edu.washington.cs.knowitall.extractor
Class ReVerbRelationExtractor

java.lang.Object
  edu.washington.cs.knowitall.extractor.Extractor<ChunkedSentence,ChunkedBinaryExtraction>
      edu.washington.cs.knowitall.extractor.RelationFirstNpChunkExtractor
          edu.washington.cs.knowitall.extractor.ReVerbRelationExtractor

Direct Known Subclasses:

R2A2, ReVerbExtractor

public abstract class ReVerbRelationExtractor
extends RelationFirstNpChunkExtractor

Field Summary

static String	LONG_RELATION_PATTERN The pattern (V(W*P)?)+
static String	PREP Definition of the "preposition" part of the relation pattern.
static String	SHORT_RELATION_PATTERN The pattern (VP?)+
static String	VERB Definition of the "verb" of the relation pattern.
static String	WORD Definition of the "non-verb/prep" part of the relation pattern.

Fields inherited from class edu.washington.cs.knowitall.extractor.RelationFirstNpChunkExtractor

allowUnary, arg1Extr, arg2Extr, relExtr

Constructor Summary

ReVerbRelationExtractor()
Constructs a new extractor using the default relation pattern, relation mappers, and argument mappers.

ReVerbRelationExtractor(int minFreq, boolean useLexSynConstraints, boolean mergeOverlapRels, boolean allowUnary)
Constructs a new extractor using the default relation pattern, relation mappers, and argument mappers.

Method Summary

Iterable<ChunkedBinaryExtraction>	extractFromHtml(String html) Extracts from the given html using the default sentence reader returned by DefaultObjects#.getDefaultSentenceReaderHtml(java.io.Reader).
Iterable<ChunkedBinaryExtraction>	extractFromString(String text) Extracts from the given text using the default sentence reader returned by DefaultObjects.getDefaultSentenceReader(java.io.Reader).
protected abstract void	initializeArgumentExtractors()
protected void	initializeRelationExtractor() Wrapper for default initialization of the reverb relation extractor.
protected void	initializeRelationExtractor(int minFreq, boolean useLexSynConstraints, boolean mergeOverlapRels, boolean allowUnary) Initialize relation extractor.

Methods inherited from class edu.washington.cs.knowitall.extractor.RelationFirstNpChunkExtractor

extractCandidates, getArgument1Extractor, getArgument2Extractor, getRelationExtractor, setAllowUnary, setArgument1Extractor, setArgument2Extractor, setRelationExtractor

Methods inherited from class edu.washington.cs.knowitall.extractor.Extractor

addMapper, compose, extract, getMappers

Methods inherited from class java.lang.Object

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

Field Detail

VERB

public static final String VERB

Definition of the "verb" of the relation pattern.

See Also:

Constant Field Values

WORD

public static final String WORD

Definition of the "non-verb/prep" part of the relation pattern.

See Also:

Constant Field Values

PREP

public static final String PREP

Definition of the "preposition" part of the relation pattern.

See Also:

Constant Field Values

LONG_RELATION_PATTERN

public static final String LONG_RELATION_PATTERN

The pattern (V(W*P)?)+

SHORT_RELATION_PATTERN

public static final String SHORT_RELATION_PATTERN

The pattern (VP?)+

Constructor Detail

ReVerbRelationExtractor

public ReVerbRelationExtractor()
                        throws ExtractorException

Constructs a new extractor using the default relation pattern, relation mappers, and argument mappers.

Throws:

ExtractorException - if unable to initialize the extractor

ReVerbRelationExtractor

public ReVerbRelationExtractor(int minFreq,
                               boolean useLexSynConstraints,
                               boolean mergeOverlapRels,
                               boolean allowUnary)
                        throws ExtractorException

Constructs a new extractor using the default relation pattern, relation mappers, and argument mappers.

Parameters:

minFreq - - The minimum distinct arguments to be observed in a large collection for the relation to be deemed valid.

useLexSynConstraints - - Use syntactic and lexical constraints that are part of Reverb?

mergeOverlapRels - - Merge overlapping relations?

allowUnary - - Allow relations with one argument to be output.

Throws:

ExtractorException - if unable to initialize the extractor

Method Detail

initializeArgumentExtractors

protected abstract void initializeArgumentExtractors()

initializeRelationExtractor

protected void initializeRelationExtractor()
                                    throws ExtractorException

Wrapper for default initialization of the reverb relation extractor. Use lexical and syntactic constraints, merge overlapping relations,require a minimum of 20 distinct arguments for support, and do not allow unary relations.

Throws:

ExtractorException

initializeRelationExtractor

protected void initializeRelationExtractor(int minFreq,
                                           boolean useLexSynConstraints,
                                           boolean mergeOverlapRels,
                                           boolean allowUnary)
                                    throws ExtractorException

Initialize relation extractor.

Parameters:

minFreq - - The minimum distinct arguments to be observed in a large collection for the relation to be deemed valid.

useLexSynConstraints - - Use syntactic and lexical constraints that are part of Reverb?

mergeOverlapRels - - Merge overlapping relations?

allowUnary - - Allow relations with one argument to be output.

Throws:

ExtractorException - if unable to initialize the extractor

extractFromString

public Iterable<ChunkedBinaryExtraction> extractFromString(String text)
                                                    throws ExtractorException

Extracts from the given text using the default sentence reader returned by DefaultObjects.getDefaultSentenceReader(java.io.Reader).

Parameters:

text -

Returns:

an iterable object over the extractions

Throws:

ExtractorException - if unable to extract

extractFromHtml

public Iterable<ChunkedBinaryExtraction> extractFromHtml(String html)
                                                  throws ExtractorException

Extracts from the given html using the default sentence reader returned by DefaultObjects#.getDefaultSentenceReaderHtml(java.io.Reader).

Parameters:

html -

Returns:

an iterable object over the extractions

Throws:

ExtractorException - if unable to extract