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