edu.stanford.nlp.patterns.surface

Class SurfacePatternFactory

java.lang.Object

edu.stanford.nlp.patterns.PatternFactory

edu.stanford.nlp.patterns.surface.SurfacePatternFactory

public class SurfacePatternFactory
extends PatternFactory

Created by sonalg on 10/27/14.

Nested Class Summary

Nested Classes

Modifier and Type	Class and Description
static class	SurfacePatternFactory.Genre

Nested classes/interfaces inherited from class edu.stanford.nlp.patterns.PatternFactory

PatternFactory.PatternType

Field Summary

Fields

Modifier and Type	Field and Description
static boolean	addPatWithoutPOS Add patterns without POS restriction as well: One of this and usePOS4Pattern has to be true.
static int	maxWindow4Pattern Consider contexts less than or equal to these many tokens -- total of left and right contexts be can double of this.
static int	minWindow4Pattern Consider contexts longer or equal to these many tokens.
static int	numMinStopWordsToAdd If the whole (either left or right) context is just stop words, add the pattern only if number of tokens is equal or more than this.
static boolean	useContextNERRestriction If the NER tag of the context tokens is not the background symbol, generalize the token with the NER tag
static boolean	useFillerWordsInPat Ignore words like "a", "an", "the" when matching a pattern.
static boolean	useNextContext Consider contexts on the right of a token.
static boolean	usePOS4Pattern Use POS tag restriction in the target term: One of this and addPatWithoutPOS has to be true.
static boolean	usePreviousContext Consider contexts on the left of a token.
static boolean	useTargetParserParentRestriction Adds the parent's tag from the parse tree to the target phrase in the patterns

Fields inherited from class edu.stanford.nlp.patterns.PatternFactory

fillerWords, ignoreWordRegex, numWordsCompound, useLemmaContextTokens, useStopWordsBeforeTerm, useTargetNERRestriction

Constructor Summary

Constructors

Constructor and Description
SurfacePatternFactory()

Method Summary

Modifier and Type	Method and Description
static Set<SurfacePattern>	getContext(List<CoreLabel> sent, int i, Set<CandidatePhrase> stopWords)
static Map<Integer,Set>	getPatternsAroundTokens(DataInstance sent, Set<CandidatePhrase> stopWords)
static boolean	isASCII(String text)
static void	setUp(Properties props)

Methods inherited from class edu.stanford.nlp.patterns.PatternFactory

doNotUse, getPatternsAroundTokens, setUp

Methods inherited from class java.lang.Object

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

Field Detail

usePOS4Pattern

@Execution.Option(name="usePOS4Pattern")
public static boolean usePOS4Pattern

Use POS tag restriction in the target term: One of this and addPatWithoutPOS has to be true.

addPatWithoutPOS

@Execution.Option(name="addPatWithoutPOS")
public static boolean addPatWithoutPOS

Add patterns without POS restriction as well: One of this and usePOS4Pattern has to be true.

minWindow4Pattern

@Execution.Option(name="minWindow4Pattern")
public static int minWindow4Pattern

Consider contexts longer or equal to these many tokens.

maxWindow4Pattern

@Execution.Option(name="maxWindow4Pattern")
public static int maxWindow4Pattern

Consider contexts less than or equal to these many tokens -- total of left and right contexts be can double of this.

usePreviousContext

@Execution.Option(name="usePreviousContext")
public static boolean usePreviousContext

Consider contexts on the left of a token.

useNextContext

@Execution.Option(name="useNextContext")
public static boolean useNextContext

Consider contexts on the right of a token.

numMinStopWordsToAdd

@Execution.Option(name="numMinStopWordsToAdd")
public static int numMinStopWordsToAdd

If the whole (either left or right) context is just stop words, add the pattern only if number of tokens is equal or more than this. This is get patterns like "I am on X" but ignore "on X".

useTargetParserParentRestriction

@Execution.Option(name="useTargetParserParentRestriction")
public static boolean useTargetParserParentRestriction

Adds the parent's tag from the parse tree to the target phrase in the patterns

useContextNERRestriction

@Execution.Option(name="useContextNERRestriction")
public static boolean useContextNERRestriction

If the NER tag of the context tokens is not the background symbol, generalize the token with the NER tag

useFillerWordsInPat

@Execution.Option(name="useFillerWordsInPat")
public static boolean useFillerWordsInPat

Ignore words like "a", "an", "the" when matching a pattern.

Constructor Detail

SurfacePatternFactory

public SurfacePatternFactory()

Method Detail

setUp

public static void setUp(Properties props)

getContext

public static Set<SurfacePattern> getContext(List<CoreLabel> sent,
                                             int i,
                                             Set<CandidatePhrase> stopWords)

isASCII

public static boolean isASCII(String text)

getPatternsAroundTokens

public static Map<Integer,Set> getPatternsAroundTokens(DataInstance sent,
                                                       Set<CandidatePhrase> stopWords)