maintenance/generateSitemap.php

   1 <?php
   2 /**
   3  * Creates a sitemap for the site
   4  *
   5  * Copyright © 2005, Ævar Arnfjörð Bjarmason, Jens Frank <jeluf@gmx.de> and
   6  * Brion Vibber <brion@pobox.com>
   7  *
   8  * This program is free software; you can redistribute it and/or modify
   9  * it under the terms of the GNU General Public License as published by
  10  * the Free Software Foundation; either version 2 of the License, or
  11  * (at your option) any later version.
  12  *
  13  * This program is distributed in the hope that it will be useful,
  14  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  15  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
  16  * GNU General Public License for more details.
  17  *
  18  * You should have received a copy of the GNU General Public License along
  19  * with this program; if not, write to the Free Software Foundation, Inc.,
  20  * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
  21  * http://www.gnu.org/copyleft/gpl.html
  22  *
  23  * @file
  24  * @ingroup Maintenance
  25  * @see http://www.sitemaps.org/
  26  * @see http://www.sitemaps.org/schemas/sitemap/0.9/sitemap.xsd
  27  */
  28
  29 require_once( dirname( __FILE__ ) . '/Maintenance.php' );
  30
  31 class GenerateSitemap extends Maintenance {
  32         const GS_MAIN = -2;
  33         const GS_TALK = -1;
  34
  35         /**
  36          * The maximum amount of urls in a sitemap file
  37          *
  38          * @link http://www.sitemaps.org/schemas/sitemap/0.9/sitemap.xsd
  39          *
  40          * @var int
  41          */
  42         var $url_limit;
  43
  44         /**
  45          * The maximum size of a sitemap file
  46          *
  47          * @link http://www.sitemaps.org/faq.php#faq_sitemap_size
  48          *
  49          * @var int
  50          */
  51         var $size_limit;
  52
  53         /**
  54          * The path to prepend to the filename
  55          *
  56          * @var string
  57          */
  58         var $fspath;
  59
  60         /**
  61          * The URL path to prepend to filenames in the index; should resolve to the same directory as $fspath
  62          *
  63          * @var string
  64          */
  65         var $urlpath;
  66
  67         /**
  68          * Whether or not to use compression
  69          *
  70          * @var bool
  71          */
  72         var $compress;
  73
  74         /**
  75          * The number of entries to save in each sitemap file
  76          *
  77          * @var array
  78          */
  79         var $limit = array();
  80
  81         /**
  82          * Key => value entries of namespaces and their priorities
  83          *
  84          * @var array
  85          */
  86         var $priorities = array();
  87
  88         /**
  89          * A one-dimensional array of namespaces in the wiki
  90          *
  91          * @var array
  92          */
  93         var $namespaces = array();
  94
  95         /**
  96          * When this sitemap batch was generated
  97          *
  98          * @var string
  99          */
 100         var $timestamp;
 101
 102         /**
 103          * A database slave object
 104          *
 105          * @var object
 106          */
 107         var $dbr;
 108
 109         /**
 110          * A resource pointing to the sitemap index file
 111          *
 112          * @var resource
 113          */
 114         var $findex;
 115
 116
 117         /**
 118          * A resource pointing to a sitemap file
 119          *
 120          * @var resource
 121          */
 122         var $file;
 123
 124         /**
 125          * Constructor
 126          */
 127         public function __construct() {
 128                 parent::__construct();
 129                 $this->mDescription = "Creates a sitemap for the site";
 130                 $this->addOption( 'fspath', 'The file system path to save to, e.g. /tmp/sitemap; defaults to current directory', false, true );
 131                 $this->addOption( 'urlpath', 'The URL path corresponding to --fspath, prepended to filenames in the index; defaults to an empty string', false, true );
 132                 $this->addOption( 'compress', 'Compress the sitemap files, can take value yes|no, default yes', false, true );
 133         }
 134
 135         /**
 136          * Execute
 137          */
 138         public function execute() {
 139                 $this->setNamespacePriorities();
 140                 $this->url_limit = 50000;
 141                 $this->size_limit = pow( 2, 20 ) * 10;
 142                 $this->fspath = self::init_path( $this->getOption( 'fspath', getcwd() ) );
 143                 $this->urlpath = $this->getOption( 'urlpath', "" );
 144                 if ( $this->urlpath !== "" && substr( $this->urlpath, -1 ) !== '/' ) {
 145                         $this->urlpath .= '/';
 146                 }
 147                 $this->compress = $this->getOption( 'compress', 'yes' ) !== 'no';
 148                 $this->dbr = wfGetDB( DB_SLAVE );
 149                 $this->generateNamespaces();
 150                 $this->timestamp = wfTimestamp( TS_ISO_8601, wfTimestampNow() );
 151                 $this->findex = fopen( "{$this->fspath}sitemap-index-" . wfWikiID() . ".xml", 'wb' );
 152                 $this->main();
 153         }
 154
 155         private function setNamespacePriorities() {
 156                 // Custom main namespaces
 157                 $this->priorities[self::GS_MAIN] = '0.5';
 158                 // Custom talk namesspaces
 159                 $this->priorities[self::GS_TALK] = '0.1';
 160                 // MediaWiki standard namespaces
 161                 $this->priorities[NS_MAIN] = '1.0';
 162                 $this->priorities[NS_TALK] = '0.1';
 163                 $this->priorities[NS_USER] = '0.5';
 164                 $this->priorities[NS_USER_TALK] = '0.1';
 165                 $this->priorities[NS_PROJECT] = '0.5';
 166                 $this->priorities[NS_PROJECT_TALK] = '0.1';
 167                 $this->priorities[NS_FILE] = '0.5';
 168                 $this->priorities[NS_FILE_TALK] = '0.1';
 169                 $this->priorities[NS_MEDIAWIKI] = '0.0';
 170                 $this->priorities[NS_MEDIAWIKI_TALK] = '0.1';
 171                 $this->priorities[NS_TEMPLATE] = '0.0';
 172                 $this->priorities[NS_TEMPLATE_TALK] = '0.1';
 173                 $this->priorities[NS_HELP] = '0.5';
 174                 $this->priorities[NS_HELP_TALK] = '0.1';
 175                 $this->priorities[NS_CATEGORY] = '0.5';
 176                 $this->priorities[NS_CATEGORY_TALK] = '0.1';
 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                 foreach ( $res as $row )
 216                         $this->namespaces[] = $row->page_namespace;
 217         }
 218
 219         /**
 220          * Get the priority of a given namespace
 221          *
 222          * @param $namespace Integer: the namespace to get the priority for
 223          * @return String
 224          */
 225         function priority( $namespace ) {
 226                 return isset( $this->priorities[$namespace] ) ? $this->priorities[$namespace] : $this->guessPriority( $namespace );
 227         }
 228
 229         /**
 230          * If the namespace isn't listed on the priority list return the
 231          * default priority for the namespace, varies depending on whether it's
 232          * a talkpage or not.
 233          *
 234          * @param $namespace Integer: the namespace to get the priority for
 235          * @return String
 236          */
 237         function guessPriority( $namespace ) {
 238                 return MWNamespace::isMain( $namespace ) ? $this->priorities[self::GS_MAIN] : $this->priorities[self::GS_TALK];
 239         }
 240
 241         /**
 242          * Return a database resolution of all the pages in a given namespace
 243          *
 244          * @param $namespace Integer: limit the query to this namespace
 245          * @return Resource
 246          */
 247         function getPageRes( $namespace ) {
 248                 return $this->dbr->select( 'page',
 249                         array(
 250                                 'page_namespace',
 251                                 'page_title',
 252                                 'page_touched',
 253                         ),
 254                         array( 'page_namespace' => $namespace ),
 255                         __METHOD__
 256                 );
 257         }
 258
 259         /**
 260          * Main loop
 261          */
 262         public function main() {
 263                 global $wgContLang;
 264
 265                 fwrite( $this->findex, $this->openIndex() );
 266
 267                 foreach ( $this->namespaces as $namespace ) {
 268                         $res = $this->getPageRes( $namespace );
 269                         $this->file = false;
 270                         $this->generateLimit( $namespace );
 271                         $length = $this->limit[0];
 272                         $i = $smcount = 0;
 273
 274                         $fns = $wgContLang->getFormattedNsText( $namespace );
 275                         $this->output( "$namespace ($fns)\n" );
 276                         foreach ( $res as $row ) {
 277                                 if ( $i++ === 0 || $i === $this->url_limit + 1 || $length + $this->limit[1] + $this->limit[2] > $this->size_limit ) {
 278                                         if ( $this->file !== false ) {
 279                                                 $this->write( $this->file, $this->closeFile() );
 280                                                 $this->close( $this->file );
 281                                         }
 282                                         $filename = $this->sitemapFilename( $namespace, $smcount++ );
 283                                         $this->file = $this->open( $this->fspath . $filename, 'wb' );
 284                                         $this->write( $this->file, $this->openFile() );
 285                                         fwrite( $this->findex, $this->indexEntry( $filename ) );
 286                                         $this->output( "\t$this->fspath$filename\n" );
 287                                         $length = $this->limit[0];
 288                                         $i = 1;
 289                                 }
 290                                 $title = Title::makeTitle( $row->page_namespace, $row->page_title );
 291                                 $date = wfTimestamp( TS_ISO_8601, $row->page_touched );
 292                                 $entry = $this->fileEntry( $title->getFullURL(), $date, $this->priority( $namespace ) );
 293                                 $length += strlen( $entry );
 294                                 $this->write( $this->file, $entry );
 295                                 // generate pages for language variants
 296                                 if ( $wgContLang->hasVariants() ) {
 297                                         $variants = $wgContLang->getVariants();
 298                                         foreach ( $variants as $vCode ) {
 299                                                 if ( $vCode == $wgContLang->getCode() ) continue; // we don't want default variant
 300                                                 $entry = $this->fileEntry( $title->getFullURL( '', $vCode ), $date, $this->priority( $namespace ) );
 301                                                 $length += strlen( $entry );
 302                                                 $this->write( $this->file, $entry );
 303                                         }
 304                                 }
 305                         }
 306                         if ( $this->file ) {
 307                                 $this->write( $this->file, $this->closeFile() );
 308                                 $this->close( $this->file );
 309                         }
 310                 }
 311                 fwrite( $this->findex, $this->closeIndex() );
 312                 fclose( $this->findex );
 313         }
 314
 315         /**
 316          * gzopen() / fopen() wrapper
 317          *
 318          * @return Resource
 319          */
 320         function open( $file, $flags ) {
 321                 return $this->compress ? gzopen( $file, $flags ) : fopen( $file, $flags );
 322         }
 323
 324         /**
 325          * gzwrite() / fwrite() wrapper
 326          */
 327         function write( &$handle, $str ) {
 328                 if ( $this->compress )
 329                         gzwrite( $handle, $str );
 330                 else
 331                         fwrite( $handle, $str );
 332         }
 333
 334         /**
 335          * gzclose() / fclose() wrapper
 336          */
 337         function close( &$handle ) {
 338                 if ( $this->compress )
 339                         gzclose( $handle );
 340                 else
 341                         fclose( $handle );
 342         }
 343
 344         /**
 345          * Get a sitemap filename
 346          *
 347          * @param $namespace Integer: the namespace
 348          * @param $count Integer: the count
 349          * @return String
 350          */
 351         function sitemapFilename( $namespace, $count ) {
 352                 $ext = $this->compress ? '.gz' : '';
 353                 return "sitemap-" . wfWikiID() . "-NS_$namespace-$count.xml$ext";
 354         }
 355
 356         /**
 357          * Return the XML required to open an XML file
 358          *
 359          * @return string
 360          */
 361         function xmlHead() {
 362                 return '<?xml version="1.0" encoding="UTF-8"?>' . "\n";
 363         }
 364
 365         /**
 366          * Return the XML schema being used
 367          *
 368          * @return String
 369          */
 370         function xmlSchema() {
 371                 return 'http://www.sitemaps.org/schemas/sitemap/0.9';
 372         }
 373
 374         /**
 375          * Return the XML required to open a sitemap index file
 376          *
 377          * @return String
 378          */
 379         function openIndex() {
 380                 return $this->xmlHead() . '<sitemapindex xmlns="' . $this->xmlSchema() . '">' . "\n";
 381         }
 382
 383         /**
 384          * Return the XML for a single sitemap indexfile entry
 385          *
 386          * @param $filename String: the filename of the sitemap file
 387          * @return String
 388          */
 389         function indexEntry( $filename ) {
 390                 return
 391                         "\t<sitemap>\n" .
 392                         "\t\t<loc>{$this->urlpath}$filename</loc>\n" .
 393                         "\t\t<lastmod>{$this->timestamp}</lastmod>\n" .
 394                         "\t</sitemap>\n";
 395         }
 396
 397         /**
 398          * Return the XML required to close a sitemap index file
 399          *
 400          * @return String
 401          */
 402         function closeIndex() {
 403                 return "</sitemapindex>\n";
 404         }
 405
 406         /**
 407          * Return the XML required to open a sitemap file
 408          *
 409          * @return String
 410          */
 411         function openFile() {
 412                 return $this->xmlHead() . '<urlset xmlns="' . $this->xmlSchema() . '">' . "\n";
 413         }
 414
 415         /**
 416          * Return the XML for a single sitemap entry
 417          *
 418          * @param $url String: an RFC 2396 compliant URL
 419          * @param $date String: a ISO 8601 date
 420          * @param $priority String: a priority indicator, 0.0 - 1.0 inclusive with a 0.1 stepsize
 421          * @return String
 422          */
 423         function fileEntry( $url, $date, $priority ) {
 424                 return
 425                         "\t<url>\n" .
 426                         "\t\t<loc>$url</loc>\n" .
 427                         "\t\t<lastmod>$date</lastmod>\n" .
 428                         "\t\t<priority>$priority</priority>\n" .
 429                         "\t</url>\n";
 430         }
 431
 432         /**
 433          * Return the XML required to close sitemap file
 434          *
 435          * @return String
 436          */
 437         function closeFile() {
 438                 return "</urlset>\n";
 439         }
 440
 441         /**
 442          * Populate $this->limit
 443          */
 444         function generateLimit( $namespace ) {
 445                 // bug 17961: make a title with the longest possible URL in this namespace
 446                 $title = Title::makeTitle( $namespace, str_repeat( "\xf0\xa8\xae\x81", 63 ) . "\xe5\x96\x83" );
 447
 448                 $this->limit = array(
 449                         strlen( $this->openFile() ),
 450                         strlen( $this->fileEntry( $title->getFullUrl(), wfTimestamp( TS_ISO_8601, wfTimestamp() ), $this->priority( $namespace ) ) ),
 451                         strlen( $this->closeFile() )
 452                 );
 453         }
 454 }
 455
 456 $maintClass = "GenerateSitemap";
 457 require_once( RUN_MAINTENANCE_IF_MAIN );