traininfoservice/banedk/TimetableFetcher.java

package dk.thoerup.traininfoservice.banedk;


import java.io.IOException;
import java.net.URL;
import java.util.ArrayList;
import java.util.List;
import java.util.logging.Logger;

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;

public class TimetableFetcher {
        
        class NullRefreshHandler implements RefreshHandler {
                public void handleRefresh(Page arg0, URL arg1, int arg2) throws IOException {                   
                }
                
        }
                
        TimeoutCache<String, List<TimetableBean>> cache = new TimeoutCache<String,List<TimetableBean>>(120 * 1000);
        
        
        Logger logger = Logger.getLogger(TimetableFetcher.class.getName());
        
        
        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 {
                        logger.info("Timetable: Cache hit " + trainID); 
                }
                return list;
        }

        List<TimetableBean> lookupTimetable(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";
                                

            final WebClient webClient = new WebClient();
            webClient.setTimeout(2500);
            webClient.setJavaScriptEnabled(false);         
            webClient.setRefreshHandler( new NullRefreshHandler() );
            webClient.setCssEnabled(false);
            
            
            final HtmlPage page = webClient.getPage(url);
            
            
            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();
                        bean.setStation( fields.get(0).asText() );
                        bean.setArrival( fields.get(1).asText() );
                        bean.setDeparture( fields.get(2).asText() );
                        
                        if (currentStation == true && currentStationSaved == false ) {
                                bean.setCurrent(currentStation);
                                currentStationSaved = true;
                        }
                        
                        timetableList.add(bean);

                }
                
            } else {
                logger.warning("No time table found, trainID=" + trainID + " type=" + type);
            }
                
                return timetableList;
        }
        
}
1	package dk.thoerup.traininfoservice.banedk;
2
3
4	import java.io.IOException;
5	import java.net.URL;
6	import java.util.ArrayList;
7	import java.util.List;
8	import java.util.logging.Logger;
9
10	import com.gargoylesoftware.htmlunit.Page;
11	import com.gargoylesoftware.htmlunit.RefreshHandler;
12	import com.gargoylesoftware.htmlunit.WebClient;
13	import com.gargoylesoftware.htmlunit.html.DomNodeList;
14	import com.gargoylesoftware.htmlunit.html.HtmlElement;
15	import com.gargoylesoftware.htmlunit.html.HtmlPage;
16
17	public class TimetableFetcher {
18
19	class NullRefreshHandler implements RefreshHandler {
20	public void handleRefresh(Page arg0, URL arg1, int arg2) throws IOException {
21	}
22
23	}
24
25	TimeoutCache<String, List<TimetableBean>> cache = new TimeoutCache<String,List<TimetableBean>>(120 * 1000);
26
27
28	Logger logger = Logger.getLogger(TimetableFetcher.class.getName());
29
30
31	List<TimetableBean> cachedLookupTimetable(String trainID, String type) throws Exception {
32	String key = trainID+type;
33	List<TimetableBean> list = cache.get(key);
34
35	if (list == null) {
36	list = lookupTimetable(trainID,type);
37	cache.put(key, list);
38	} else {
39	logger.info("Timetable: Cache hit " + trainID);
40	}
41	return list;
42	}
43
44	List<TimetableBean> lookupTimetable(String trainID, String type) throws Exception {
45	List<TimetableBean> timetableList = new ArrayList<TimetableBean>();
46
47	String url = "http://www.bane.dk/visRute.asp?W=" + type + "&TogNr=" + trainID + "&artikelId=4276";
48
49
50	final WebClient webClient = new WebClient();
51	webClient.setTimeout(2500);
52	webClient.setJavaScriptEnabled(false);
53	webClient.setRefreshHandler( new NullRefreshHandler() );
54	webClient.setCssEnabled(false);
55
56
57	final HtmlPage page = webClient.getPage(url);
58
59
60	boolean currentStation = false;
61	boolean currentStationSaved = false;
62
63	List<HtmlElement> tables = page.getDocumentElement().getElementsByAttribute("table", "class", "Rute");
64	if (tables.size() == 1) {
65	HtmlElement timetable = tables.get(0);
66	DomNodeList<HtmlElement> rows = timetable.getElementsByTagName("tr");
67
68	for (int i=0; i<rows.size(); i++) {
69	if (i==0) //First row is column headers
70	continue;
71
72
73	HtmlElement row = rows.get(i);
74	DomNodeList<HtmlElement> fields = row.getElementsByTagName("td");
75
76	if (currentStationSaved == false && fields.get(0).getAttribute("class").equalsIgnoreCase("Tidsstreg")) {
77	currentStation = true;
78	continue;
79	}
80
81	TimetableBean bean = new TimetableBean();
82	bean.setStation( fields.get(0).asText() );
83	bean.setArrival( fields.get(1).asText() );
84	bean.setDeparture( fields.get(2).asText() );
85
86	if (currentStation == true && currentStationSaved == false ) {
87	bean.setCurrent(currentStation);
88	currentStationSaved = true;
89	}
90
91	timetableList.add(bean);
92
93	}
94
95	} else {
96	logger.warning("No time table found, trainID=" + trainID + " type=" + type);
97	}
98
99	return timetableList;
100	}
101
102	}