src/skrueger/Sitemap.java

package skrueger;

import java.io.FileWriter;
import java.io.IOException;
import java.io.StringWriter;
import java.io.Writer;
import java.net.MalformedURLException;
import java.net.URL;
import java.net.URLEncoder;
import java.util.Date;

import javax.xml.parsers.DocumentBuilderFactory;
import javax.xml.transform.OutputKeys;
import javax.xml.transform.Result;
import javax.xml.transform.Transformer;
import javax.xml.transform.TransformerFactory;
import javax.xml.transform.dom.DOMSource;
import javax.xml.transform.stream.StreamResult;

import net.sf.antcontrib.property.URLEncodeTask;

import org.apache.log4j.Logger;
import org.w3c.dom.Document;
import org.w3c.dom.Element;

/**
 * This class provides an easy way to list all pages of your site (including
 * generates ones) ina sitemap.xml. See http://sitemaps.org/protocol.php for the
 * XML specification.
 * 
 * <br/>
 * XML output is based on org.w3c.dom
 * 
 * 
 * @author Stefan Tzeggai
 */
public class Sitemap {

        static String[] SEARCHENGINES = new String[] {
                        "http://www.sitemapwriter.com/notify.php?crawler=all&url=",
                        "http://submissions.ask.com/ping?sitemap=",
                        "http://www.google.com/webmasters/tools/ping?sitemap=",
                        "http://api.moreover.com/ping?u=",
                        "http://www.bing.com/webmaster/ping.aspx?siteMap=",
                        "http://search.yahooapis.com/SiteExplorerService/V1/ping?sitemap=" };

        final static Logger log = Logger.getLogger(Sitemap.class);

        private Document document;

        final private Element urlsetElement;

        /**
         * Returns the number of URLs. May not be more than 50,000 by spec.
         */
        private int size = 0;

        final static String NSURL = "http://www.sitemaps.org/schemas/sitemap/0.9";

        public enum CHANGEFREQ {
                monthly, daily, weekly
        }

        public Sitemap() {

                try {

                        // Create a DOM builder and parse the fragment
                        final DocumentBuilderFactory factory = DocumentBuilderFactory
                                        .newInstance();
                        document = factory.newDocumentBuilder().newDocument();

                        // XML root element
                        urlsetElement = document.createElementNS(NSURL, "urlset");
                        document.appendChild(urlsetElement);

                        // // Linking this XML to the AtlasML Schema
                        // final Attr namespaces = document.createAttributeNS(
                        // "http://www.w3.org/2001/XMLSchema-instance", "schemaLocation");
                        // namespaces
                        // .setValue(NSURL+" http://localhost:"
                        // + Webserver.DEFAULTPORT
                        // + "/skrueger/atlas/resource/AtlasML.xsd");
                        // urlset.setAttributeNode(namespaces);

                } catch (Exception e) {
                        throw new RuntimeException(
                                        "Sitemap org.w3c.xml newDocumentBuilder failed:", e);
                }
        }

        /**
         * 
         * @param locString
         * @param lastmod
         *            when was the page modified the last time? May be
         *            <code>null</code>.
         * @param changefreq
         *            How often does the page change. May be <code>null</code>.
         * @param priority
         *            Priority 0.0 - 1.0. May be <code>null</code>.
         * @return <code>true</code> if the {@link Sitemap} contains less than 50000
         *         urls and the page was added.
         */
        public boolean addUrl(String locString, Date lastmod,
                        CHANGEFREQ changefreq, Double priority) {
                size++;
                if (size >= 50000)
                        return false;

                Element urlElement = document.createElement("url");

                if (locString == null)
                        throw new IllegalArgumentException("location must be provided");
                else {
                        Element e = document.createElement("loc");
                        e.appendChild(document.createTextNode(locString));
                        urlElement.appendChild(e);
                }

                if (lastmod != null) {
                        Element e = document.createElement("lastmod");
                        e.appendChild(document.createTextNode(lastmod.toString()));
                        urlElement.appendChild(e);
                }

                if (changefreq != null) {
                        Element e = document.createElement("changefreq");
                        e.appendChild(document.createTextNode(changefreq.toString()));
                        urlElement.appendChild(e);
                }

                if (priority != null) {
                        Element e = document.createElement("priority");
                        e.appendChild(document.createTextNode(priority.toString()));
                        urlElement.appendChild(e);
                }
                urlsetElement.appendChild(urlElement);
                return true;
        }

        public Document getDocument() {
                return document;
        }

        public String getXmlString() {
                StringWriter stringWriter = new StringWriter();
                outputToWriter(stringWriter);
                return stringWriter.toString();
        }

        /**
         * Can be used to write to a File with {@link FileWriter}.
         */
        public void outputToWriter(Writer stringWriter) {
                try {

                        try { // close outputStreamWriter.close();

                                // ****************************************************************************
                                // Create the XML
                                // ****************************************************************************
                                final Result result = new StreamResult(stringWriter);

                                // with indenting to make it human-readable
                                final TransformerFactory tf = TransformerFactory.newInstance();

                                // TODO Ging mit xerces, geht nicht mehr mit xalan ?!
                                // tf.setAttribute("indent-number", new Integer(2));

                                final Transformer xformer = tf.newTransformer();
                                xformer.setOutputProperty(OutputKeys.INDENT, "yes");
                                xformer.setOutputProperty(
                                                "{http://xml.apache.org/xalan}indent-amount", "2");

                                // Write the DOM document to the file
                                xformer.transform(new DOMSource(document), result);

                        } finally {
                                stringWriter.close();
                        }

                } catch (Exception e) {
                        log.error("Failed to create sitemap.XML-String", e);
                        throw new RuntimeException(e);
                }
        }

        public int getSize() {
                return size;
        }

        public static void submitToSearchEngines(String urlToSitemap) {
                for (String se : SEARCHENGINES) {
                        log.info("Submitting " + urlToSitemap + " to " + se);
                        try {
                                URL url = new URL(se + urlToSitemap);

                                urlToSitemap = URLEncoder.encode(urlToSitemap, "UTF8");

                                url.openConnection().connect();

                        } catch (Exception e) {
                                log.error("failed to submit " + urlToSitemap + " to " + se, e);
                        }
                }
        }
}
1	package skrueger;
2
3	import java.io.FileWriter;
4	import java.io.IOException;
5	import java.io.StringWriter;
6	import java.io.Writer;
7	import java.net.MalformedURLException;
8	import java.net.URL;
9	import java.net.URLEncoder;
10	import java.util.Date;
11
12	import javax.xml.parsers.DocumentBuilderFactory;
13	import javax.xml.transform.OutputKeys;
14	import javax.xml.transform.Result;
15	import javax.xml.transform.Transformer;
16	import javax.xml.transform.TransformerFactory;
17	import javax.xml.transform.dom.DOMSource;
18	import javax.xml.transform.stream.StreamResult;
19
20	import net.sf.antcontrib.property.URLEncodeTask;
21
22	import org.apache.log4j.Logger;
23	import org.w3c.dom.Document;
24	import org.w3c.dom.Element;
25
26	/**
27	* This class provides an easy way to list all pages of your site (including
28	* generates ones) ina sitemap.xml. See http://sitemaps.org/protocol.php for the
29	* XML specification.
30	*
31	* <br/>
32	* XML output is based on org.w3c.dom
33	*
34	*
35	* @author Stefan Tzeggai
36	*/
37	public class Sitemap {
38
39	static String[] SEARCHENGINES = new String[] {
40	"http://www.sitemapwriter.com/notify.php?crawler=all&url=",
41	"http://submissions.ask.com/ping?sitemap=",
42	"http://www.google.com/webmasters/tools/ping?sitemap=",
43	"http://api.moreover.com/ping?u=",
44	"http://www.bing.com/webmaster/ping.aspx?siteMap=",
45	"http://search.yahooapis.com/SiteExplorerService/V1/ping?sitemap=" };
46
47	final static Logger log = Logger.getLogger(Sitemap.class);
48
49	private Document document;
50
51	final private Element urlsetElement;
52
53	/**
54	* Returns the number of URLs. May not be more than 50,000 by spec.
55	*/
56	private int size = 0;
57
58	final static String NSURL = "http://www.sitemaps.org/schemas/sitemap/0.9";
59
60	public enum CHANGEFREQ {
61	monthly, daily, weekly
62	}
63
64	public Sitemap() {
65
66	try {
67
68	// Create a DOM builder and parse the fragment
69	final DocumentBuilderFactory factory = DocumentBuilderFactory
70	.newInstance();
71	document = factory.newDocumentBuilder().newDocument();
72
73	// XML root element
74	urlsetElement = document.createElementNS(NSURL, "urlset");
75	document.appendChild(urlsetElement);
76
77	// // Linking this XML to the AtlasML Schema
78	// final Attr namespaces = document.createAttributeNS(
79	// "http://www.w3.org/2001/XMLSchema-instance", "schemaLocation");
80	// namespaces
81	// .setValue(NSURL+" http://localhost:"
82	// + Webserver.DEFAULTPORT
83	// + "/skrueger/atlas/resource/AtlasML.xsd");
84	// urlset.setAttributeNode(namespaces);
85
86	} catch (Exception e) {
87	throw new RuntimeException(
88	"Sitemap org.w3c.xml newDocumentBuilder failed:", e);
89	}
90	}
91
92	/**
93	*
94	* @param locString
95	* @param lastmod
96	* when was the page modified the last time? May be
97	* <code>null</code>.
98	* @param changefreq
99	* How often does the page change. May be <code>null</code>.
100	* @param priority
101	* Priority 0.0 - 1.0. May be <code>null</code>.
102	* @return <code>true</code> if the {@link Sitemap} contains less than 50000
103	* urls and the page was added.
104	*/
105	public boolean addUrl(String locString, Date lastmod,
106	CHANGEFREQ changefreq, Double priority) {
107	size++;
108	if (size >= 50000)
109	return false;
110
111	Element urlElement = document.createElement("url");
112
113	if (locString == null)
114	throw new IllegalArgumentException("location must be provided");
115	else {
116	Element e = document.createElement("loc");
117	e.appendChild(document.createTextNode(locString));
118	urlElement.appendChild(e);
119	}
120
121	if (lastmod != null) {
122	Element e = document.createElement("lastmod");
123	e.appendChild(document.createTextNode(lastmod.toString()));
124	urlElement.appendChild(e);
125	}
126
127	if (changefreq != null) {
128	Element e = document.createElement("changefreq");
129	e.appendChild(document.createTextNode(changefreq.toString()));
130	urlElement.appendChild(e);
131	}
132
133	if (priority != null) {
134	Element e = document.createElement("priority");
135	e.appendChild(document.createTextNode(priority.toString()));
136	urlElement.appendChild(e);
137	}
138	urlsetElement.appendChild(urlElement);
139	return true;
140	}
141
142	public Document getDocument() {
143	return document;
144	}
145
146	public String getXmlString() {
147	StringWriter stringWriter = new StringWriter();
148	outputToWriter(stringWriter);
149	return stringWriter.toString();
150	}
151
152	/**
153	* Can be used to write to a File with {@link FileWriter}.
154	*/
155	public void outputToWriter(Writer stringWriter) {
156	try {
157
158	try { // close outputStreamWriter.close();
159
160	// ****************************************************************************
161	// Create the XML
162	// ****************************************************************************
163	final Result result = new StreamResult(stringWriter);
164
165	// with indenting to make it human-readable
166	final TransformerFactory tf = TransformerFactory.newInstance();
167
168	// TODO Ging mit xerces, geht nicht mehr mit xalan ?!
169	// tf.setAttribute("indent-number", new Integer(2));
170
171	final Transformer xformer = tf.newTransformer();
172	xformer.setOutputProperty(OutputKeys.INDENT, "yes");
173	xformer.setOutputProperty(
174	"{http://xml.apache.org/xalan}indent-amount", "2");
175
176	// Write the DOM document to the file
177	xformer.transform(new DOMSource(document), result);
178
179	} finally {
180	stringWriter.close();
181	}
182
183	} catch (Exception e) {
184	log.error("Failed to create sitemap.XML-String", e);
185	throw new RuntimeException(e);
186	}
187	}
188
189	public int getSize() {
190	return size;
191	}
192
193	public static void submitToSearchEngines(String urlToSitemap) {
194	for (String se : SEARCHENGINES) {
195	log.info("Submitting " + urlToSitemap + " to " + se);
196	try {
197	URL url = new URL(se + urlToSitemap);
198
199	urlToSitemap = URLEncoder.encode(urlToSitemap, "UTF8");
200
201	url.openConnection().connect();
202
203	} catch (Exception e) {
204	log.error("failed to submit " + urlToSitemap + " to " + se, e);
205	}
206	}
207	}
208	}
Name	Value
svn:eol-style	native
svn:keywords	Id URL
svn:mime-type	text/plain