edu.stanford.nlp.trees.tregex
Class TreeMatcher

java.lang.Object
  edu.stanford.nlp.trees.tregex.TreeMatcher

public class TreeMatcher
extends Object

Matches a TreePattern against a Tree. For programmatic usage, the important methods are:

• reset() resets the TreeMatcher

• find() attempts to find the next match of the TreePattern to the Tree.

• matches(edu.stanford.nlp.trees.Tree) determines whether the TreePattern matches the Tree at the root of the Tree.

• getMatch() once a match has been found, returns the Tree node corresponding to the root of the TreePattern.

• getNode(java.lang.Object) once a match has been found, allows you to use the name given to a node description in the TreePattern to retrieve the node matched to that node description

For running it from the command line, see the main method for a description of command line arguments and the TreePattern class for a description of supported tree patterns.

Author:

Roger Levy

Nested Class Summary

static class

TreeMatcher.TRegexTreeReaderFactory

Method Summary

boolean	find() Attempts to find the next match, if any.
Tree	getMatch() Returns the root node currently matched by the TreePattern
Tree	getNode(Object name) Returns the relevant node in the pattern by the name it was given in the TreePattern
static void	main(String[] args) Use to match a tree pattern to the trees in files.
boolean	matches(Tree t) Attempts to match the input TreePattern against the specified Tree, at the supplied Tree node t.
TreeMatcher	reset() Resets the TreeMatcher.

Methods inherited from class java.lang.Object

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

Method Detail

matches

public boolean matches(Tree t)

Attempts to match the input TreePattern against the specified Tree, at the supplied Tree node t. Stops at the first full match, if any.

At the moment, matches() is fairly inefficiently implemented: for cases where the root tree-pattern node description matches t but the rest of the pattern does not match given that assignment, it actually traverses the entire tree and the entire pattern to look for a deeper complete match before rejecting any that it may find. Hopefully this will be fixed in the future.

Parameters:

t - the supplied tree node for a match

reset

public TreeMatcher reset()

Resets the TreeMatcher.

find

public boolean find()

Attempts to find the next match, if any.

getMatch

public Tree getMatch()

Returns the root node currently matched by the TreePattern

getNode

public Tree getNode(Object name)

Returns the relevant node in the pattern by the name it was given in the TreePattern

main

public static void main(String[] args)

Use to match a tree pattern to the trees in files. Usage:

java edu.stanford.nlp.trees.tregex.TreeMatcher [-T] [-C] [-w] [-f] [-tn TN] pattern [handle] filepath

It prints out all the matches of the tree pattern to every tree in the treebank rooted at the filepath.

Parameters:

args - Command line arguments: Argument 1 is the tree pattern which might name a node with =name (for some arbitrary string "name"), argument 2 is an optional name =name, and argument 3 is a filepath to files with trees. A -T flag causes all trees to be printed as processed. Otherwise just matches are printed. The -C flag suppresses printing of matches, so only a number of matches is printed. The -w flag causes the whole of a tree that matches to be printed. The -f flag prints the filename of the file containing each match. The -tn flag is followed by the name of a TreeNormalizer class, which can be used to normalize trees before they are matched (e.g., to strip functional tags)