Revision 7764
Added by Matt Jones almost 12 years ago
| src/edu/ucsb/nceas/metacat/Sitemap.java | ||
|---|---|---|
| 25 | 25 |
package edu.ucsb.nceas.metacat; |
| 26 | 26 |
|
| 27 | 27 |
import java.io.File; |
| 28 |
import java.io.FileOutputStream; |
|
| 28 | 29 |
import java.io.FileWriter; |
| 29 | 30 |
import java.io.IOException; |
| 31 |
import java.io.OutputStreamWriter; |
|
| 30 | 32 |
import java.io.Writer; |
| 33 |
import java.nio.charset.Charset; |
|
| 31 | 34 |
import java.sql.PreparedStatement; |
| 32 | 35 |
import java.sql.ResultSet; |
| 33 | 36 |
import java.sql.SQLException; |
| ... | ... | |
| 131 | 134 |
// Loop through all of the documents, and write them to a |
| 132 | 135 |
// sitemap |
| 133 | 136 |
File sitemapFile = null; |
| 134 |
FileWriter sitemap = null;
|
|
| 137 |
OutputStreamWriter sitemap = null;
|
|
| 135 | 138 |
int counter = 0; |
| 136 | 139 |
int fileNumber = 0; |
| 137 | 140 |
while (rs.next()) {
|
| ... | ... | |
| 148 | 151 |
fileNumber++; |
| 149 | 152 |
sitemapFile = new File(directory, fileRoot + fileNumber |
| 150 | 153 |
+ ".xml"); |
| 151 |
sitemap = new FileWriter(sitemapFile);
|
|
| 154 |
sitemap = new OutputStreamWriter(new FileOutputStream(sitemapFile), Charset.forName("UTF-8"));
|
|
| 152 | 155 |
|
| 153 | 156 |
// Write the sitemap document header for the new file |
| 154 | 157 |
writeSitemapHeader(sitemap); |
| ... | ... | |
| 162 | 165 |
} |
| 163 | 166 |
stmt.close(); |
| 164 | 167 |
writeSitemapFooter(sitemap); |
| 168 |
writeSitemapIndex(fileNumber); |
|
| 165 | 169 |
} catch (SQLException e) {
|
| 166 | 170 |
logMetacat.warn("Error while writing to the sitemap file: "
|
| 167 | 171 |
+ e.getMessage()); |
| ... | ... | |
| 195 | 199 |
String header = "<urlset xmlns=\"http://www.sitemaps.org/schemas/sitemap/0.9\"\n" + |
| 196 | 200 |
"xmlns:sm=\"http://www.sitemaps.org/schemas/sitemap/0.9\"\n" + |
| 197 | 201 |
"xmlns:xsi=\"http://www.w3.org/2001/XMLSchema-instance\"\n" + |
| 198 |
"xsi:schemaLocation=\"http://www.sitemaps.org/schemas/sitemap/0.9 http://www.sitemaps.org/schemas/sitemap/0.9/sitemap.xsd\">"; |
|
| 202 |
"xsi:schemaLocation=\"http://www.sitemaps.org/schemas/sitemap/0.9 http://www.sitemaps.org/schemas/sitemap/0.9/sitemap.xsd\">\n";
|
|
| 199 | 203 |
|
| 200 | 204 |
sitemap.write(header); |
| 201 | 205 |
sitemap.flush(); |
| ... | ... | |
| 259 | 263 |
} |
| 260 | 264 |
} |
| 261 | 265 |
|
| 266 |
/** |
|
| 267 |
* Create an index file listing all of the sitemap files that were created. |
|
| 268 |
* @param fileNumber the number of sitemap files that were created. |
|
| 269 |
*/ |
|
| 270 |
private void writeSitemapIndex(int fileNumber) {
|
|
| 271 |
|
|
| 272 |
// Open a new sitemapIndex file for writing |
|
| 273 |
File sitemapIndexFile = null; |
|
| 274 |
OutputStreamWriter sitemapIndex = null; |
|
| 275 |
sitemapIndexFile = new File(directory, indexFilename); |
|
| 276 |
try {
|
|
| 277 |
sitemapIndex = new OutputStreamWriter(new FileOutputStream(sitemapIndexFile), Charset.forName("UTF-8"));
|
|
| 278 |
|
|
| 279 |
// Write the sitemap index header for the new file |
|
| 280 |
sitemapIndex.write(PROLOG); |
|
| 281 |
String header = "<sitemapindex xmlns=\"http://www.sitemaps.org/schemas/sitemap/0.9\"\n" |
|
| 282 |
+ "xmlns:sm=\"http://www.sitemaps.org/schemas/sitemap/0.9\"\n" + "xmlns:xsi=\"http://www.w3.org/2001/XMLSchema-instance\"\n" |
|
| 283 |
+ "xsi:schemaLocation=\"http://www.sitemaps.org/schemas/sitemap/0.9 http://www.sitemaps.org/schemas/sitemap/0.9/siteindex.xsd\">\n"; |
|
| 284 |
sitemapIndex.write(header); |
|
| 285 |
sitemapIndex.flush(); |
|
| 286 |
|
|
| 287 |
// Write out one index entry for each sitemap file |
|
| 288 |
for (int fn = 1; fn <= fileNumber; fn++) {
|
|
| 289 |
String filename = fileRoot + fileNumber + ".xml"; |
|
| 290 |
writeSitemapIndexEntry(sitemapIndex, filename); |
|
| 291 |
} |
|
| 292 |
|
|
| 293 |
// Write the sitemap index footer content |
|
| 294 |
if (sitemapIndex != null) {
|
|
| 295 |
String footer = "</sitemapindex>\n"; |
|
| 296 |
sitemapIndex.write(footer); |
|
| 297 |
sitemapIndex.close(); |
|
| 298 |
} |
|
| 299 |
|
|
| 300 |
// Close the index file |
|
| 301 |
if (sitemapIndex != null) {
|
|
| 302 |
sitemapIndex.close(); |
|
| 303 |
} |
|
| 304 |
|
|
| 305 |
} catch (IOException e) {
|
|
| 306 |
logMetacat.warn("Could not open or write to the sitemap index file." + e.getMessage());
|
|
| 307 |
} |
|
| 308 |
} |
|
| 309 |
|
|
| 310 |
/** |
|
| 311 |
* Write a single line of the sitemap index file containing the URL to a specific sitemap file. |
|
| 312 |
* @param sitemapIndex the writer to which the index information is written |
|
| 313 |
* @param filename the name of the index file to be used |
|
| 314 |
* @throws IOException on error writing to the index file |
|
| 315 |
*/ |
|
| 316 |
private void writeSitemapIndexEntry(Writer sitemapIndex, String filename) |
|
| 317 |
throws IOException {
|
|
| 318 |
if (sitemapIndex != null && filename != null && urlRoot != null) {
|
|
| 319 |
StringBuffer url = new StringBuffer(); |
|
| 320 |
url.append(urlRoot); |
|
| 321 |
if (!urlRoot.endsWith("/")) {
|
|
| 322 |
url.append("/");
|
|
| 323 |
} |
|
| 324 |
url.append(filename); |
|
| 325 |
sitemapIndex.write("<sitemap><loc>");
|
|
| 326 |
sitemapIndex.write(url.toString()); |
|
| 327 |
sitemapIndex.write("</loc>");
|
|
| 328 |
sitemapIndex.write("</sitemap>");
|
|
| 329 |
sitemapIndex.write("\n");
|
|
| 330 |
sitemapIndex.flush(); |
|
| 331 |
} |
|
| 332 |
} |
|
| 333 |
|
|
| 262 | 334 |
// Member variables |
| 263 | 335 |
|
| 264 | 336 |
/** The directory in which sitemaps are written. */ |
| ... | ... | |
| 275 | 347 |
|
| 276 | 348 |
/** The root name to be used in naming sitemap files. */ |
| 277 | 349 |
static final String fileRoot = "metacat"; |
| 350 |
|
|
| 351 |
/** The name to give to the sitemap index file */ |
|
| 352 |
static final String indexFilename = "metacatSitemapIndex.xml"; |
|
| 278 | 353 |
|
| 279 | 354 |
/** A String constant containing the XML prolog to be written in files. */ |
| 280 | 355 |
static final String PROLOG = "<?xml version=\"1.0\" encoding=\"UTF-8\" ?>\n"; |
Also available in: Unified diff
Mofdified Sitemap class to also generate the sitemap index file that is needed when more than one sitemap file is provided.