/*

NOTE: This doesn't work perfectly, but it's pretty darn close.
      Feel free to debug it and send me your results.
      I would be grateful and heap praise upon you.

      Adam Shaw, 5/16/06, U Chicago: CS 102

*/

package scraper;

import java.io.BufferedReader;
import java.io.IOException;
import java.io.InputStream;
import java.io.InputStreamReader;
import java.net.URL;
import java.util.ArrayList;
import java.util.List;

public class PageGrabber {

    private String webaddr;
	    	
    public PageGrabber(String a) {
	this.webaddr = a;
    }
			    	
    public List<String> getContents() throws IOException {

	List<String> ss = new ArrayList<String>();
							
	URL url = new URL(webaddr);
	InputStream is = url.openStream();
	InputStreamReader isr = new InputStreamReader(is);
	BufferedReader br = new BufferedReader(isr);
													
	while (br.ready()) {
	    String s = br.readLine();
	    ss.add(s);
	}

	br.close();
	isr.close();
	is.close();

	return ss;		
    }
}

/* ************************* */

package scraper;

import java.io.FileWriter;
import java.io.IOException;
import java.io.PrintWriter;
import java.util.List;

public class NBARosterScraper {

    private String teamname;
    private String webaddr;
    private PageGrabber g;

    public NBARosterScraper(String t) {
	this.teamname = t;
	this.webaddr = "http://www.nba.com/" + teamname + "/roster/index.html";
	g = new PageGrabber(webaddr);
    }

    private boolean isPlayerBlock(String s) {
	String t = s.trim();
	if (t.startsWith("<td class=\"gSGRowOdd"))
	    return true;
	if (t.startsWith("<td class=\"gSGRowEven"))
	    return true;
	return false;
    }
    
    private static String discardTags(String h) {
	h = h.trim();
	if (h.charAt(0) != '<')
	    throw new IllegalArgumentException("I can't handle that.");
	int firstGT = h.indexOf(">");
	int secondLT = h.indexOf("<",firstGT);
	String contents = h.substring(firstGT+1,secondLT);
	return contents.trim();
    }

    private static String dropLastTag(String h) {
	h = h.trim();
	int firstLT = h.indexOf("<");
	String s = h.substring(0,firstLT);
	return s.trim();
    }

    private static int inches(String ht) {
	int firstHyphen = ht.indexOf('-');
	String ft = ht.substring(0, firstHyphen);
	String in = ht.substring(firstHyphen+1);
	int i = Integer.parseInt(ft) * 12 + Integer.parseInt(in);
	return i;
    }

    // takes a date like "06/27/1987" and reformats it to "'1987-06-27'"
    private static String mySQLDate(String dt) {
	int firstSlash = dt.indexOf('/');
	int secondSlash = dt.indexOf("/",firstSlash+1);
	String m = dt.substring(0,firstSlash);
	String d = dt.substring(firstSlash+1,secondSlash);
	String y = dt.substring(secondSlash+1);
	return ("'" + y + '-' + m + '-' + d + "'");
    }

    private static int yrsToInt(String y) {
	if (y.equals("R"))
	    return 0;
	return Integer.parseInt(y);
    }

    public void scrape(PrintWriter pw) throws IOException {
	List<String> lines = g.getContents();
	for (int i=0; i<lines.size(); i++) {

	    String line = lines.get(i);

	    if (isPlayerBlock(line)) {

		// get player's number
		String num = discardTags(line);
		pw.print(num + ",");

		// get player's name
		i += 3;
		String name = dropLastTag(lines.get(i));
		pw.print(name + ",");
		
		// get player's position
		i += 3;
		String pos = dropLastTag(lines.get(i));
		pw.print(pos + ",");
		
		// get height
		i += 2;
		String ht = dropLastTag(lines.get(i));
		int ht_inches = inches(ht);
		pw.print(ht_inches + ",");
		
		// get weight
		i += 2;
		String wt = dropLastTag(lines.get(i));
		pw.print(wt + ",");
		
		// dob
		i += 2;
		String dob = mySQLDate(dropLastTag(lines.get(i)));
		pw.print(dob + ",");
		
		// college
		i += 2;
		String college = dropLastTag(lines.get(i)).substring(7);
		pw.print(college + ",");

		// yrs
		i += 2;
		int yrs = yrsToInt(dropLastTag(lines.get(i)));
		pw.print(yrs + ",");

		pw.println(this.teamname);
	    }
	}
    }

    private static String capital(String s) {
	return s.substring(0,1).toUpperCase() + s.substring(1);
    }

    public static void main(String[] args) throws IOException {

	System.out.println("scraping...");

	FileWriter fw = new FileWriter("C:\\NBA.txt");
	PrintWriter pw = new PrintWriter(fw);

	String[] teams = {"mavericks", "rockets", "grizzlies", "hornets", "spurs",
			  "warriors", "clippers", "lakers", "suns", "kings",
			  "nuggets", "timberwolves", "blazers", "sonics", "jazz",
			  "celtics", "nets", "knicks", "sixers", "raptors", 
			  "bulls", "cavaliers", "pistons", "pacers", "bucks", 
			  "hawks","bobcats", "heat", "magic", "wizards"};

	for (String team : teams) {
	    NBARosterScraper s = new NBARosterScraper(team);
	    s.scrape(pw);
	    String Team = capital(team);
	    System.out.println("Wrote data for " + Team + ".");
	}

	pw.close();
	fw.close();

	System.out.println("Finished.");		
    }
}

/* ************************* */

DROP TABLE BBPlayers;

CREATE TABLE BBPlayers (
  PlayerID INT AUTO_INCREMENT,
  Number INT,
  Name VARCHAR(150),
  Pos VARCHAR(5),
  Height INT,
  Weight INT,  
  DOB DATE,
  College VARCHAR(150),
  Yrs0506 INT,
  Team VARCHAR(150),
  PRIMARY KEY (PlayerID)
);

LOAD DATA
LOCAL INFILE "NBA.txt"
REPLACE INTO TABLE BBPlayers
FIELDS TERMINATED BY ","
(Number, Name, Pos, Height, Weight, DOB, College, Yrs0506, Team);