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 $namespace Integer: the namespace to get the priority for
 225          * @return String
 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 $namespace Integer: the namespace to get the priority for
 237          * @return String
 238          */
 239         function guessPriority( $namespace ) {
 240                 return MWNamespace::isMain( $namespace ) ? $this->priorities[GS_MAIN] : $this->priorities[GS_TALK];
 241         }
 242
 243         /**
 244          * Return a database resolution of all the pages in a given namespace
 245          *
 246          * @param $namespace Integer: limit the query to this namespace
 247          * @return Resource
 248          */
 249         function getPageRes( $namespace ) {
 250                 return $this->dbr->select( 'page',
 251                         array(
 252                                 'page_namespace',
 253                                 'page_title',
 254                                 'page_touched',
 255                         ),
 256                         array( 'page_namespace' => $namespace ),
 257                         __METHOD__
 258                 );
 259         }
 260
 261         /**
 262          * Main loop
 263          */
 264         public function main() {
 265                 global $wgContLang;
 266
 267                 fwrite( $this->findex, $this->openIndex() );
 268
 269                 foreach ( $this->namespaces as $namespace ) {
 270                         $res = $this->getPageRes( $namespace );
 271                         $this->file = false;
 272                         $this->generateLimit( $namespace );
 273                         $length = $this->limit[0];
 274                         $i = $smcount = 0;
 275
 276                         $fns = $wgContLang->getFormattedNsText( $namespace );
 277                         $this->output( "$namespace ($fns)" );
 278                         foreach ( $res as $row ) {
 279                                 if ( $i++ === 0 || $i === $this->url_limit + 1 || $length + $this->limit[1] + $this->limit[2] > $this->size_limit ) {
 280                                         if ( $this->file !== false ) {
 281                                                 $this->write( $this->file, $this->closeFile() );
 282                                                 $this->close( $this->file );
 283                                         }
 284                                         $filename = $this->sitemapFilename( $namespace, $smcount++ );
 285                                         $this->file = $this->open( $this->fspath . $filename, 'wb' );
 286                                         $this->write( $this->file, $this->openFile() );
 287                                         fwrite( $this->findex, $this->indexEntry( $filename ) );
 288                                         $this->output( "\t$this->fspath$filename\n" );
 289                                         $length = $this->limit[0];
 290                                         $i = 1;
 291                                 }
 292                                 $title = Title::makeTitle( $row->page_namespace, $row->page_title );
 293                                 $date = wfTimestamp( TS_ISO_8601, $row->page_touched );
 294                                 $entry = $this->fileEntry( $title->getFullURL(), $date, $this->priority( $namespace ) );
 295                                 $length += strlen( $entry );
 296                                 $this->write( $this->file, $entry );
 297                                 // generate pages for language variants
 298                                 if ( $wgContLang->hasVariants() ) {
 299                                         $variants = $wgContLang->getVariants();
 300                                         foreach ( $variants as $vCode ) {
 301                                                 if ( $vCode == $wgContLang->getCode() ) continue; // we don't want default variant
 302                                                 $entry = $this->fileEntry( $title->getFullURL( '', $vCode ), $date, $this->priority( $namespace ) );
 303                                                 $length += strlen( $entry );
 304                                                 $this->write( $this->file, $entry );
 305                                         }
 306                                 }
 307                         }
 308                         if ( $this->file ) {
 309                                 $this->write( $this->file, $this->closeFile() );
 310                                 $this->close( $this->file );
 311                         }
 312                 }
 313                 fwrite( $this->findex, $this->closeIndex() );
 314                 fclose( $this->findex );
 315         }
 316
 317         /**
 318          * gzopen() / fopen() wrapper
 319          *
 320          * @return Resource
 321          */
 322         function open( $file, $flags ) {
 323                 return $this->compress ? gzopen( $file, $flags ) : fopen( $file, $flags );
 324         }
 325
 326         /**
 327          * gzwrite() / fwrite() wrapper
 328          */
 329         function write( &$handle, $str ) {
 330                 if ( $this->compress )
 331                         gzwrite( $handle, $str );
 332                 else
 333                         fwrite( $handle, $str );
 334         }
 335
 336         /**
 337          * gzclose() / fclose() wrapper
 338          */
 339         function close( &$handle ) {
 340                 if ( $this->compress )
 341                         gzclose( $handle );
 342                 else
 343                         fclose( $handle );
 344         }
 345
 346         /**
 347          * Get a sitemap filename
 348          *
 349          * @param $namespace Integer: the namespace
 350          * @param $count Integer: the count
 351          * @return String
 352          */
 353         function sitemapFilename( $namespace, $count ) {
 354                 $ext = $this->compress ? '.gz' : '';
 355                 return "sitemap-" . wfWikiID() . "-NS_$namespace-$count.xml$ext";
 356         }
 357
 358         /**
 359          * Return the XML required to open an XML file
 360          *
 361          * @return string
 362          */
 363         function xmlHead() {
 364                 return '<?xml version="1.0" encoding="UTF-8"?>' . "\n";
 365         }
 366
 367         /**
 368          * Return the XML schema being used
 369          *
 370          * @return String
 371          */
 372         function xmlSchema() {
 373                 return 'http://www.sitemaps.org/schemas/sitemap/0.9';
 374         }
 375
 376         /**
 377          * Return the XML required to open a sitemap index file
 378          *
 379          * @return String
 380          */
 381         function openIndex() {
 382                 return $this->xmlHead() . '<sitemapindex xmlns="' . $this->xmlSchema() . '">' . "\n";
 383         }
 384
 385         /**
 386          * Return the XML for a single sitemap indexfile entry
 387          *
 388          * @param $filename String: the filename of the sitemap file
 389          * @return String
 390          */
 391         function indexEntry( $filename ) {
 392                 return
 393                         "\t<sitemap>\n" .
 394                         "\t\t<loc>$filename</loc>\n" .
 395                         "\t\t<lastmod>{$this->timestamp}</lastmod>\n" .
 396                         "\t</sitemap>\n";
 397         }
 398
 399         /**
 400          * Return the XML required to close a sitemap index file
 401          *
 402          * @return String
 403          */
 404         function closeIndex() {
 405                 return "</sitemapindex>\n";
 406         }
 407
 408         /**
 409          * Return the XML required to open a sitemap file
 410          *
 411          * @return String
 412          */
 413         function openFile() {
 414                 return $this->xmlHead() . '<urlset xmlns="' . $this->xmlSchema() . '">' . "\n";
 415         }
 416
 417         /**
 418          * Return the XML for a single sitemap entry
 419          *
 420          * @param $url String: an RFC 2396 compliant URL
 421          * @param $date String: a ISO 8601 date
 422          * @param $priority String: a priority indicator, 0.0 - 1.0 inclusive with a 0.1 stepsize
 423          * @return String
 424          */
 425         function fileEntry( $url, $date, $priority ) {
 426                 return
 427                         "\t<url>\n" .
 428                         "\t\t<loc>$url</loc>\n" .
 429                         "\t\t<lastmod>$date</lastmod>\n" .
 430                         "\t\t<priority>$priority</priority>\n" .
 431                         "\t</url>\n";
 432         }
 433
 434         /**
 435          * Return the XML required to close sitemap file
 436          *
 437          * @return String
 438          */
 439         function closeFile() {
 440                 return "</urlset>\n";
 441         }
 442
 443         /**
 444          * Populate $this->limit
 445          */
 446         function generateLimit( $namespace ) {
 447                 $title = Title::makeTitle( $namespace, str_repeat( "\xf0\xa8\xae\x81", 63 ) . "\xe5\x96\x83" );
 448
 449                 $this->limit = array(
 450                         strlen( $this->openFile() ),
 451                         strlen( $this->fileEntry( $title->getFullUrl(), wfTimestamp( TS_ISO_8601, wfTimestamp() ), $this->priority( $namespace ) ) ),
 452                         strlen( $this->closeFile() )
 453                 );
 454         }
 455 }
 456
 457 $maintClass = "GenerateSitemap";
 458 require_once( DO_MAINTENANCE );