WordStatistics.java

// WordStatistics.java
import java.util.*;
import java.io.*;
import java.util.regex.*;

enum Pronoun
{
	I ("I"),
	YOU ("you"),
	SHE ("she"),
	HE ("he"),
	THEY ("they"),
	WE ("we");
	
	String s;
	
	Pronoun(String s)
	{
		this.s = s;
	}
	String getPronoun()
	{ return s; }

}
public class WordStatistics
{
	String fileName;
	StringBuilder text = new StringBuilder();
	Map<String, Integer> wordMap = new HashMap<String, Integer>();
	List<Word> wordList = new ArrayList<Word>();
	
	public static void main(String[] args)
	{	
		
        if(args.length < 1)
        {
        	System.out.println("Error: needs at least one argument. \n Usage: java WordStatistics textFileName\n")
        }

		WordStatistics ws = new WordStatistics();
		ws.fileName = args[0];
		
		try
		{
			
			ws.read();
			ws.addWords();
			ws.printSorted();
		}
		catch (IOException e) {e.printStackTrace(); }
		
	}
	void printSorted()
	{
		Collections.sort(wordList);
		
		// Print only if word repeats more than five times
		for(Word w: wordList)
		{
			if(w.getReps() >= 5)
			System.out.println("[" + w.getWord() + "] : [" + w.getReps() + "]");
		}
	}

	boolean checkIfPronoun(String word)
	{
		boolean foundPronoun = false;
		for (Pronoun p : Pronoun.values())
        {
			if(p.getPronoun().equals(word))
				return true;
		}
		return false;
	}
	// Retrieves and adds words to wordMap
	void addWords()
	{
		Pattern p = Pattern.compile("[\\w&&\\D]{3,}");
		Matcher m = p.matcher(text);
		
		// Look for the next word that matches the pattern
		while(m.find())
		{
			// Get the word from the last search
			String word = m.group().toLowerCase();
			
			// Check whether word is a pronoun
			if(!checkIfPronoun(word))
			{
				// Check the key value (repetitions) for current word
				Integer val = wordMap.get(word);
			
				// Value is null, add new word to map
				if(val == null)
				{
					wordMap.put(word, 1);
				}
				// Increase the number of repetitions
				else
				{
					val++;
					wordMap.put(word, val);
				}
			}
		}
		// Add words to list
		for (Map.Entry<String, Integer> entry : wordMap.entrySet()) 
			{
				String w = entry.getKey();
				Integer r = entry.getValue();
				
				wordList.add(new Word(w, r));

			}
		
	}
	void read() throws IOException 
	{
		log("Reading from file.");
		String NL = System.getProperty("line.separator");
		Scanner scanner = new Scanner(new FileInputStream(fileName), "utf-8");
		try {
		while (scanner.hasNextLine()){
			text.append(scanner.nextLine() + NL);
		}
		}
		finally{
		scanner.close();
	}
  }
  private void log(String aMessage)
  {
    System.out.println(aMessage);
  }
	
}