to the same entity and the train method which compiles all the information provided via calls to setEntities into a model.

Field Summary

static String	COMBINED_NPS String constant used to label a mention which consists of two or more noun phrases.
static String	DESCRIPTOR String constant used to label a mention which is a description.
static String	ISA String constant used to label an mention in an appositive relationship.
static String	NP String constant used to label a mention which consists of a single noun phrase.
static String	PRONOUN_MODIFIER String constant used to label a mention which is a pronoun.
static String	PROPER_NOUN_MODIFIER String constant used to label a mention which is a proper noun modifying another noun.

Method Summary

MentionContext[]	constructMentionContexts(Mention[] mentions) Creates mention contexts for the specified mention exents.
DiscourseEntity[]	getEntities(Mention[] mentions) Returns a list of entities which group the mentions into entity classes.
HeadFinder	getHeadFinder() Returns the head finder associated with this linker.
MentionFinder	getMentionFinder() Returns the mention finder for this linker.
void	setEntities(Mention[] mentions) Indicated that the specified mentions can be used to train this linker.
void	train() Trains the linker based on the data specified via calls to setEntities.

Field Detail

DESCRIPTOR

static final String DESCRIPTOR

String constant used to label a mention which is a description.

See Also:

Constant Field Values

ISA

static final String ISA

String constant used to label an mention in an appositive relationship.

See Also:

Constant Field Values

COMBINED_NPS

static final String COMBINED_NPS

String constant used to label a mention which consists of two or more noun phrases.

See Also:

Constant Field Values

NP

static final String NP

String constant used to label a mention which consists of a single noun phrase.

See Also:

Constant Field Values

PROPER_NOUN_MODIFIER

static final String PROPER_NOUN_MODIFIER

String constant used to label a mention which is a proper noun modifying another noun.

See Also:

Constant Field Values

PRONOUN_MODIFIER

static final String PRONOUN_MODIFIER

String constant used to label a mention which is a pronoun.

See Also:

Constant Field Values

Method Detail

setEntities

void setEntities(Mention[] mentions)

Indicated that the specified mentions can be used to train this linker. This requires that the coreference relationship between the mentions have been labeled in the mention's id field.

Parameters:

mentions - The mentions to be used to train the linker.

getEntities

DiscourseEntity[] getEntities(Mention[] mentions)

Returns a list of entities which group the mentions into entity classes.

Parameters:

mentions - A array of mentions.

Returns:

An array of discourse entities.

constructMentionContexts

MentionContext[] constructMentionContexts(Mention[] mentions)

Creates mention contexts for the specified mention exents. These are used to compute coreference features over.

Parameters:

mentions - The mention of a document.

Returns:

mention contexts for the specified mention exents.

train

void train()
           throws IOException

Trains the linker based on the data specified via calls to setEntities.

Throws:

IOException

getMentionFinder

MentionFinder getMentionFinder()

Returns the mention finder for this linker. This can be used to get the mentions of a Parse.

Returns:

The object which finds mentions for this linker.

getHeadFinder

HeadFinder getHeadFinder()

Returns the head finder associated with this linker.

Returns:

The head finder associated with this linker.