package mpqareader;

import java.io.BufferedReader;
import java.io.FileReader;
import java.io.IOException;
import java.util.ArrayList;

import se.lth.cs.nlp.nlputils.core.IntHistogram;

public class TestPosition {

	private static ArrayList<String> readSentence(BufferedReader br) throws IOException {
		ArrayList<String> out = new ArrayList();
		String line = br.readLine().trim();
		while(!line.isEmpty()) {
			out.add(line);
			line = br.readLine().trim();			
		}
		return out;
	}
	
	public static void main(String[] argv) {

		String fileName = argv[0];

		IntHistogram subjHist = new IntHistogram();
		IntHistogram allHist = new IntHistogram();
		
		try {
			BufferedReader br = new BufferedReader(new FileReader(fileName));

			int pos = 0;
			
			while(true) {
				String line = br.readLine();
				if(line == null)
					break;
				if(line.startsWith("<DOC")) {
					line = br.readLine();
					pos = 0;
				}
				if(!line.startsWith("<SUBJ"))
					throw new RuntimeException("bug");
				pos++;
				boolean isSubjective = line.equals("<SUBJ=\"true\"/>");
				readSentence(br);
				
				if(isSubjective)
					subjHist.add(pos);
				allHist.add(pos);
				
			}
			
			//System.out.println(subjHist.asSortedList());
			//System.out.println(allHist.asSortedList());
			
			for(int i = 0; i < 100; i++) {
				double p = (double) subjHist.getFrequency(i) / allHist.getFrequency(i);
				System.out.println(i + ". " + p);
			}
			
		} catch(Exception e) {
			e.printStackTrace();
		}
	}
}