/* Use, modification, and distribution are subject to the Boost Software 
License, Version 1.0. (See accompanying file LICENSE_1.0.txt or copy at 
www.boost.org/LICENSE_1.0.txt) */

import java.util.*;

class Limiter
{
    protected static long LAST_PAUSE_TIME;
    public static long EARLIEST_CRAWL_TIME;
    private static Calendar calendar;

    public static boolean index(String url)
    {
	return url.matches(".*carleton\\.edu.*");
	//return !url.matches(".*mathcs\\.carleton\\.edu.*");
	//return true;
    }

    public static boolean follow(String url)
    {
	//return url.matches("http://violet\\.mathcs\\.carleton\\.edu.*");
	return (url.matches("https?://[^/]*carleton\\.edu.*") &&
		!url.matches(".*muse\\.library\\.carleton\\.edu/search/.*") &&
		!url.matches(".*webkiosk\\.carleton\\.edu.*") &&
		!url.matches("http://webapps.acs.carleton.edu/calendar/academic/\\?.*"));
    }
    
    public static boolean parse(String url)
    {
	if (url.matches(".*&.*"))
	    return false;
	if (url.matches(".*textonly=1.*"))
	    return false;
	if (url.matches("(?i)\\Ahttps?://[^/]*\\Z"))
	    return true;
	if (url.matches("(?i)\\Ahttps?://.*(htm|html|php|php3|php4|php5|pl|asp|/)(\\?[^?]*){0,1}\\Z"))
	    return true;
	return false;
    }

    public static boolean fetch(String host)
    {
	if(calendar == null) {
	    String [] ids = TimeZone.getAvailableIDs(-6 * 60 * 60 * 1000);
	    SimpleTimeZone cdt = new SimpleTimeZone(-6 * 60 * 60 * 1000, ids[0]);

	    cdt.setStartRule(Calendar.APRIL, 1, Calendar.SUNDAY, 2 * 60 * 60 * 1000);
	    cdt.setEndRule(Calendar.OCTOBER, -1, Calendar.SUNDAY, 2 * 60 * 60 * 1000);

	    calendar = new GregorianCalendar(cdt);
	}

	Date trialTime = new Date();
	calendar.setTime(trialTime);

	if(!host.matches(".*carleton\\.edu.*")) {
	    return true;
	}
	
	long h = calendar.get(Calendar.HOUR_OF_DAY);
	long m = calendar.get(Calendar.MINUTE);
	long d = calendar.get(Calendar.DAY_OF_WEEK);

	// peek times - all carleton servers
	
	
	if((d > Calendar.SUNDAY && d < Calendar.SATURDAY) && (h >= 8 && h < 17)) {
	    LAST_PAUSE_TIME = System.currentTimeMillis();
	    if(17 < EARLIEST_CRAWL_TIME) EARLIEST_CRAWL_TIME = 17;
	    return false;
	}
	
// maintenance
// 	if(d == calendar.FRIDAY && h >= 21) {
// 	    LAST_PAUSE_TIME = System.currentTimeMillis();
// 	    return false;
// 	}
// 	if((d == calendar.SATURDAY || d == calendar.SUNDAY) && (h >= 12 && h < 17)) {
// 	    LAST_PAUSE_TIME = System.currentTimeMillis();
// 	    return false;
// 	}
	LAST_PAUSE_TIME = -1;
	return true;
    }

    public static boolean isEarliestCrawlTime()
    {
	if(calendar == null) {
	    String [] ids = TimeZone.getAvailableIDs(-6 * 60 * 60 * 1000);
	    SimpleTimeZone cdt = new SimpleTimeZone(-6 * 60 * 60 * 1000, ids[0]);
	    
	    cdt.setStartRule(Calendar.APRIL, 1, Calendar.SUNDAY, 2 * 60 * 60 * 1000);
	    cdt.setEndRule(Calendar.OCTOBER, -1, Calendar.SUNDAY, 2 * 60 * 60 * 1000);
	    
	    calendar = new GregorianCalendar(cdt);
	}

	Date trialTime = new Date();
	calendar.setTime(trialTime);

	long h = calendar.get(Calendar.HOUR_OF_DAY);
	long m = calendar.get(Calendar.MINUTE);
	long d = calendar.get(Calendar.DAY_OF_WEEK);

	if (h >= EARLIEST_CRAWL_TIME) {
	    return true;
	}
	else {
	    return false;
	}
    }
}