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