import java.sql.*;
import java.util.*;
import java.io.*;
import java.text.*;

public class AddBook {
    
    public static final String url = "[url omitted]";
    public static final String username = "[username omitted]";
    public static final String passwd = "[password omitted]";
    public static final String database = "[database omitted]";
    public static final String driver = "org.gjt.mm.mysql.Driver";
    
    private static final int NUM_TALLIES=3;
    private long start = 0;
    private int[] tally = new int[NUM_TALLIES + 1];

    private Connection con;
    private Statement stmt;

    public AddBook() {
	tally[0] = tally[1] = 0;
	try {
	    Class.forName(driver);
	    con = DriverManager.getConnection(url, username, passwd);
	} catch (Exception e) {
	    e.printStackTrace(System.err);
	}
    }


    private int addNewPage(int work, int pageInWork,
			   int globalPosition, int lastPageId) 
	throws SQLException {
	startTally();
	if (lastPageId >= 0) {
	    stmt.executeUpdate("UPDATE UniquePage SET global_end = " +
			       globalPosition + " WHERE id = " + lastPageId);
	}
	stmt.executeUpdate("INSERT INTO UniquePage VALUES (0, " +
			   work + ", " + pageInWork + ", " + 
			   (globalPosition + 1) + ", -1)");
	endTally(2);
	startTally();
	ResultSet rs = stmt.executeQuery("SELECT last_insert_id() FROM UniquePage LIMIT 1");
	rs.next();
	int pageId = rs.getInt(1);
	endTally(2);
	return pageId;
    }


    private void insertIntoPCW(int wordId, int pageId) throws SQLException {
	startTally();
	ResultSet rs = stmt.executeQuery
	    ("SELECT * FROM PageContainsWord WHERE " +
	     "word=" + wordId + " AND page=" + pageId);
	endTally(0);
	if (rs.next()) return;
	startTally();
	stmt.executeUpdate("INSERT INTO PageContainsWord VALUES (" +
			   wordId + ", " + pageId + ")");
	endTally(1);
    }


    private void insertIntoWI(int gid,
			     int wordId, int prevWord) throws SQLException {
	startTally();
	stmt.executeUpdate("INSERT INTO WordInstance VALUES(" + wordId + ", " +
			   gid + ", " + prevWord + ")");
	endTally(1);
    }


    private int getWordId(String root) throws SQLException {
	int id;
	startTally();
	ResultSet rs = stmt.executeQuery
	    ("SELECT * FROM Word where word=" + formatString(root));
	
	if (!rs.next()) {
	    rs = stmt.executeQuery("SELECT (MAX(id)+1) FROM Word");
	    rs.next();
	    id = rs.getInt(1);
	    endTally(0);
	    startTally();
	    stmt.executeUpdate("INSERT INTO Word VALUES (" +
			       id + ", " + formatString(root) + ")");
	    endTally(1);
	    System.err.println("Inserted: " + formatString(root) +
			       " (" + id + ")");
	} else {
	    id = rs.getInt(1);
	    endTally(0);
	}
	return id;
    }


    private void insertIntoWD(int globalPosition, String word) 
	throws SQLException {
	startTally();
	stmt.executeUpdate("INSERT INTO WordDetails VALUES (" +
			   globalPosition + ", " + formatString(word) + ")");
	endTally(1);
    }
	

    private static final int LINES_PER_PAGE = 70;

    public void parse(int workId) {
	int globalPosition = -1;
	int pageId = -1;
	int pageInWork = 0;
	int prevWord = -1;
	int lineOnPage = LINES_PER_PAGE;

	BufferedReader in = null;
	try {
	    stmt = con.createStatement();
	    ResultSet rs = stmt.executeQuery("SELECT MAX(gpos) FROM WordInstance");
	    rs.next();
	    globalPosition = rs.getInt(1);
	    in = new BufferedReader(new FileReader(workId + ".txt"));
	} catch (Exception e) { e.printStackTrace(System.err); }

	String line;
	while (true) {
	    try {
	        try {
		    line = in.readLine();
		} catch (NullPointerException e) { break; }
		if (line == null) {
		    //end of the work, close off the last page
		    stmt.executeUpdate("UPDATE UniquePage SET global_end = " +
			       globalPosition + " WHERE id = " + pageId);
		    break;
		}
		
		if (lineOnPage == LINES_PER_PAGE) {
		    pageInWork++;
		    lineOnPage = 1;
		    pageId = addNewPage(workId, pageInWork, globalPosition, 
					pageId);

		    int selectPct = tally[0] * 100 / tally[NUM_TALLIES];
		    int updatePct = tally[1] * 100 / tally[NUM_TALLIES];
		    int pgPct =  tally[2] * 100 / tally[NUM_TALLIES];
		    System.err.println("Starting new page: id:" + pageId +
				       ", p:" + pageInWork + ", gp:" +
				       globalPosition + ", s:" + selectPct +", up:" + updatePct + ", pg:" + pgPct);
		}

		StringTokenizer st = new StringTokenizer(line, "\t\n\r -");
		
		while (st.hasMoreTokens()) {
		    String word = st.nextToken();
		    String root = PorterStemmer.stem(word);
		    if (!st.hasMoreTokens()) word += "-";  //end of line

		    int wordId = getWordId(root);
		    globalPosition++;
		    insertIntoPCW(wordId, pageId);
		    insertIntoWI(globalPosition, wordId, prevWord); 
		    insertIntoWD(globalPosition, word);
		    prevWord = wordId;
		}

	    } catch (Exception e) {
		e.printStackTrace(System.err);
	    }
	    lineOnPage++;

	    try {
		con.commit();
	    } catch (Exception e) {}
	}
    }


    public String formatString(String s) {
	if (s == null) return "null";
	
	StringBuffer buf = new StringBuffer(s);
	for (int i = 0; i < buf.length(); i++) {
	    if (buf.charAt(i) == '\'') {
		buf.insert(i, '\'');
		i++;
	    }
	}
	buf.insert(0, '\'');
	buf.append('\'');
	return buf.toString();
    }


    private void startTally() {
	start = System.currentTimeMillis();
    }

    private void endTally(int idx) {
	long duration = System.currentTimeMillis() - start;
	tally[idx] += (int) duration;
	tally[NUM_TALLIES] += (int) duration; 
    }


  private static SimpleDateFormat df = new SimpleDateFormat("[yyyy-MM-dd HH:mm:ss]  ");

  public static void main(String[] arg) {
    try {
        PrintWriter log = new PrintWriter(new FileWriter("addlog.txt"));
        int i = Integer.parseInt(arg[0]);
	log.println(df.format(new java.util.Date()) + "Starting run at " + i);
	log.flush();
        BufferedReader r = new BufferedReader(new FileReader("filenames"));
	String line;
	ArrayList a = new ArrayList();
	AddBook adder = new AddBook();
	while ((line = r.readLine()) != null) {
	  if (line.indexOf('|') > 0) {
	    String id = line.substring(0, line.indexOf('|'));
	    a.add(id);
	  }
	}

	Iterator itr = a.iterator();
	while (itr.hasNext() && i-- > 1) { itr.next(); }
	int sequence = 1;
	while (itr.hasNext()) {
	    int id = Integer.parseInt((String) itr.next());
	    log.println(df.format(new java.util.Date()) + "Starting parse of work " + id + "(" + sequence + ")");
	    log.flush();
	    sequence++;
	    adder.parse(id);
	    System.out.println("Finished work : " + id);

	    Thread.sleep(10000);
	}
	log.close();
    } catch (Exception e) {
      e.printStackTrace(System.err);
    }
  }
}