//WordFreqTest.java
//read File of strings.  determine fequency of each word.  Similar to case study of chapter 8.
//Usage: java WordFreqTest inputFileName   >redirected_output_file
//all-in-in file includes class WordFreq (string and int counter).
// TreeSet used to hold WordFreq objects

import java.io.*;
import java.util.*;

public class WordFreqTest {

    public static void main( String[] args ) {

	String fileName=null;
	BufferedReader wordFile;
	TreeSet wordFreqs = new TreeSet();
	WordFreq nextWordFreq;
	String line, word;
	StringTokenizer tokens;
	int totalWords=0;
	if (args.length > 0)   //there is a command line argument
	    fileName = args[0];
	else {
	    System.out.println("Usage: java WordFreq inputFileName");
	    System.exit(1);
	}
	    

	try {
	    wordFile = new BufferedReader( new FileReader(fileName));
	    
	    line = wordFile.readLine();
	    while (line != null) {
		//tokens = new StringTokenizer(line); //space separated
		//strip all punctuation:
		tokens = new StringTokenizer(line," \t\r\n\\\"!@#$%^&*()-_=+[]{}';:|/?.>,<`~");
		while (tokens.hasMoreTokens()) {
		    word = tokens.nextToken().toLowerCase();
		    totalWords++;
		    nextWordFreq = new WordFreq(word);
		    //add word if not already in tree.  
		    if (!wordFreqs.add(nextWordFreq)) {
			//if already in tree, retrieve view of it as subset of one, 
			//access it as the first element and increment it.
			//???there's no other way to access an object in a TreeSet??
			//the subSet returns from arg1 to arg2 not including arg2,
			//not including either if arg1=arg2!  so make arg2 a bogus
			//wordfreq of same string but add nul char to it.

			SortedSet subOneWordFreq = 
			    wordFreqs.subSet(nextWordFreq,new WordFreq(nextWordFreq.wordIs()+"\0"));
			((WordFreq)(subOneWordFreq.first())).another();
		    }
		}
		line = wordFile.readLine();
	    }
	    wordFile.close();

	}
	catch (IOException e) {
	    System.out.println( "threw an exception...File does not exist?" );
	}

	Iterator it = wordFreqs.iterator();
	while (it.hasNext())
	    System.out.println(it.next());    
	System.out.println("total words: "+totalWords+"   Different words: "+
			   wordFreqs.size());
    }
}




//should be in a public class in its own file...
//***needs to say implements Comparable so that works with TreeSet ??
class WordFreq implements Comparable {
    private String word;
    private int freq;

    public WordFreq(String newWord) {
	word = newWord;
	freq = 1;
    }

    public String wordIs() {
	return word;
    }

    public int freqIs() {
	return freq;
    }

    public String toString() {
	return word + " " + freq;
    }

    public void another() {
	freq++;
    }

    //*** TreeSet uses compareTo of objects to determine equality and relative order.
    //argument must be Object, so cast it to this class and use String's compareTo
    //to compare the String members
    public int compareTo(Object otherWordFreq) {
	return this.word.compareTo(((WordFreq)otherWordFreq).word);
    }


}