resources.jape.sse.jape Maven / Gradle / Ivy

Go to download
Show more of this group Show more artifacts with this name
Show all versions of linguistic-simplifier Show documentation
Linguistic based techniques for text simplification
The newest version!
/*
 * sse.jape
 *
 * Copyright (c) 2004-2013, The University of Sheffield.
 * 
 * This file is part of GATE (see http://gate.ac.uk/), and is free software,
 * Licensed under the GNU Library General Public License, Version 3, June 2007
 * (in the distribution as file licence.html, and also available at
 * http://gate.ac.uk/gate/licence.html).
 * 
 * Original version developed by Mark A. Greenwood as part of his PhD thesis.
 *
 * This grammar implements some of the ideas found in:
 * http://owl.english.purdue.edu/owl/resource/572/1/
 *
 * In the examples redundant is  while replace is 
 */

Phase: Redundant
Input: Token NounChunk
Options: control = appelt

//Tries to implement 4. Avoid Overusing Expletives at the Beginning of Sentences
//Examlpe sentences include
//   It is the governor who signs or vetoes bills.
//   There are four rules that should be observed.
//   There was a big explosion, which shook the windows, and people ran into the street. 
Rule: SSE1
(
	(({Token.string == "It"}|{Token.string == "There"})
	{Token.root == "be"}):r1
	(({Token})?
	({Token})?
	({Token})?
	({Token})?
	({Token})?
	({Token})?):nouns
	(({Token.string == ","})?
	({Token.string == "that"}|{Token.string == "which"}|{Token.string == "who"})):r2
):sse
-->
{
	// Get the tokens we used to spot this possible gerund cluase
	AnnotationSet sse = bindings.get("sse");
	
	AnnotationSet nt = bindings.get("nouns");
	AnnotationSet nouns = inputAS.get(nt.firstNode().getOffset(),nt.lastNode().getOffset());
	
	if (!nouns.getAllTypes().contains("NounChunk")) return;
	
	Iterator nit = nouns.iterator();
	while (nit.hasNext())
	{
		gate.Annotation na = (gate.Annotation)nit.next();
		
		if (na.getType().equals("Token"))
		{
			String cna = (String)na.getFeatures().get("category");
			
			if (cna.startsWith("W") || cna.startsWith("V")) return;
		}
	}
	
	AnnotationSet r1 = bindings.get("r1");
	AnnotationSet r2 = bindings.get("r2");
	
	
	// get the start and end of this section
	Node start = sse.firstNode();
	Node end = sse.lastNode();
	
	// Get any overlapping sentences
	AnnotationSet sentences = inputAS.get("Sentence",start.getOffset(),end.getOffset());
	
	// if there aren't any sentences or more than 1 do nothing
	if (sentences == null || sentences.size() != 1) return;

	// Get the single overlapping sentence
	Annotation sentence = sentences.iterator().next();

	// Get all the tokens in the sentence that appear after the tokens we matched	
	AnnotationSet ts = inputAS.get("Token",end.getOffset(),sentences.lastNode().getOffset());

	// Put these tokens in a list and sort them into the correct order
	List tokens = new ArrayList();
	tokens.addAll(ts);
	Collections.sort(tokens);

	// Get the first token we matched against
	Annotation token = (Annotation)tokens.get(0);
	
	// If the first token is at the beginning of the
	// sentence then this something we want to change
	if (!start.getOffset().equals(sentence.getStartNode().getOffset())) return;

	String matchID = Gate.genSym();

	// Create a feature map and store the rule name
	// within it
	FeatureMap params = Factory.newFeatureMap();
	params.put("rule","SSE1");
	params.put("part","be-verb");
	params.put("matchID",matchID);
	
	outputAS.add(r1.firstNode(),r1.lastNode(),"Redundant",params);

	params = Factory.newFeatureMap();
	params.put("rule","SSE1");
	params.put("part","pronoun");
	params.put("matchID",matchID);
	outputAS.add(r2.firstNode(),r2.lastNode(),"Redundant",params);
}