traininfoservice/banedk/TimetableFetcher.java

package dk.thoerup.traininfoservice.banedk;


import java.io.IOException;
import java.net.URL;
import java.sql.SQLException;
import java.util.ArrayList;
import java.util.List;
import java.util.Map;
import java.util.logging.Level;
import java.util.logging.Logger;

import com.gargoylesoftware.htmlunit.BrowserVersion;
import com.gargoylesoftware.htmlunit.Page;
import com.gargoylesoftware.htmlunit.RefreshHandler;
import com.gargoylesoftware.htmlunit.WebClient;
import com.gargoylesoftware.htmlunit.html.DomNodeList;
import com.gargoylesoftware.htmlunit.html.HtmlElement;
import com.gargoylesoftware.htmlunit.html.HtmlPage;

import dk.thoerup.circuitbreaker.CircuitBreaker;
import dk.thoerup.circuitbreaker.CircuitBreakerManager;
import dk.thoerup.traininfoservice.StationDAO;
import dk.thoerup.traininfoservice.Statistics;

public class TimetableFetcher {
        
        class NullRefreshHandler implements RefreshHandler {
                public void handleRefresh(Page arg0, URL arg1, int arg2) throws IOException {                   
                }
                
        }
                
        Map<String, List<TimetableBean>> cache; 
        Map<String, Integer> stationCache;

        StationDAO stationDao = new StationDAO();
        
        
        Logger logger = Logger.getLogger(TimetableFetcher.class.getName());
        
        private boolean useTempSite;
        
        public TimetableFetcher(boolean tmpSite, int cacheTimeout) {
                useTempSite = tmpSite;
                
                cache = new TimeoutMap<String,List<TimetableBean>>(cacheTimeout);
                stationCache = new TimeoutMap<String,Integer>( 3*60*60*1000 );
        }
        
        
        List<TimetableBean> cachedLookupTimetable(String trainID, String type) throws Exception {
                String key = trainID+type;
                List<TimetableBean> list = cache.get(key);
                
                if (list == null) {
                        list = lookupTimetable(trainID,type);
                        cache.put(key, list);
                } else {
                        Statistics.getInstance().incrementTimetableCacheHits();
                        logger.info("Timetable: Cache hit " + trainID); 
                }
                return list;
        }
        
        List<TimetableBean> lookupTimetable(String trainID, String type) throws Exception {
                if (useTempSite == false ){
                        return lookupTimetableRealSite(trainID, type);
                } else {
                        return new ArrayList<TimetableBean>(); // no timetable data on temp site
                }
        }
        
        int getStationId(String name) {
                Integer id = stationCache.get(name);
                
                if (id == null) {
                        try {
                                id = stationDao.getIdByName(name);
                                stationCache.put(name, id);
                        } catch (SQLException e) {
                                logger.log(Level.SEVERE, "getStationId failed", e);
                                id = -1;
                        }
                } 

                return id;
        }

        List<TimetableBean> lookupTimetableRealSite(String trainID, String type) throws Exception {             
                List<TimetableBean> timetableList = new ArrayList<TimetableBean>();
                
                //String url = "http://www.bane.dk/visRute.asp?W=" + type + "&TogNr=" + trainID + "&artikelId=4276";
                String url = "http://trafikinfo.bane.dk/TrafikInformation/Ruteplan/" + trainID;                 

            final WebClient webClient = new WebClient(BrowserVersion.FIREFOX_3);
            webClient.setTimeout(2500);
            webClient.setJavaScriptEnabled(false);         
            webClient.setRefreshHandler( new NullRefreshHandler() );
            webClient.setCssEnabled(false);
            
            
            HtmlunitInvocation wrapper = new HtmlunitInvocation(webClient, url);
            CircuitBreaker breaker = CircuitBreakerManager.getManager().getCircuitBreaker("banedk");
            
            HtmlPage page = (HtmlPage) breaker.invoke(wrapper);
                    
            
            boolean currentStation = false;
            boolean currentStationSaved = false;
            
            List<HtmlElement> tables = page.getDocumentElement().getElementsByAttribute("table", "class", "Rute");
            if (tables.size() == 1) {
                HtmlElement timetable = tables.get(0);
                DomNodeList<HtmlElement> rows = timetable.getElementsByTagName("tr");
                
                for (int i=0; i<rows.size(); i++) {
                        if (i==0) //First row is column headers
                                continue;
                        
                        
                        HtmlElement row = rows.get(i);
                        DomNodeList<HtmlElement> fields = row.getElementsByTagName("td");
                        
                        if (currentStationSaved == false && fields.get(0).getAttribute("class").equalsIgnoreCase("Tidsstreg")) {
                                currentStation = true;
                                continue;
                        }
                        
                        TimetableBean bean = new TimetableBean();
                        
                        String station = fields.get(0).asText() ;
                        if (station.equals("København"))
                                station = "København H"; //correct inconsistency in naming
                        
                        bean.setStation( station );
                        bean.setArrival( fields.get(1).asText() );
                        bean.setDeparture( fields.get(2).asText() );
                        
                        boolean cancelled = fields.get(3).asText().equalsIgnoreCase("aflyst");
                        bean.setCancelled(cancelled);
                        
                        if (currentStation == true && currentStationSaved == false ) {
                                bean.setCurrent(currentStation);
                                currentStationSaved = true;
                        }
                        
                        bean.setStationId( getStationId( station ));
                        
                        timetableList.add(bean);
                }
                
                final String cancelledString = "Aflyst";
                for (int i=0;i<timetableList.size(); i++) { //handle cancelled labels
                        final int lastIdx = (timetableList.size() - 1);
                        
                        TimetableBean current = timetableList.get(i);
                        if (current.isCancelled() && i == 0) {
                                current.setDeparture(cancelledString);
                        } else if (current.isCancelled() && i == lastIdx) {
                                current.setArrival(cancelledString);
                        } else if (current.isCancelled() && i>0 && i<lastIdx) {
                                TimetableBean next = timetableList.get(i+1);
                                TimetableBean prev = timetableList.get(i-1);
                                
                                if (next.isCancelled())
                                        current.setDeparture(cancelledString);
                                if (prev.isCancelled())
                                        current.setArrival(cancelledString);
                        }
                }
                
            } else {
                logger.warning("No time table found, trainID=" + trainID + " type=" + type);
            }
            webClient.closeAllWindows();
                
                return timetableList;
        }
        
}
1	torben	350	package dk.thoerup.traininfoservice.banedk;
2
3
4			import java.io.IOException;
5			import java.net.URL;
6	torben	836	import java.sql.SQLException;
7	torben	350	import java.util.ArrayList;
8			import java.util.List;
9	torben	428	import java.util.Map;
10	torben	836	import java.util.logging.Level;
11	torben	350	import java.util.logging.Logger;
12
13	torben	591	import com.gargoylesoftware.htmlunit.BrowserVersion;
14	torben	387	import com.gargoylesoftware.htmlunit.Page;
15			import com.gargoylesoftware.htmlunit.RefreshHandler;
16			import com.gargoylesoftware.htmlunit.WebClient;
17			import com.gargoylesoftware.htmlunit.html.DomNodeList;
18			import com.gargoylesoftware.htmlunit.html.HtmlElement;
19			import com.gargoylesoftware.htmlunit.html.HtmlPage;
20	torben	350
21	torben	468	import dk.thoerup.circuitbreaker.CircuitBreaker;
22			import dk.thoerup.circuitbreaker.CircuitBreakerManager;
23	torben	836	import dk.thoerup.traininfoservice.StationDAO;
24	torben	711	import dk.thoerup.traininfoservice.Statistics;
25	torben	421
26	torben	350	public class TimetableFetcher {
27
28			class NullRefreshHandler implements RefreshHandler {
29			public void handleRefresh(Page arg0, URL arg1, int arg2) throws IOException {
30			}
31
32			}
33	torben	387
34	torben	584	Map<String, List<TimetableBean>> cache;
35	torben	836	Map<String, Integer> stationCache;
36
37			StationDAO stationDao = new StationDAO();
38	torben	350
39	torben	387
40	torben	350	Logger logger = Logger.getLogger(TimetableFetcher.class.getName());
41	torben	387
42	torben	581	private boolean useTempSite;
43	torben	387
44	torben	584	public TimetableFetcher(boolean tmpSite, int cacheTimeout) {
45	torben	581	useTempSite = tmpSite;
46	torben	584
47			cache = new TimeoutMap<String,List<TimetableBean>>(cacheTimeout);
48	torben	836	stationCache = new TimeoutMap<String,Integer>( 36060*1000 );
49	torben	581	}
50
51
52	torben	451	List<TimetableBean> cachedLookupTimetable(String trainID, String type) throws Exception {
53	torben	387	String key = trainID+type;
54	torben	837	List<TimetableBean> list = cache.get(key);
55	torben	387
56			if (list == null) {
57			list = lookupTimetable(trainID,type);
58			cache.put(key, list);
59			} else {
60	torben	711	Statistics.getInstance().incrementTimetableCacheHits();
61	torben	389	logger.info("Timetable: Cache hit " + trainID);
62	torben	387	}
63			return list;
64			}
65	torben	581
66			List<TimetableBean> lookupTimetable(String trainID, String type) throws Exception {
67			if (useTempSite == false ){
68			return lookupTimetableRealSite(trainID, type);
69			} else {
70			return new ArrayList<TimetableBean>(); // no timetable data on temp site
71			}
72			}
73	torben	836
74			int getStationId(String name) {
75			Integer id = stationCache.get(name);
76
77			if (id == null) {
78			try {
79	torben	842	id = stationDao.getIdByName(name);
80	torben	836	stationCache.put(name, id);
81			} catch (SQLException e) {
82			logger.log(Level.SEVERE, "getStationId failed", e);
83			id = -1;
84			}
85			}
86	torben	350
87	torben	836	return id;
88			}
89
90	torben	581	List<TimetableBean> lookupTimetableRealSite(String trainID, String type) throws Exception {
91	torben	350	List<TimetableBean> timetableList = new ArrayList<TimetableBean>();
92
93	torben	970	//String url = "http://www.bane.dk/visRute.asp?W=" + type + "&TogNr=" + trainID + "&artikelId=4276";
94			String url = "http://trafikinfo.bane.dk/TrafikInformation/Ruteplan/" + trainID;
95	torben	350
96	torben	591	final WebClient webClient = new WebClient(BrowserVersion.FIREFOX_3);
97	torben	386	webClient.setTimeout(2500);
98	torben	350	webClient.setJavaScriptEnabled(false);
99			webClient.setRefreshHandler( new NullRefreshHandler() );
100			webClient.setCssEnabled(false);
101
102
103	torben	939	HtmlunitInvocation wrapper = new HtmlunitInvocation(webClient, url);
104	torben	421	CircuitBreaker breaker = CircuitBreakerManager.getManager().getCircuitBreaker("banedk");
105	torben	350
106	torben	421	HtmlPage page = (HtmlPage) breaker.invoke(wrapper);
107
108	torben	350
109			boolean currentStation = false;
110			boolean currentStationSaved = false;
111
112			List<HtmlElement> tables = page.getDocumentElement().getElementsByAttribute("table", "class", "Rute");
113			if (tables.size() == 1) {
114			HtmlElement timetable = tables.get(0);
115			DomNodeList<HtmlElement> rows = timetable.getElementsByTagName("tr");
116
117			for (int i=0; i<rows.size(); i++) {
118			if (i==0) //First row is column headers
119			continue;
120
121
122			HtmlElement row = rows.get(i);
123			DomNodeList<HtmlElement> fields = row.getElementsByTagName("td");
124
125			if (currentStationSaved == false && fields.get(0).getAttribute("class").equalsIgnoreCase("Tidsstreg")) {
126			currentStation = true;
127			continue;
128			}
129
130			TimetableBean bean = new TimetableBean();
131	torben	836
132			String station = fields.get(0).asText() ;
133			if (station.equals("København"))
134			station = "København H"; //correct inconsistency in naming
135
136			bean.setStation( station );
137	torben	350	bean.setArrival( fields.get(1).asText() );
138			bean.setDeparture( fields.get(2).asText() );
139
140	torben	975	boolean cancelled = fields.get(3).asText().equalsIgnoreCase("aflyst");
141			bean.setCancelled(cancelled);
142
143	torben	350	if (currentStation == true && currentStationSaved == false ) {
144			bean.setCurrent(currentStation);
145			currentStationSaved = true;
146			}
147
148	torben	836	bean.setStationId( getStationId( station ));
149
150	torben	350	timetableList.add(bean);
151			}
152
153	torben	975	final String cancelledString = "Aflyst";
154			for (int i=0;i<timetableList.size(); i++) { //handle cancelled labels
155			final int lastIdx = (timetableList.size() - 1);
156
157			TimetableBean current = timetableList.get(i);
158			if (current.isCancelled() && i == 0) {
159			current.setDeparture(cancelledString);
160			} else if (current.isCancelled() && i == lastIdx) {
161			current.setArrival(cancelledString);
162			} else if (current.isCancelled() && i>0 && i<lastIdx) {
163			TimetableBean next = timetableList.get(i+1);
164			TimetableBean prev = timetableList.get(i-1);
165
166			if (next.isCancelled())
167			current.setDeparture(cancelledString);
168			if (prev.isCancelled())
169			current.setArrival(cancelledString);
170			}
171			}
172
173	torben	350	} else {
174			logger.warning("No time table found, trainID=" + trainID + " type=" + type);
175			}
176	torben	591	webClient.closeAllWindows();
177	torben	350
178			return timetableList;
179			}
180
181			}