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( "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                 // Custom main namespaces
  90                 GS_MAIN                 => '0.5',
  91                 // Custom talk namesspaces
  92                 GS_TALK                 => '0.1',
  93                 // MediaWiki standard namespaces
  94                 NS_MAIN                 => '1.0',
  95                 NS_TALK                 => '0.1',
  96                 NS_USER                 => '0.5',
  97                 NS_USER_TALK            => '0.1',
  98                 NS_PROJECT              => '0.5',
  99                 NS_PROJECT_TALK         => '0.1',
 100                 NS_FILE                 => '0.5',
 101                 NS_FILE_TALK            => '0.1',
 102                 NS_MEDIAWIKI            => '0.0',
 103                 NS_MEDIAWIKI_TALK       => '0.1',
 104                 NS_TEMPLATE             => '0.0',
 105                 NS_TEMPLATE_TALK        => '0.1',
 106                 NS_HELP                 => '0.5',
 107                 NS_HELP_TALK            => '0.1',
 108                 NS_CATEGORY             => '0.5',
 109                 NS_CATEGORY_TALK        => '0.1',
 110         );
 111
 112         /**
 113          * A one-dimensional array of namespaces in the wiki
 114          *
 115          * @var array
 116          */
 117         var $namespaces = array();
 118
 119         /**
 120          * When this sitemap batch was generated
 121          *
 122          * @var string
 123          */
 124         var $timestamp;
 125
 126         /**
 127          * A database slave object
 128          *
 129          * @var object
 130          */
 131         var $dbr;
 132
 133         /**
 134          * A resource pointing to the sitemap index file
 135          *
 136          * @var resource
 137          */
 138         var $findex;
 139
 140
 141         /**
 142          * A resource pointing to a sitemap file
 143          *
 144          * @var resource
 145          */
 146         var $file;
 147
 148         /**
 149          * Constructor
 150          */
 151         public function __construct() {
 152                 parent::__construct();
 153                 $this->mDescription = "Creates a sitemap for the site";
 154                 $this->addOption( 'fspath', 'The file system path to save to, e.g. /tmp/sitemap' .
 155                                                                         "\n\t\tdefaults to current directory", false, true );
 156                 $this->addOption( 'server', "The protocol and server name to use in URLs, e.g.\n" .
 157                                                                         "\t\thttp://en.wikipedia.org. This is sometimes necessary because\n" .
 158                                                                         "\t\tserver name detection may fail in command line scripts.", false, true );
 159                 $this->addOption( 'compress', 'Compress the sitemap files, can take value yes|no, default yes' );
 160         }
 161
 162         /**
 163          * Execute
 164          */
 165         public function execute() {
 166                 global $wgScriptPath;
 167
 168                 $this->url_limit = 50000;
 169                 $this->size_limit = pow( 2, 20 ) * 10;
 170                 $this->fspath = self::init_path( $this->getOption( 'fspath', getcwd() ) );
 171                 $this->compress = $this->getOption( 'compress', 'yes' ) !== 'no';
 172                 $this->dbr = wfGetDB( DB_SLAVE );
 173                 $this->generateNamespaces();
 174                 $this->timestamp = wfTimestamp( TS_ISO_8601, wfTimestampNow() );
 175                 $this->findex = fopen( "{$this->fspath}sitemap-index-" . wfWikiID() . ".xml", 'wb' );
 176                 $this->main();
 177         }
 178
 179         /**
 180          * Create directory if it does not exist and return pathname with a trailing slash
 181          */
 182         private static function init_path( $fspath ) {
 183                 if( !isset( $fspath ) ) {
 184                         return null;
 185                 }
 186                 # Create directory if needed
 187                 if( $fspath && !is_dir( $fspath ) ) {
 188                         wfMkdirParents( $fspath ) or die("Can not create directory $fspath.\n");
 189                 }
 190
 191                 return realpath( $fspath ). DIRECTORY_SEPARATOR ;
 192         }
 193
 194         /**
 195          * Generate a one-dimensional array of existing namespaces
 196          */
 197         function generateNamespaces() {
 198                 // Only generate for specific namespaces if $wgSitemapNamespaces is an array.
 199                 global $wgSitemapNamespaces;
 200                 if( is_array( $wgSitemapNamespaces ) ) {
 201                         $this->namespaces = $wgSitemapNamespaces;
 202                         return;
 203                 }
 204
 205                 $res = $this->dbr->select( 'page',
 206                         array( 'page_namespace' ),
 207                         array(),
 208                         __METHOD__,
 209                         array(
 210                                 'GROUP BY' => 'page_namespace',
 211                                 'ORDER BY' => 'page_namespace',
 212                         )
 213                 );
 214
 215                 while ( $row = $this->dbr->fetchObject( $res ) )
 216                         $this->namespaces[] = $row->page_namespace;
 217         }
 218
 219         /**
 220          * Get the priority of a given namespace
 221          *
 222          * @param int $namespace The namespace to get the priority for
 223          +
 224          * @return string
 225          */
 226
 227         function priority( $namespace ) {
 228                 return isset( $this->priorities[$namespace] ) ? $this->priorities[$namespace] : $this->guessPriority( $namespace );
 229         }
 230
 231         /**
 232          * If the namespace isn't listed on the priority list return the
 233          * default priority for the namespace, varies depending on whether it's
 234          * a talkpage or not.
 235          *
 236          * @param int $namespace The namespace to get the priority for
 237          *
 238          * @return string
 239          */
 240         function guessPriority( $namespace ) {
 241                 return MWNamespace::isMain( $namespace ) ? $this->priorities[GS_MAIN] : $this->priorities[GS_TALK];
 242         }
 243
 244         /**
 245          * Return a database resolution of all the pages in a given namespace
 246          *
 247          * @param int $namespace Limit the query to this namespace
 248          *
 249          * @return resource
 250          */
 251         function getPageRes( $namespace ) {
 252                 return $this->dbr->select( 'page',
 253                         array(
 254                                 'page_namespace',
 255                                 'page_title',
 256                                 'page_touched',
 257                         ),
 258                         array( 'page_namespace' => $namespace ),
 259                         __METHOD__
 260                 );
 261         }
 262
 263         /**
 264          * Main loop
 265          *
 266          * @access public
 267          */
 268         function main() {
 269                 global $wgContLang;
 270
 271                 fwrite( $this->findex, $this->openIndex() );
 272
 273                 foreach ( $this->namespaces as $namespace ) {
 274                         $res = $this->getPageRes( $namespace );
 275                         $this->file = false;
 276                         $this->generateLimit( $namespace );
 277                         $length = $this->limit[0];
 278                         $i = $smcount = 0;
 279
 280                         $fns = $wgContLang->getFormattedNsText( $namespace );
 281                         $this->output( "$namespace ($fns)" );
 282                         while ( $row = $this->dbr->fetchObject( $res ) ) {
 283                                 if ( $i++ === 0 || $i === $this->url_limit + 1 || $length + $this->limit[1] + $this->limit[2] > $this->size_limit ) {
 284                                         if ( $this->file !== false ) {
 285                                                 $this->write( $this->file, $this->closeFile() );
 286                                                 $this->close( $this->file );
 287                                         }
 288                                         $filename = $this->sitemapFilename( $namespace, $smcount++ );
 289                                         $this->file = $this->open( $this->fspath . $filename, 'wb' );
 290                                         $this->write( $this->file, $this->openFile() );
 291                                         fwrite( $this->findex, $this->indexEntry( $filename ) );
 292                                         $this->output( "\t$this->fspath$filename" );
 293                                         $length = $this->limit[0];
 294                                         $i = 1;
 295                                 }
 296                                 $title = Title::makeTitle( $row->page_namespace, $row->page_title );
 297                                 $date = wfTimestamp( TS_ISO_8601, $row->page_touched );
 298                                 $entry = $this->fileEntry( $title->getFullURL(), $date, $this->priority( $namespace ) );
 299                                 $length += strlen( $entry );
 300                                 $this->write( $this->file, $entry );
 301                                 // generate pages for language variants
 302                                 if($wgContLang->hasVariants()){
 303                                         $variants = $wgContLang->getVariants();
 304                                         foreach($variants as $vCode){
 305                                                 if($vCode==$wgContLang->getCode()) continue; // we don't want default variant
 306                                                 $entry = $this->fileEntry( $title->getFullURL('',$vCode), $date, $this->priority( $namespace ) );
 307                                                 $length += strlen( $entry );
 308                                                 $this->write( $this->file, $entry );
 309                                         }
 310                                 }
 311                         }
 312                         if ( $this->file ) {
 313                                 $this->write( $this->file, $this->closeFile() );
 314                                 $this->close( $this->file );
 315                         }
 316                 }
 317                 fwrite( $this->findex, $this->closeIndex() );
 318                 fclose( $this->findex );
 319         }
 320
 321         /**
 322          * gzopen() / fopen() wrapper
 323          *
 324          * @return resource
 325          */
 326         function open( $file, $flags ) {
 327                 return $this->compress ? gzopen( $file, $flags ) : fopen( $file, $flags );
 328         }
 329
 330         /**
 331          * gzwrite() / fwrite() wrapper
 332          */
 333         function write( &$handle, $str ) {
 334                 if ( $this->compress )
 335                         gzwrite( $handle, $str );
 336                 else
 337                         fwrite( $handle, $str );
 338         }
 339
 340         /**
 341          * gzclose() / fclose() wrapper
 342          */
 343         function close( &$handle ) {
 344                 if ( $this->compress )
 345                         gzclose( $handle );
 346                 else
 347                         fclose( $handle );
 348         }
 349
 350         /**
 351          * Get a sitemap filename
 352          *
 353          * @static
 354          *
 355          * @param int $namespace The namespace
 356          * @param int $count The count
 357          *
 358          * @return string
 359          */
 360         function sitemapFilename( $namespace, $count ) {
 361                 $ext = $this->compress ? '.gz' : '';
 362                 return "sitemap-".wfWikiID()."-NS_$namespace-$count.xml$ext";
 363         }
 364
 365         /**
 366          * Return the XML required to open an XML file
 367          *
 368          * @static
 369          *
 370          * @return string
 371          */
 372         function xmlHead() {
 373                 return '<?xml version="1.0" encoding="UTF-8"?>' . "\n";
 374         }
 375
 376         /**
 377          * Return the XML schema being used
 378          *
 379          * @static
 380          *
 381          * @returns string
 382          */
 383         function xmlSchema() {
 384                 return 'http://www.sitemaps.org/schemas/sitemap/0.9';
 385         }
 386
 387         /**
 388          * Return the XML required to open a sitemap index file
 389          *
 390          * @return string
 391          */
 392         function openIndex() {
 393                 return $this->xmlHead() . '<sitemapindex xmlns="' . $this->xmlSchema() . '">' . "\n";
 394         }
 395
 396         /**
 397          * Return the XML for a single sitemap indexfile entry
 398          *
 399          * @static
 400          *
 401          * @param string $filename The filename of the sitemap file
 402          *
 403          * @return string
 404          */
 405         function indexEntry( $filename ) {
 406                 return
 407                         "\t<sitemap>\n" .
 408                         "\t\t<loc>$filename</loc>\n" .
 409                         "\t\t<lastmod>{$this->timestamp}</lastmod>\n" .
 410                         "\t</sitemap>\n";
 411         }
 412
 413         /**
 414          * Return the XML required to close a sitemap index file
 415          *
 416          * @static
 417          *
 418          * @return string
 419          */
 420         function closeIndex() {
 421                 return "</sitemapindex>\n";
 422         }
 423
 424         /**
 425          * Return the XML required to open a sitemap file
 426          *
 427          * @return string
 428          */
 429         function openFile() {
 430                 return $this->xmlHead() . '<urlset xmlns="' . $this->xmlSchema() . '">' . "\n";
 431         }
 432
 433         /**
 434          * Return the XML for a single sitemap entry
 435          *
 436          * @static
 437          *
 438          * @param string $url An RFC 2396 compliant URL
 439          * @param string $date A ISO 8601 date
 440          * @param string $priority A priority indicator, 0.0 - 1.0 inclusive with a 0.1 stepsize
 441          *
 442          * @return string
 443          */
 444         function fileEntry( $url, $date, $priority ) {
 445                 return
 446                         "\t<url>\n" .
 447                         "\t\t<loc>$url</loc>\n" .
 448                         "\t\t<lastmod>$date</lastmod>\n" .
 449                         "\t\t<priority>$priority</priority>\n" .
 450                         "\t</url>\n";
 451         }
 452
 453         /**
 454          * Return the XML required to close sitemap file
 455          *
 456          * @static
 457          * @return string
 458          */
 459         function closeFile() {
 460                 return "</urlset>\n";
 461         }
 462
 463         /**
 464          * Populate $this->limit
 465          */
 466         function generateLimit( $namespace ) {
 467                 $title = Title::makeTitle( $namespace, str_repeat( "\xf0\xa8\xae\x81", 63 ) . "\xe5\x96\x83" );
 468
 469                 $this->limit = array(
 470                         strlen( $this->openFile() ),
 471                         strlen( $this->fileEntry( $title->getFullUrl(), wfTimestamp( TS_ISO_8601, wfTimestamp() ), $this->priority( $namespace ) ) ),
 472                         strlen( $this->closeFile() )
 473                 );
 474         }
 475 }
 476
 477 $maintClass = "GenerateSitemap";
 478 require_once( DO_MAINTENANCE );