maintenance/generateSitemap.php

   1 <?php
   2 define( 'GS_MAIN', -2 );
   3 define( 'GS_TALK', -1 );
   4 /**
   5  * Creates a sitemap for the site
   6  *
   7  * This program is free software; you can redistribute it and/or modify
   8  * it under the terms of the GNU General Public License as published by
   9  * the Free Software Foundation; either version 2 of the License, or
  10  * (at your option) any later version.
  11  *
  12  * This program is distributed in the hope that it will be useful,
  13  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  14  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
  15  * GNU General Public License for more details.
  16  *
  17  * You should have received a copy of the GNU General Public License along
  18  * with this program; if not, write to the Free Software Foundation, Inc.,
  19  * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
  20  * http://www.gnu.org/copyleft/gpl.html
  21  *
  22  * @ingroup Maintenance
  23  *
  24  * @copyright Copyright © 2005, Ævar Arnfjörð Bjarmason
  25  * @copyright Copyright © 2005, Jens Frank <jeluf@gmx.de>
  26  * @copyright Copyright © 2005, Brion Vibber <brion@pobox.com>
  27  *
  28  * @see http://www.sitemaps.org/
  29  * @see http://www.sitemaps.org/schemas/sitemap/0.9/sitemap.xsd
  30  *
  31  * @license http://www.gnu.org/copyleft/gpl.html GNU General Public License 2.0 or later
  32  */
  33
  34 require_once( dirname(__FILE__) . '/Maintenance.php' );
  35
  36 class GenerateSitemap extends Maintenance {
  37         /**
  38          * The maximum amount of urls in a sitemap file
  39          *
  40          * @link http://www.sitemaps.org/schemas/sitemap/0.9/sitemap.xsd
  41          *
  42          * @var int
  43          */
  44         var $url_limit;
  45
  46         /**
  47          * The maximum size of a sitemap file
  48          *
  49          * @link http://www.sitemaps.org/faq.php#faq_sitemap_size
  50          *
  51          * @var int
  52          */
  53         var $size_limit;
  54
  55         /**
  56          * The path to prepend to the filename
  57          *
  58          * @var string
  59          */
  60         var $fspath;
  61
  62         /**
  63          * The path to append to the domain name
  64          *
  65          * @var string
  66          */
  67         var $path;
  68
  69         /**
  70          * Whether or not to use compression
  71          *
  72          * @var bool
  73          */
  74         var $compress;
  75
  76         /**
  77          * The number of entries to save in each sitemap file
  78          *
  79          * @var array
  80          */
  81         var $limit = array();
  82
  83         /**
  84          * Key => value entries of namespaces and their priorities
  85          *
  86          * @var array
  87          */
  88         var $priorities = array();
  89
  90         /**
  91          * A one-dimensional array of namespaces in the wiki
  92          *
  93          * @var array
  94          */
  95         var $namespaces = array();
  96
  97         /**
  98          * When this sitemap batch was generated
  99          *
 100          * @var string
 101          */
 102         var $timestamp;
 103
 104         /**
 105          * A database slave object
 106          *
 107          * @var object
 108          */
 109         var $dbr;
 110
 111         /**
 112          * A resource pointing to the sitemap index file
 113          *
 114          * @var resource
 115          */
 116         var $findex;
 117
 118
 119         /**
 120          * A resource pointing to a sitemap file
 121          *
 122          * @var resource
 123          */
 124         var $file;
 125
 126         /**
 127          * Constructor
 128          */
 129         public function __construct() {
 130                 parent::__construct();
 131                 $this->mDescription = "Creates a sitemap for the site";
 132                 $this->addOption( 'fspath', 'The file system path to save to, e.g. /tmp/sitemap' .
 133                                                                         "\n\t\tdefaults to current directory", false, true );
 134                 $this->addOption( 'server', "The protocol and server name to use in URLs, e.g.\n" .
 135                                                                         "\t\thttp://en.wikipedia.org. This is sometimes necessary because\n" .
 136                                                                         "\t\tserver name detection may fail in command line scripts.", false, true );
 137                 $this->addOption( 'compress', 'Compress the sitemap files, can take value yes|no, default yes', false, true );
 138         }
 139
 140         /**
 141          * Execute
 142          */
 143         public function execute() {
 144                 global $wgScriptPath;
 145                 $this->setNamespacePriorities();
 146                 $this->url_limit = 50000;
 147                 $this->size_limit = pow( 2, 20 ) * 10;
 148                 $this->fspath = self::init_path( $this->getOption( 'fspath', getcwd() ) );
 149                 $this->compress = $this->getOption( 'compress', 'yes' ) !== 'no';
 150                 $this->dbr = wfGetDB( DB_SLAVE );
 151                 $this->generateNamespaces();
 152                 $this->timestamp = wfTimestamp( TS_ISO_8601, wfTimestampNow() );
 153                 $this->findex = fopen( "{$this->fspath}sitemap-index-" . wfWikiID() . ".xml", 'wb' );
 154                 $this->main();
 155         }
 156
 157         private function setNamespacePriorities() {
 158                 // Custom main namespaces
 159                 $this->priorities[GS_MAIN] = '0.5';
 160                 // Custom talk namesspaces
 161                 $this->priorities[GS_TALK] = '0.1';
 162                 // MediaWiki standard namespaces
 163                 $this->priorities[NS_MAIN] = '1.0';
 164                 $this->priorities[NS_TALK] = '0.1';
 165                 $this->priorities[NS_USER] = '0.5';
 166                 $this->priorities[NS_USER_TALK] = '0.1';
 167                 $this->priorities[NS_PROJECT] = '0.5';
 168                 $this->priorities[NS_PROJECT_TALK] = '0.1';
 169                 $this->priorities[NS_FILE] = '0.5';
 170                 $this->priorities[NS_FILE_TALK] = '0.1';
 171                 $this->priorities[NS_MEDIAWIKI] = '0.0';
 172                 $this->priorities[NS_MEDIAWIKI_TALK] = '0.1';
 173                 $this->priorities[NS_TEMPLATE] = '0.0';
 174                 $this->priorities[NS_TEMPLATE_TALK] = '0.1';
 175                 $this->priorities[NS_HELP] = '0.5';
 176                 $this->priorities[NS_HELP_TALK] = '0.1';
 177                 $this->priorities[NS_CATEGORY] = '0.5';
 178                 $this->priorities[NS_CATEGORY_TALK] = '0.1';
 179         }
 180
 181         /**
 182          * Create directory if it does not exist and return pathname with a trailing slash
 183          */
 184         private static function init_path( $fspath ) {
 185                 if( !isset( $fspath ) ) {
 186                         return null;
 187                 }
 188                 # Create directory if needed
 189                 if( $fspath && !is_dir( $fspath ) ) {
 190                         wfMkdirParents( $fspath ) or die("Can not create directory $fspath.\n");
 191                 }
 192
 193                 return realpath( $fspath ). DIRECTORY_SEPARATOR ;
 194         }
 195
 196         /**
 197          * Generate a one-dimensional array of existing namespaces
 198          */
 199         function generateNamespaces() {
 200                 // Only generate for specific namespaces if $wgSitemapNamespaces is an array.
 201                 global $wgSitemapNamespaces;
 202                 if( is_array( $wgSitemapNamespaces ) ) {
 203                         $this->namespaces = $wgSitemapNamespaces;
 204                         return;
 205                 }
 206
 207                 $res = $this->dbr->select( 'page',
 208                         array( 'page_namespace' ),
 209                         array(),
 210                         __METHOD__,
 211                         array(
 212                                 'GROUP BY' => 'page_namespace',
 213                                 'ORDER BY' => 'page_namespace',
 214                         )
 215                 );
 216
 217                 foreach ( $res as $row )
 218                         $this->namespaces[] = $row->page_namespace;
 219         }
 220
 221         /**
 222          * Get the priority of a given namespace
 223          *
 224          * @param int $namespace The namespace to get the priority for
 225          +
 226          * @return string
 227          */
 228
 229         function priority( $namespace ) {
 230                 return isset( $this->priorities[$namespace] ) ? $this->priorities[$namespace] : $this->guessPriority( $namespace );
 231         }
 232
 233         /**
 234          * If the namespace isn't listed on the priority list return the
 235          * default priority for the namespace, varies depending on whether it's
 236          * a talkpage or not.
 237          *
 238          * @param int $namespace The namespace to get the priority for
 239          *
 240          * @return string
 241          */
 242         function guessPriority( $namespace ) {
 243                 return MWNamespace::isMain( $namespace ) ? $this->priorities[GS_MAIN] : $this->priorities[GS_TALK];
 244         }
 245
 246         /**
 247          * Return a database resolution of all the pages in a given namespace
 248          *
 249          * @param int $namespace Limit the query to this namespace
 250          *
 251          * @return resource
 252          */
 253         function getPageRes( $namespace ) {
 254                 return $this->dbr->select( 'page',
 255                         array(
 256                                 'page_namespace',
 257                                 'page_title',
 258                                 'page_touched',
 259                         ),
 260                         array( 'page_namespace' => $namespace ),
 261                         __METHOD__
 262                 );
 263         }
 264
 265         /**
 266          * Main loop
 267          *
 268          * @access public
 269          */
 270         function main() {
 271                 global $wgContLang;
 272
 273                 fwrite( $this->findex, $this->openIndex() );
 274
 275                 foreach ( $this->namespaces as $namespace ) {
 276                         $res = $this->getPageRes( $namespace );
 277                         $this->file = false;
 278                         $this->generateLimit( $namespace );
 279                         $length = $this->limit[0];
 280                         $i = $smcount = 0;
 281
 282                         $fns = $wgContLang->getFormattedNsText( $namespace );
 283                         $this->output( "$namespace ($fns)" );
 284                         foreach ( $res as $row ) {
 285                                 if ( $i++ === 0 || $i === $this->url_limit + 1 || $length + $this->limit[1] + $this->limit[2] > $this->size_limit ) {
 286                                         if ( $this->file !== false ) {
 287                                                 $this->write( $this->file, $this->closeFile() );
 288                                                 $this->close( $this->file );
 289                                         }
 290                                         $filename = $this->sitemapFilename( $namespace, $smcount++ );
 291                                         $this->file = $this->open( $this->fspath . $filename, 'wb' );
 292                                         $this->write( $this->file, $this->openFile() );
 293                                         fwrite( $this->findex, $this->indexEntry( $filename ) );
 294                                         $this->output( "\t$this->fspath$filename\n" );
 295                                         $length = $this->limit[0];
 296                                         $i = 1;
 297                                 }
 298                                 $title = Title::makeTitle( $row->page_namespace, $row->page_title );
 299                                 $date = wfTimestamp( TS_ISO_8601, $row->page_touched );
 300                                 $entry = $this->fileEntry( $title->getFullURL(), $date, $this->priority( $namespace ) );
 301                                 $length += strlen( $entry );
 302                                 $this->write( $this->file, $entry );
 303                                 // generate pages for language variants
 304                                 if($wgContLang->hasVariants()){
 305                                         $variants = $wgContLang->getVariants();
 306                                         foreach($variants as $vCode){
 307                                                 if($vCode==$wgContLang->getCode()) continue; // we don't want default variant
 308                                                 $entry = $this->fileEntry( $title->getFullURL('',$vCode), $date, $this->priority( $namespace ) );
 309                                                 $length += strlen( $entry );
 310                                                 $this->write( $this->file, $entry );
 311                                         }
 312                                 }
 313                         }
 314                         if ( $this->file ) {
 315                                 $this->write( $this->file, $this->closeFile() );
 316                                 $this->close( $this->file );
 317                         }
 318                 }
 319                 fwrite( $this->findex, $this->closeIndex() );
 320                 fclose( $this->findex );
 321         }
 322
 323         /**
 324          * gzopen() / fopen() wrapper
 325          *
 326          * @return resource
 327          */
 328         function open( $file, $flags ) {
 329                 return $this->compress ? gzopen( $file, $flags ) : fopen( $file, $flags );
 330         }
 331
 332         /**
 333          * gzwrite() / fwrite() wrapper
 334          */
 335         function write( &$handle, $str ) {
 336                 if ( $this->compress )
 337                         gzwrite( $handle, $str );
 338                 else
 339                         fwrite( $handle, $str );
 340         }
 341
 342         /**
 343          * gzclose() / fclose() wrapper
 344          */
 345         function close( &$handle ) {
 346                 if ( $this->compress )
 347                         gzclose( $handle );
 348                 else
 349                         fclose( $handle );
 350         }
 351
 352         /**
 353          * Get a sitemap filename
 354          *
 355          * @static
 356          *
 357          * @param int $namespace The namespace
 358          * @param int $count The count
 359          *
 360          * @return string
 361          */
 362         function sitemapFilename( $namespace, $count ) {
 363                 $ext = $this->compress ? '.gz' : '';
 364                 return "sitemap-".wfWikiID()."-NS_$namespace-$count.xml$ext";
 365         }
 366
 367         /**
 368          * Return the XML required to open an XML file
 369          *
 370          * @static
 371          *
 372          * @return string
 373          */
 374         function xmlHead() {
 375                 return '<?xml version="1.0" encoding="UTF-8"?>' . "\n";
 376         }
 377
 378         /**
 379          * Return the XML schema being used
 380          *
 381          * @static
 382          *
 383          * @returns string
 384          */
 385         function xmlSchema() {
 386                 return 'http://www.sitemaps.org/schemas/sitemap/0.9';
 387         }
 388
 389         /**
 390          * Return the XML required to open a sitemap index file
 391          *
 392          * @return string
 393          */
 394         function openIndex() {
 395                 return $this->xmlHead() . '<sitemapindex xmlns="' . $this->xmlSchema() . '">' . "\n";
 396         }
 397
 398         /**
 399          * Return the XML for a single sitemap indexfile entry
 400          *
 401          * @static
 402          *
 403          * @param string $filename The filename of the sitemap file
 404          *
 405          * @return string
 406          */
 407         function indexEntry( $filename ) {
 408                 return
 409                         "\t<sitemap>\n" .
 410                         "\t\t<loc>$filename</loc>\n" .
 411                         "\t\t<lastmod>{$this->timestamp}</lastmod>\n" .
 412                         "\t</sitemap>\n";
 413         }
 414
 415         /**
 416          * Return the XML required to close a sitemap index file
 417          *
 418          * @static
 419          *
 420          * @return string
 421          */
 422         function closeIndex() {
 423                 return "</sitemapindex>\n";
 424         }
 425
 426         /**
 427          * Return the XML required to open a sitemap file
 428          *
 429          * @return string
 430          */
 431         function openFile() {
 432                 return $this->xmlHead() . '<urlset xmlns="' . $this->xmlSchema() . '">' . "\n";
 433         }
 434
 435         /**
 436          * Return the XML for a single sitemap entry
 437          *
 438          * @static
 439          *
 440          * @param string $url An RFC 2396 compliant URL
 441          * @param string $date A ISO 8601 date
 442          * @param string $priority A priority indicator, 0.0 - 1.0 inclusive with a 0.1 stepsize
 443          *
 444          * @return string
 445          */
 446         function fileEntry( $url, $date, $priority ) {
 447                 return
 448                         "\t<url>\n" .
 449                         "\t\t<loc>$url</loc>\n" .
 450                         "\t\t<lastmod>$date</lastmod>\n" .
 451                         "\t\t<priority>$priority</priority>\n" .
 452                         "\t</url>\n";
 453         }
 454
 455         /**
 456          * Return the XML required to close sitemap file
 457          *
 458          * @static
 459          * @return string
 460          */
 461         function closeFile() {
 462                 return "</urlset>\n";
 463         }
 464
 465         /**
 466          * Populate $this->limit
 467          */
 468         function generateLimit( $namespace ) {
 469                 $title = Title::makeTitle( $namespace, str_repeat( "\xf0\xa8\xae\x81", 63 ) . "\xe5\x96\x83" );
 470
 471                 $this->limit = array(
 472                         strlen( $this->openFile() ),
 473                         strlen( $this->fileEntry( $title->getFullUrl(), wfTimestamp( TS_ISO_8601, wfTimestamp() ), $this->priority( $namespace ) ) ),
 474                         strlen( $this->closeFile() )
 475                 );
 476         }
 477 }
 478
 479 $maintClass = "GenerateSitemap";
 480 require_once( DO_MAINTENANCE );