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.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 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);

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