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