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 URL path to prepend to filenames in the index; should resolve to the same directory as $fspath
  61          *
  62          * @var string
  63          */
  64         var $urlpath;
  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; defaults to current directory', false, true );
 130                 $this->addOption( 'urlpath', 'The URL path corresponding to --fspath, prepended to filenames in the index; defaults to an empty string', 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->urlpath = $this->getOption( 'urlpath', "" );
 143                 if ( $this->urlpath !== "" && substr( $this->urlpath, -1 ) !== '/' ) {
 144                         $this->urlpath .= '/';
 145                 }
 146                 $this->compress = $this->getOption( 'compress', 'yes' ) !== 'no';
 147                 $this->dbr = wfGetDB( DB_SLAVE );
 148                 $this->generateNamespaces();
 149                 $this->timestamp = wfTimestamp( TS_ISO_8601, wfTimestampNow() );
 150                 $this->findex = fopen( "{$this->fspath}sitemap-index-" . wfWikiID() . ".xml", 'wb' );
 151                 $this->main();
 152         }
 153
 154         private function setNamespacePriorities() {
 155                 // Custom main namespaces
 156                 $this->priorities[GS_MAIN] = '0.5';
 157                 // Custom talk namesspaces
 158                 $this->priorities[GS_TALK] = '0.1';
 159                 // MediaWiki standard namespaces
 160                 $this->priorities[NS_MAIN] = '1.0';
 161                 $this->priorities[NS_TALK] = '0.1';
 162                 $this->priorities[NS_USER] = '0.5';
 163                 $this->priorities[NS_USER_TALK] = '0.1';
 164                 $this->priorities[NS_PROJECT] = '0.5';
 165                 $this->priorities[NS_PROJECT_TALK] = '0.1';
 166                 $this->priorities[NS_FILE] = '0.5';
 167                 $this->priorities[NS_FILE_TALK] = '0.1';
 168                 $this->priorities[NS_MEDIAWIKI] = '0.0';
 169                 $this->priorities[NS_MEDIAWIKI_TALK] = '0.1';
 170                 $this->priorities[NS_TEMPLATE] = '0.0';
 171                 $this->priorities[NS_TEMPLATE_TALK] = '0.1';
 172                 $this->priorities[NS_HELP] = '0.5';
 173                 $this->priorities[NS_HELP_TALK] = '0.1';
 174                 $this->priorities[NS_CATEGORY] = '0.5';
 175                 $this->priorities[NS_CATEGORY_TALK] = '0.1';
 176         }
 177
 178         /**
 179          * Create directory if it does not exist and return pathname with a trailing slash
 180          */
 181         private static function init_path( $fspath ) {
 182                 if ( !isset( $fspath ) ) {
 183                         return null;
 184                 }
 185                 # Create directory if needed
 186                 if ( $fspath && !is_dir( $fspath ) ) {
 187                         wfMkdirParents( $fspath ) or die( "Can not create directory $fspath.\n" );
 188                 }
 189
 190                 return realpath( $fspath ) . DIRECTORY_SEPARATOR ;
 191         }
 192
 193         /**
 194          * Generate a one-dimensional array of existing namespaces
 195          */
 196         function generateNamespaces() {
 197                 // Only generate for specific namespaces if $wgSitemapNamespaces is an array.
 198                 global $wgSitemapNamespaces;
 199                 if ( is_array( $wgSitemapNamespaces ) ) {
 200                         $this->namespaces = $wgSitemapNamespaces;
 201                         return;
 202                 }
 203
 204                 $res = $this->dbr->select( 'page',
 205                         array( 'page_namespace' ),
 206                         array(),
 207                         __METHOD__,
 208                         array(
 209                                 'GROUP BY' => 'page_namespace',
 210                                 'ORDER BY' => 'page_namespace',
 211                         )
 212                 );
 213
 214                 foreach ( $res as $row )
 215                         $this->namespaces[] = $row->page_namespace;
 216         }
 217
 218         /**
 219          * Get the priority of a given namespace
 220          *
 221          * @param $namespace Integer: the namespace to get the priority for
 222          * @return String
 223          */
 224         function priority( $namespace ) {
 225                 return isset( $this->priorities[$namespace] ) ? $this->priorities[$namespace] : $this->guessPriority( $namespace );
 226         }
 227
 228         /**
 229          * If the namespace isn't listed on the priority list return the
 230          * default priority for the namespace, varies depending on whether it's
 231          * a talkpage or not.
 232          *
 233          * @param $namespace Integer: the namespace to get the priority for
 234          * @return String
 235          */
 236         function guessPriority( $namespace ) {
 237                 return MWNamespace::isMain( $namespace ) ? $this->priorities[GS_MAIN] : $this->priorities[GS_TALK];
 238         }
 239
 240         /**
 241          * Return a database resolution of all the pages in a given namespace
 242          *
 243          * @param $namespace Integer: limit the query to this namespace
 244          * @return Resource
 245          */
 246         function getPageRes( $namespace ) {
 247                 return $this->dbr->select( 'page',
 248                         array(
 249                                 'page_namespace',
 250                                 'page_title',
 251                                 'page_touched',
 252                         ),
 253                         array( 'page_namespace' => $namespace ),
 254                         __METHOD__
 255                 );
 256         }
 257
 258         /**
 259          * Main loop
 260          */
 261         public function main() {
 262                 global $wgContLang;
 263
 264                 fwrite( $this->findex, $this->openIndex() );
 265
 266                 foreach ( $this->namespaces as $namespace ) {
 267                         $res = $this->getPageRes( $namespace );
 268                         $this->file = false;
 269                         $this->generateLimit( $namespace );
 270                         $length = $this->limit[0];
 271                         $i = $smcount = 0;
 272
 273                         $fns = $wgContLang->getFormattedNsText( $namespace );
 274                         $this->output( "$namespace ($fns)" );
 275                         foreach ( $res as $row ) {
 276                                 if ( $i++ === 0 || $i === $this->url_limit + 1 || $length + $this->limit[1] + $this->limit[2] > $this->size_limit ) {
 277                                         if ( $this->file !== false ) {
 278                                                 $this->write( $this->file, $this->closeFile() );
 279                                                 $this->close( $this->file );
 280                                         }
 281                                         $filename = $this->sitemapFilename( $namespace, $smcount++ );
 282                                         $this->file = $this->open( $this->fspath . $filename, 'wb' );
 283                                         $this->write( $this->file, $this->openFile() );
 284                                         fwrite( $this->findex, $this->indexEntry( $filename ) );
 285                                         $this->output( "\t$this->fspath$filename\n" );
 286                                         $length = $this->limit[0];
 287                                         $i = 1;
 288                                 }
 289                                 $title = Title::makeTitle( $row->page_namespace, $row->page_title );
 290                                 $date = wfTimestamp( TS_ISO_8601, $row->page_touched );
 291                                 $entry = $this->fileEntry( $title->getFullURL(), $date, $this->priority( $namespace ) );
 292                                 $length += strlen( $entry );
 293                                 $this->write( $this->file, $entry );
 294                                 // generate pages for language variants
 295                                 if ( $wgContLang->hasVariants() ) {
 296                                         $variants = $wgContLang->getVariants();
 297                                         foreach ( $variants as $vCode ) {
 298                                                 if ( $vCode == $wgContLang->getCode() ) continue; // we don't want default variant
 299                                                 $entry = $this->fileEntry( $title->getFullURL( '', $vCode ), $date, $this->priority( $namespace ) );
 300                                                 $length += strlen( $entry );
 301                                                 $this->write( $this->file, $entry );
 302                                         }
 303                                 }
 304                         }
 305                         if ( $this->file ) {
 306                                 $this->write( $this->file, $this->closeFile() );
 307                                 $this->close( $this->file );
 308                         }
 309                 }
 310                 fwrite( $this->findex, $this->closeIndex() );
 311                 fclose( $this->findex );
 312         }
 313
 314         /**
 315          * gzopen() / fopen() wrapper
 316          *
 317          * @return Resource
 318          */
 319         function open( $file, $flags ) {
 320                 return $this->compress ? gzopen( $file, $flags ) : fopen( $file, $flags );
 321         }
 322
 323         /**
 324          * gzwrite() / fwrite() wrapper
 325          */
 326         function write( &$handle, $str ) {
 327                 if ( $this->compress )
 328                         gzwrite( $handle, $str );
 329                 else
 330                         fwrite( $handle, $str );
 331         }
 332
 333         /**
 334          * gzclose() / fclose() wrapper
 335          */
 336         function close( &$handle ) {
 337                 if ( $this->compress )
 338                         gzclose( $handle );
 339                 else
 340                         fclose( $handle );
 341         }
 342
 343         /**
 344          * Get a sitemap filename
 345          *
 346          * @param $namespace Integer: the namespace
 347          * @param $count Integer: the count
 348          * @return String
 349          */
 350         function sitemapFilename( $namespace, $count ) {
 351                 $ext = $this->compress ? '.gz' : '';
 352                 return "sitemap-" . wfWikiID() . "-NS_$namespace-$count.xml$ext";
 353         }
 354
 355         /**
 356          * Return the XML required to open an XML file
 357          *
 358          * @return string
 359          */
 360         function xmlHead() {
 361                 return '<?xml version="1.0" encoding="UTF-8"?>' . "\n";
 362         }
 363
 364         /**
 365          * Return the XML schema being used
 366          *
 367          * @return String
 368          */
 369         function xmlSchema() {
 370                 return 'http://www.sitemaps.org/schemas/sitemap/0.9';
 371         }
 372
 373         /**
 374          * Return the XML required to open a sitemap index file
 375          *
 376          * @return String
 377          */
 378         function openIndex() {
 379                 return $this->xmlHead() . '<sitemapindex xmlns="' . $this->xmlSchema() . '">' . "\n";
 380         }
 381
 382         /**
 383          * Return the XML for a single sitemap indexfile entry
 384          *
 385          * @param $filename String: the filename of the sitemap file
 386          * @return String
 387          */
 388         function indexEntry( $filename ) {
 389                 return
 390                         "\t<sitemap>\n" .
 391                         "\t\t<loc>{$this->urlpath}$filename</loc>\n" .
 392                         "\t\t<lastmod>{$this->timestamp}</lastmod>\n" .
 393                         "\t</sitemap>\n";
 394         }
 395
 396         /**
 397          * Return the XML required to close a sitemap index file
 398          *
 399          * @return String
 400          */
 401         function closeIndex() {
 402                 return "</sitemapindex>\n";
 403         }
 404
 405         /**
 406          * Return the XML required to open a sitemap file
 407          *
 408          * @return String
 409          */
 410         function openFile() {
 411                 return $this->xmlHead() . '<urlset xmlns="' . $this->xmlSchema() . '">' . "\n";
 412         }
 413
 414         /**
 415          * Return the XML for a single sitemap entry
 416          *
 417          * @param $url String: an RFC 2396 compliant URL
 418          * @param $date String: a ISO 8601 date
 419          * @param $priority String: a priority indicator, 0.0 - 1.0 inclusive with a 0.1 stepsize
 420          * @return String
 421          */
 422         function fileEntry( $url, $date, $priority ) {
 423                 return
 424                         "\t<url>\n" .
 425                         "\t\t<loc>$url</loc>\n" .
 426                         "\t\t<lastmod>$date</lastmod>\n" .
 427                         "\t\t<priority>$priority</priority>\n" .
 428                         "\t</url>\n";
 429         }
 430
 431         /**
 432          * Return the XML required to close sitemap file
 433          *
 434          * @return String
 435          */
 436         function closeFile() {
 437                 return "</urlset>\n";
 438         }
 439
 440         /**
 441          * Populate $this->limit
 442          */
 443         function generateLimit( $namespace ) {
 444                 $title = Title::makeTitle( $namespace, str_repeat( "\xf0\xa8\xae\x81", 63 ) . "\xe5\x96\x83" );
 445
 446                 $this->limit = array(
 447                         strlen( $this->openFile() ),
 448                         strlen( $this->fileEntry( $title->getFullUrl(), wfTimestamp( TS_ISO_8601, wfTimestamp() ), $this->priority( $namespace ) ) ),
 449                         strlen( $this->closeFile() )
 450                 );
 451         }
 452 }
 453
 454 $maintClass = "GenerateSitemap";
 455 require_once( DO_MAINTENANCE );