package mpqareader;

import se.lth.cs.nlp.nlputils.annotations.*;

import java.util.*;

public class ExtractSubjSent {

	private static int nclue1, nclue2, nobj;
	
	private static boolean isXMLTag(String s) {
		return s.contains("<") || s.contains(">");
	}
	
	private static void printSentence(String[] tokens,
			int start, int end, String[] cls) {

		if(end - start > 2 && tokens[start].equals("AR") && tokens[start+1].equals(":")) {
			return;
		}

		if(end - start > 2 && tokens[start].equals("EN") && tokens[start+1].equals(":")) {
			start += 2;
		}		

		while(start < end && start < tokens.length && isXMLTag(tokens[start]))
			start++;
		
		if(start >= end)
			return;

		if(cls == null)
			System.out.println("<SUBJ subj=\"false\"/>");
		else {
			System.out.print("<SUBJ subj=\"true\"");
			System.out.print(" posInt=\"" + cls[0] + "\"");
			System.out.print(" neuInt=\"" + cls[1] + "\"");
			System.out.print(" negInt=\"" + cls[2] + "\"");
			System.out.println("/>");
		}
		
		for(int i = start; i < end; i++) {
			String t = tokens[i];
			if(!isXMLTag(t))
				System.out.println(tokens[i]);
		}
		System.out.println();
	}
	
	private static void processText(AnnotatedText text) {
		System.err.println(text.getProperty("mpqa_file"));
		System.out.println("<DOC name=\"" + text.getProperty("mpqa_file") + "\"/>");
		
		AnnotationLayer senLayer = text.layers.get(0);						
		for(Span sen: (AnnotationLayer<Span>) senLayer) {
			
			String[] cls = MPQAToLK.classifySubjSen(sen, text);
			
			if(cls == null)
				nobj++;
			
			printSentence(text.tokens, sen.tokenStart, sen.tokenEnd, cls);
			
		}		
	}
	
	public static void main(String[] argv) {
		Iterator<AnnotatedText> i = MPQAReader.processDirectory(argv[0]);
		while(i.hasNext()) {
			AnnotatedText t = i.next();
			processText(t);
		}
		
		System.err.println(nclue1);
		System.err.println(nclue2);
		System.err.println(nobj);		
	}
	
}