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          * Identifier to use in filenames, default $wgDBname
 126          *
 127          * @var string
 128          */
 129         private $identifier;
 130
 131         /**
 132          * Constructor
 133          */
 134         public function __construct() {
 135                 parent::__construct();
 136                 global $wgDBname;
 137                 $this->mDescription = "Creates a sitemap for the site";
 138                 $this->addOption( 'fspath', 'The file system path to save to, e.g. /tmp/sitemap; defaults to current directory', false, true );
 139                 $this->addOption( 'urlpath', 'The URL path corresponding to --fspath, prepended to filenames in the index; defaults to an empty string', false, true );
 140                 $this->addOption( 'compress', 'Compress the sitemap files, can take value yes|no, default yes', false, true );
 141                 $this->addOption( 'identifier', 'What site identifier to use for the wiki, defaults to $wgDBname', false, true );
 142         }
 143
 144         /**
 145          * Execute
 146          */
 147         public function execute() {
 148                 $this->setNamespacePriorities();
 149                 $this->url_limit = 50000;
 150                 $this->size_limit = pow( 2, 20 ) * 10;
 151                 $this->fspath = self::init_path( $this->getOption( 'fspath', getcwd() ) );
 152                 $this->urlpath = $this->getOption( 'urlpath', "" );
 153                 if ( $this->urlpath !== "" && substr( $this->urlpath, -1 ) !== '/' ) {
 154                         $this->urlpath .= '/';
 155                 }
 156                 $this->identifier = $this->getOption( 'identifier', wfWikiID() );
 157                 $this->compress = $this->getOption( 'compress', 'yes' ) !== 'no';
 158                 $this->dbr = wfGetDB( DB_SLAVE );
 159                 $this->generateNamespaces();
 160                 $this->timestamp = wfTimestamp( TS_ISO_8601, wfTimestampNow() );
 161                 $this->findex = fopen( "{$this->fspath}sitemap-index-{$this->identifier}.xml", 'wb' );
 162                 $this->main();
 163         }
 164
 165         private function setNamespacePriorities() {
 166                 // Custom main namespaces
 167                 $this->priorities[self::GS_MAIN] = '0.5';
 168                 // Custom talk namesspaces
 169                 $this->priorities[self::GS_TALK] = '0.1';
 170                 // MediaWiki standard namespaces
 171                 $this->priorities[NS_MAIN] = '1.0';
 172                 $this->priorities[NS_TALK] = '0.1';
 173                 $this->priorities[NS_USER] = '0.5';
 174                 $this->priorities[NS_USER_TALK] = '0.1';
 175                 $this->priorities[NS_PROJECT] = '0.5';
 176                 $this->priorities[NS_PROJECT_TALK] = '0.1';
 177                 $this->priorities[NS_FILE] = '0.5';
 178                 $this->priorities[NS_FILE_TALK] = '0.1';
 179                 $this->priorities[NS_MEDIAWIKI] = '0.0';
 180                 $this->priorities[NS_MEDIAWIKI_TALK] = '0.1';
 181                 $this->priorities[NS_TEMPLATE] = '0.0';
 182                 $this->priorities[NS_TEMPLATE_TALK] = '0.1';
 183                 $this->priorities[NS_HELP] = '0.5';
 184                 $this->priorities[NS_HELP_TALK] = '0.1';
 185                 $this->priorities[NS_CATEGORY] = '0.5';
 186                 $this->priorities[NS_CATEGORY_TALK] = '0.1';
 187         }
 188
 189         /**
 190          * Create directory if it does not exist and return pathname with a trailing slash
 191          */
 192         private static function init_path( $fspath ) {
 193                 if ( !isset( $fspath ) ) {
 194                         return null;
 195                 }
 196                 # Create directory if needed
 197                 if ( $fspath && !is_dir( $fspath ) ) {
 198                         wfMkdirParents( $fspath ) or die( "Can not create directory $fspath.\n" );
 199                 }
 200
 201                 return realpath( $fspath ) . DIRECTORY_SEPARATOR ;
 202         }
 203
 204         /**
 205          * Generate a one-dimensional array of existing namespaces
 206          */
 207         function generateNamespaces() {
 208                 // Only generate for specific namespaces if $wgSitemapNamespaces is an array.
 209                 global $wgSitemapNamespaces;
 210                 if ( is_array( $wgSitemapNamespaces ) ) {
 211                         $this->namespaces = $wgSitemapNamespaces;
 212                         return;
 213                 }
 214
 215                 $res = $this->dbr->select( 'page',
 216                         array( 'page_namespace' ),
 217                         array(),
 218                         __METHOD__,
 219                         array(
 220                                 'GROUP BY' => 'page_namespace',
 221                                 'ORDER BY' => 'page_namespace',
 222                         )
 223                 );
 224
 225                 foreach ( $res as $row )
 226                         $this->namespaces[] = $row->page_namespace;
 227         }
 228
 229         /**
 230          * Get the priority of a given namespace
 231          *
 232          * @param $namespace Integer: the namespace to get the priority for
 233          * @return String
 234          */
 235         function priority( $namespace ) {
 236                 return isset( $this->priorities[$namespace] ) ? $this->priorities[$namespace] : $this->guessPriority( $namespace );
 237         }
 238
 239         /**
 240          * If the namespace isn't listed on the priority list return the
 241          * default priority for the namespace, varies depending on whether it's
 242          * a talkpage or not.
 243          *
 244          * @param $namespace Integer: the namespace to get the priority for
 245          * @return String
 246          */
 247         function guessPriority( $namespace ) {
 248                 return MWNamespace::isMain( $namespace ) ? $this->priorities[self::GS_MAIN] : $this->priorities[self::GS_TALK];
 249         }
 250
 251         /**
 252          * Return a database resolution of all the pages in a given namespace
 253          *
 254          * @param $namespace Integer: limit the query to this namespace
 255          * @return Resource
 256          */
 257         function getPageRes( $namespace ) {
 258                 return $this->dbr->select( 'page',
 259                         array(
 260                                 'page_namespace',
 261                                 'page_title',
 262                                 'page_touched',
 263                         ),
 264                         array( 'page_namespace' => $namespace ),
 265                         __METHOD__
 266                 );
 267         }
 268
 269         /**
 270          * Main loop
 271          */
 272         public function main() {
 273                 global $wgContLang;
 274
 275                 fwrite( $this->findex, $this->openIndex() );
 276
 277                 foreach ( $this->namespaces as $namespace ) {
 278                         $res = $this->getPageRes( $namespace );
 279                         $this->file = false;
 280                         $this->generateLimit( $namespace );
 281                         $length = $this->limit[0];
 282                         $i = $smcount = 0;
 283
 284                         $fns = $wgContLang->getFormattedNsText( $namespace );
 285                         $this->output( "$namespace ($fns)\n" );
 286                         foreach ( $res as $row ) {
 287                                 if ( $i++ === 0 || $i === $this->url_limit + 1 || $length + $this->limit[1] + $this->limit[2] > $this->size_limit ) {
 288                                         if ( $this->file !== false ) {
 289                                                 $this->write( $this->file, $this->closeFile() );
 290                                                 $this->close( $this->file );
 291                                         }
 292                                         $filename = $this->sitemapFilename( $namespace, $smcount++ );
 293                                         $this->file = $this->open( $this->fspath . $filename, 'wb' );
 294                                         $this->write( $this->file, $this->openFile() );
 295                                         fwrite( $this->findex, $this->indexEntry( $filename ) );
 296                                         $this->output( "\t$this->fspath$filename\n" );
 297                                         $length = $this->limit[0];
 298                                         $i = 1;
 299                                 }
 300                                 $title = Title::makeTitle( $row->page_namespace, $row->page_title );
 301                                 $date = wfTimestamp( TS_ISO_8601, $row->page_touched );
 302                                 $entry = $this->fileEntry( $title->getFullURL(), $date, $this->priority( $namespace ) );
 303                                 $length += strlen( $entry );
 304                                 $this->write( $this->file, $entry );
 305                                 // generate pages for language variants
 306                                 if ( $wgContLang->hasVariants() ) {
 307                                         $variants = $wgContLang->getVariants();
 308                                         foreach ( $variants as $vCode ) {
 309                                                 if ( $vCode == $wgContLang->getCode() ) continue; // we don't want default variant
 310                                                 $entry = $this->fileEntry( $title->getFullURL( '', $vCode ), $date, $this->priority( $namespace ) );
 311                                                 $length += strlen( $entry );
 312                                                 $this->write( $this->file, $entry );
 313                                         }
 314                                 }
 315                         }
 316                         if ( $this->file ) {
 317                                 $this->write( $this->file, $this->closeFile() );
 318                                 $this->close( $this->file );
 319                         }
 320                 }
 321                 fwrite( $this->findex, $this->closeIndex() );
 322                 fclose( $this->findex );
 323         }
 324
 325         /**
 326          * gzopen() / fopen() wrapper
 327          *
 328          * @return Resource
 329          */
 330         function open( $file, $flags ) {
 331                 return $this->compress ? gzopen( $file, $flags ) : fopen( $file, $flags );
 332         }
 333
 334         /**
 335          * gzwrite() / fwrite() wrapper
 336          */
 337         function write( &$handle, $str ) {
 338                 if ( $this->compress )
 339                         gzwrite( $handle, $str );
 340                 else
 341                         fwrite( $handle, $str );
 342         }
 343
 344         /**
 345          * gzclose() / fclose() wrapper
 346          */
 347         function close( &$handle ) {
 348                 if ( $this->compress )
 349                         gzclose( $handle );
 350                 else
 351                         fclose( $handle );
 352         }
 353
 354         /**
 355          * Get a sitemap filename
 356          *
 357          * @param $namespace Integer: the namespace
 358          * @param $count Integer: the count
 359          * @return String
 360          */
 361         function sitemapFilename( $namespace, $count ) {
 362                 $ext = $this->compress ? '.gz' : '';
 363                 return "sitemap-{$this->identifier}-NS_$namespace-$count.xml$ext";
 364         }
 365
 366         /**
 367          * Return the XML required to open an XML file
 368          *
 369          * @return string
 370          */
 371         function xmlHead() {
 372                 return '<?xml version="1.0" encoding="UTF-8"?>' . "\n";
 373         }
 374
 375         /**
 376          * Return the XML schema being used
 377          *
 378          * @return String
 379          */
 380         function xmlSchema() {
 381                 return 'http://www.sitemaps.org/schemas/sitemap/0.9';
 382         }
 383
 384         /**
 385          * Return the XML required to open a sitemap index file
 386          *
 387          * @return String
 388          */
 389         function openIndex() {
 390                 return $this->xmlHead() . '<sitemapindex xmlns="' . $this->xmlSchema() . '">' . "\n";
 391         }
 392
 393         /**
 394          * Return the XML for a single sitemap indexfile entry
 395          *
 396          * @param $filename String: the filename of the sitemap file
 397          * @return String
 398          */
 399         function indexEntry( $filename ) {
 400                 return
 401                         "\t<sitemap>\n" .
 402                         "\t\t<loc>{$this->urlpath}$filename</loc>\n" .
 403                         "\t\t<lastmod>{$this->timestamp}</lastmod>\n" .
 404                         "\t</sitemap>\n";
 405         }
 406
 407         /**
 408          * Return the XML required to close a sitemap index file
 409          *
 410          * @return String
 411          */
 412         function closeIndex() {
 413                 return "</sitemapindex>\n";
 414         }
 415
 416         /**
 417          * Return the XML required to open a sitemap file
 418          *
 419          * @return String
 420          */
 421         function openFile() {
 422                 return $this->xmlHead() . '<urlset xmlns="' . $this->xmlSchema() . '">' . "\n";
 423         }
 424
 425         /**
 426          * Return the XML for a single sitemap entry
 427          *
 428          * @param $url String: an RFC 2396 compliant URL
 429          * @param $date String: a ISO 8601 date
 430          * @param $priority String: a priority indicator, 0.0 - 1.0 inclusive with a 0.1 stepsize
 431          * @return String
 432          */
 433         function fileEntry( $url, $date, $priority ) {
 434                 return
 435                         "\t<url>\n" .
 436                         "\t\t<loc>$url</loc>\n" .
 437                         "\t\t<lastmod>$date</lastmod>\n" .
 438                         "\t\t<priority>$priority</priority>\n" .
 439                         "\t</url>\n";
 440         }
 441
 442         /**
 443          * Return the XML required to close sitemap file
 444          *
 445          * @return String
 446          */
 447         function closeFile() {
 448                 return "</urlset>\n";
 449         }
 450
 451         /**
 452          * Populate $this->limit
 453          */
 454         function generateLimit( $namespace ) {
 455                 // bug 17961: make a title with the longest possible URL in this namespace
 456                 $title = Title::makeTitle( $namespace, str_repeat( "\xf0\xa8\xae\x81", 63 ) . "\xe5\x96\x83" );
 457
 458                 $this->limit = array(
 459                         strlen( $this->openFile() ),
 460                         strlen( $this->fileEntry( $title->getFullUrl(), wfTimestamp( TS_ISO_8601, wfTimestamp() ), $this->priority( $namespace ) ) ),
 461                         strlen( $this->closeFile() )
 462                 );
 463         }
 464 }
 465
 466 $maintClass = "GenerateSitemap";
 467 require_once( RUN_MAINTENANCE_IF_MAIN );