edu.stanford.nlp.ling.tokensregex

Class MultiPatternMatcher<T>

java.lang.Object

edu.stanford.nlp.ling.tokensregex.MultiPatternMatcher<T>

public class MultiPatternMatcher<T>
extends Object

Matcher that takes in multiple patterns

Author:

Angel Chang

Nested Class Summary

Nested Classes

Modifier and Type	Class and Description
static class	MultiPatternMatcher.BasicSequencePatternTrigger<T> Simple SequencePatternTrigger that looks at each node, and identifies which patterns may potentially match each node, and then aggregates (union) all these patterns together.
static interface	MultiPatternMatcher.NodePatternTrigger<T> A function which returns a collections of patterns that may match when given a single node from a larger sequence.
static interface	MultiPatternMatcher.SequencePatternTrigger<T> A function which returns a collections of patterns that may match when a sequence of nodes.

Constructor Summary

Constructors

Constructor and Description
MultiPatternMatcher(Collection<SequencePattern<T>> patterns)
MultiPatternMatcher(MultiPatternMatcher.SequencePatternTrigger<T> patternTrigger, Collection<? extends SequencePattern<T>> patterns)
MultiPatternMatcher(MultiPatternMatcher.SequencePatternTrigger<T> patternTrigger, SequencePattern<T>... patterns)
MultiPatternMatcher(SequencePattern<T>... patterns)

Method Summary

Modifier and Type	Method and Description
List<SequenceMatchResult<T>>	find(List<? extends T> elements, SequenceMatcher.FindType findType) Given a sequence, applies our patterns over the sequence and returns all matches, depending on the findType.
Iterable<SequenceMatchResult<T>>	findAllNonOverlappingMatchesPerPattern(List<? extends T> elements) Given a sequence, applies each of our patterns over the sequence and returns all non overlapping matches for each of the patterns.
List<SequenceMatchResult<T>>	findNonOverlapping(List<? extends T> elements) Given a sequence, applies our patterns over the sequence and returns all non overlapping matches.
List<SequenceMatchResult<T>>	findNonOverlapping(List<? extends T> elements, Comparator<? super SequenceMatchResult> cmp) Given a sequence, applies our patterns over the sequence and returns all non overlapping matches.
List<SequenceMatchResult<T>>	findNonOverlappingMaxScore(List<? extends T> elements) Given a sequence, applies our patterns over the sequence and returns all non overlapping matches.
List<SequenceMatchResult<T>>	findNonOverlappingMaxScore(List<? extends T> elements, java.util.function.Function<? super SequenceMatchResult,Double> scorer) Given a sequence, applies our patterns over the sequence and returns all non overlapping matches.
Collection<SequencePattern<T>>	getTriggeredPatterns(List<? extends T> elements) Given a sequence, return the collection of patterns that are triggered by the sequence (these patterns are the ones that may potentially match a subsequence in the sequence)

Methods inherited from class java.lang.Object

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

Constructor Detail

MultiPatternMatcher

public MultiPatternMatcher(MultiPatternMatcher.SequencePatternTrigger<T> patternTrigger,
                           Collection<? extends SequencePattern<T>> patterns)

MultiPatternMatcher

public MultiPatternMatcher(MultiPatternMatcher.SequencePatternTrigger<T> patternTrigger,
                           SequencePattern<T>... patterns)

MultiPatternMatcher

public MultiPatternMatcher(Collection<SequencePattern<T>> patterns)

MultiPatternMatcher

public MultiPatternMatcher(SequencePattern<T>... patterns)

Method Detail

findNonOverlapping

public List<SequenceMatchResult<T>> findNonOverlapping(List<? extends T> elements)

Given a sequence, applies our patterns over the sequence and returns all non overlapping matches. When multiple patterns overlaps, matched patterns are selected by the highest priority/score is selected, then the longest pattern, then the starting offset, then the original order

Parameters:

elements - input sequence to match against

Returns:

list of match results that are non-overlapping

findNonOverlapping

public List<SequenceMatchResult<T>> findNonOverlapping(List<? extends T> elements,
                                                       Comparator<? super SequenceMatchResult> cmp)

Given a sequence, applies our patterns over the sequence and returns all non overlapping matches. When multiple patterns overlaps, matched patterns are selected by order specified by the comparator

Parameters:

elements - input sequence to match against

cmp - comparator indicating order that overlapped sequences should be selected.

Returns:

list of match results that are non-overlapping

find

public List<SequenceMatchResult<T>> find(List<? extends T> elements,
                                         SequenceMatcher.FindType findType)

Given a sequence, applies our patterns over the sequence and returns all matches, depending on the findType. When multiple patterns overlaps, matched patterns are selected by order specified by the comparator

Parameters:

elements - input sequence to match against

findType - whether FindType.FIND_ALL or FindType.FIND_NONOVERLAPPING

Returns:

list of match results

findNonOverlappingMaxScore

public List<SequenceMatchResult<T>> findNonOverlappingMaxScore(List<? extends T> elements)

Given a sequence, applies our patterns over the sequence and returns all non overlapping matches. When multiple patterns overlaps, matched patterns are selected to give the overall maximum score

Parameters:

elements - input sequence to match against

Returns:

list of match results that are non-overlapping

findNonOverlappingMaxScore

public List<SequenceMatchResult<T>> findNonOverlappingMaxScore(List<? extends T> elements,
                                                               java.util.function.Function<? super SequenceMatchResult,Double> scorer)

Given a sequence, applies our patterns over the sequence and returns all non overlapping matches. When multiple patterns overlaps, matched patterns are selected to give the overall maximum score

Parameters:

elements - input sequence to match against

scorer - scorer for scoring each match

Returns:

list of match results that are non-overlapping

findAllNonOverlappingMatchesPerPattern

public Iterable<SequenceMatchResult<T>> findAllNonOverlappingMatchesPerPattern(List<? extends T> elements)

Given a sequence, applies each of our patterns over the sequence and returns all non overlapping matches for each of the patterns. Unlike #findAllNonOverlapping, overlapping matches from different patterns are kept

Parameters:

elements - input sequence to match against

Returns:

iterable of match results that are non-overlapping

getTriggeredPatterns

public Collection<SequencePattern<T>> getTriggeredPatterns(List<? extends T> elements)

Given a sequence, return the collection of patterns that are triggered by the sequence (these patterns are the ones that may potentially match a subsequence in the sequence)

Parameters:

elements - Input sequence

Returns:

Collection of triggered patterns