includes/api/ApiQuery.php

   1 <?php
   2 /**
   3  *
   4  *
   5  * Created on Sep 7, 2006
   6  *
   7  * Copyright © 2006 Yuri Astrakhan "<Firstname><Lastname>@gmail.com"
   8  *
   9  * This program is free software; you can redistribute it and/or modify
  10  * it under the terms of the GNU General Public License as published by
  11  * the Free Software Foundation; either version 2 of the License, or
  12  * (at your option) any later version.
  13  *
  14  * This program is distributed in the hope that it will be useful,
  15  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  16  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
  17  * GNU General Public License for more details.
  18  *
  19  * You should have received a copy of the GNU General Public License along
  20  * with this program; if not, write to the Free Software Foundation, Inc.,
  21  * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
  22  * http://www.gnu.org/copyleft/gpl.html
  23  *
  24  * @file
  25  */
  26
  27 /**
  28  * This is the main query class. It behaves similar to ApiMain: based on the
  29  * parameters given, it will create a list of titles to work on (an ApiPageSet
  30  * object), instantiate and execute various property/list/meta modules, and
  31  * assemble all resulting data into a single ApiResult object.
  32  *
  33  * In generator mode, a generator will be executed first to populate a second
  34  * ApiPageSet object, and that object will be used for all subsequent modules.
  35  *
  36  * @ingroup API
  37  */
  38 class ApiQuery extends ApiBase {
  39
  40         /**
  41          * List of Api Query prop modules
  42          * @var array
  43          */
  44         private static $QueryPropModules = array(
  45                 'categories' => 'ApiQueryCategories',
  46                 'categoryinfo' => 'ApiQueryCategoryInfo',
  47                 'duplicatefiles' => 'ApiQueryDuplicateFiles',
  48                 'extlinks' => 'ApiQueryExternalLinks',
  49                 'images' => 'ApiQueryImages',
  50                 'imageinfo' => 'ApiQueryImageInfo',
  51                 'info' => 'ApiQueryInfo',
  52                 'links' => 'ApiQueryLinks',
  53                 'iwlinks' => 'ApiQueryIWLinks',
  54                 'langlinks' => 'ApiQueryLangLinks',
  55                 'pageprops' => 'ApiQueryPageProps',
  56                 'revisions' => 'ApiQueryRevisions',
  57                 'stashimageinfo' => 'ApiQueryStashImageInfo',
  58                 'templates' => 'ApiQueryLinks',
  59         );
  60
  61         /**
  62          * List of Api Query list modules
  63          * @var array
  64          */
  65         private static $QueryListModules = array(
  66                 'allcategories' => 'ApiQueryAllCategories',
  67                 'allimages' => 'ApiQueryAllImages',
  68                 'alllinks' => 'ApiQueryAllLinks',
  69                 'allpages' => 'ApiQueryAllPages',
  70                 'alltransclusions' => 'ApiQueryAllLinks',
  71                 'allusers' => 'ApiQueryAllUsers',
  72                 'backlinks' => 'ApiQueryBacklinks',
  73                 'blocks' => 'ApiQueryBlocks',
  74                 'categorymembers' => 'ApiQueryCategoryMembers',
  75                 'deletedrevs' => 'ApiQueryDeletedrevs',
  76                 'embeddedin' => 'ApiQueryBacklinks',
  77                 'exturlusage' => 'ApiQueryExtLinksUsage',
  78                 'filearchive' => 'ApiQueryFilearchive',
  79                 'imageusage' => 'ApiQueryBacklinks',
  80                 'iwbacklinks' => 'ApiQueryIWBacklinks',
  81                 'langbacklinks' => 'ApiQueryLangBacklinks',
  82                 'logevents' => 'ApiQueryLogEvents',
  83                 'protectedtitles' => 'ApiQueryProtectedTitles',
  84                 'querypage' => 'ApiQueryQueryPage',
  85                 'random' => 'ApiQueryRandom',
  86                 'recentchanges' => 'ApiQueryRecentChanges',
  87                 'search' => 'ApiQuerySearch',
  88                 'tags' => 'ApiQueryTags',
  89                 'usercontribs' => 'ApiQueryContributions',
  90                 'users' => 'ApiQueryUsers',
  91                 'watchlist' => 'ApiQueryWatchlist',
  92                 'watchlistraw' => 'ApiQueryWatchlistRaw',
  93         );
  94
  95         /**
  96          * List of Api Query meta modules
  97          * @var array
  98          */
  99         private static $QueryMetaModules = array(
 100                 'allmessages' => 'ApiQueryAllMessages',
 101                 'siteinfo' => 'ApiQuerySiteinfo',
 102                 'userinfo' => 'ApiQueryUserInfo',
 103         );
 104
 105         /**
 106          * @var ApiPageSet
 107          */
 108         private $mPageSet;
 109
 110         private $params;
 111         private $iwUrl;
 112         private $mNamedDB = array();
 113         private $mModuleMgr;
 114
 115         /**
 116          * @param $main ApiMain
 117          * @param $action string
 118          */
 119         public function __construct( $main, $action ) {
 120                 parent::__construct( $main, $action );
 121
 122                 $this->mModuleMgr = new ApiModuleManager( $this );
 123
 124                 // Allow custom modules to be added in LocalSettings.php
 125                 global $wgAPIPropModules, $wgAPIListModules, $wgAPIMetaModules;
 126                 $this->mModuleMgr->addModules( self::$QueryPropModules, 'prop' );
 127                 $this->mModuleMgr->addModules( $wgAPIPropModules, 'prop' );
 128                 $this->mModuleMgr->addModules( self::$QueryListModules, 'list' );
 129                 $this->mModuleMgr->addModules( $wgAPIListModules, 'list' );
 130                 $this->mModuleMgr->addModules( self::$QueryMetaModules, 'meta' );
 131                 $this->mModuleMgr->addModules( $wgAPIMetaModules, 'meta' );
 132
 133                 // Create PageSet that will process titles/pageids/revids/generator
 134                 $this->mPageSet = new ApiPageSet( $this );
 135         }
 136
 137         /**
 138          * Overrides to return this instance's module manager.
 139          * @return ApiModuleManager
 140          */
 141         public function getModuleManager() {
 142                 return $this->mModuleMgr;
 143         }
 144
 145         /**
 146          * Get the query database connection with the given name.
 147          * If no such connection has been requested before, it will be created.
 148          * Subsequent calls with the same $name will return the same connection
 149          * as the first, regardless of the values of $db and $groups
 150          * @param $name string Name to assign to the database connection
 151          * @param $db int One of the DB_* constants
 152          * @param $groups array Query groups
 153          * @return DatabaseBase
 154          */
 155         public function getNamedDB( $name, $db, $groups ) {
 156                 if ( !array_key_exists( $name, $this->mNamedDB ) ) {
 157                         $this->profileDBIn();
 158                         $this->mNamedDB[$name] = wfGetDB( $db, $groups );
 159                         $this->profileDBOut();
 160                 }
 161                 return $this->mNamedDB[$name];
 162         }
 163
 164         /**
 165          * Gets the set of pages the user has requested (or generated)
 166          * @return ApiPageSet
 167          */
 168         public function getPageSet() {
 169                 return $this->mPageSet;
 170         }
 171
 172         /**
 173          * Get the array mapping module names to class names
 174          * @deprecated since 1.21, use getModuleManager()'s methods instead
 175          * @return array array(modulename => classname)
 176          */
 177         public function getModules() {
 178                 wfDeprecated( __METHOD__, '1.21' );
 179                 return $this->getModuleManager()->getNamesWithClasses();
 180         }
 181
 182         /**
 183          * Get the generators array mapping module names to class names
 184          * @deprecated since 1.21, list of generators is maintained by ApiPageSet
 185          * @return array array(modulename => classname)
 186          */
 187         public function getGenerators() {
 188                 wfDeprecated( __METHOD__, '1.21' );
 189                 $gens = array();
 190                 foreach ( $this->mModuleMgr->getNamesWithClasses() as $name => $class ) {
 191                         if ( is_subclass_of( $class, 'ApiQueryGeneratorBase' ) ) {
 192                                 $gens[$name] = $class;
 193                         }
 194                 }
 195                 return $gens;
 196         }
 197
 198         /**
 199          * Get whether the specified module is a prop, list or a meta query module
 200          * @deprecated since 1.21, use getModuleManager()->getModuleGroup()
 201          * @param $moduleName string Name of the module to find type for
 202          * @return mixed string or null
 203          */
 204         function getModuleType( $moduleName ) {
 205                 return $this->getModuleManager()->getModuleGroup( $moduleName );
 206         }
 207
 208         /**
 209          * @return ApiFormatRaw|null
 210          */
 211         public function getCustomPrinter() {
 212                 // If &exportnowrap is set, use the raw formatter
 213                 if ( $this->getParameter( 'export' ) &&
 214                                 $this->getParameter( 'exportnowrap' ) )
 215                 {
 216                         return new ApiFormatRaw( $this->getMain(),
 217                                 $this->getMain()->createPrinterByName( 'xml' ) );
 218                 } else {
 219                         return null;
 220                 }
 221         }
 222
 223         /**
 224          * Query execution happens in the following steps:
 225          * #1 Create a PageSet object with any pages requested by the user
 226          * #2 If using a generator, execute it to get a new ApiPageSet object
 227          * #3 Instantiate all requested modules.
 228          *    This way the PageSet object will know what shared data is required,
 229          *    and minimize DB calls.
 230          * #4 Output all normalization and redirect resolution information
 231          * #5 Execute all requested modules
 232          */
 233         public function execute() {
 234                 $this->params = $this->extractRequestParams();
 235                 $this->iwUrl = $this->params['iwurl'];
 236
 237                 // Instantiate requested modules
 238                 $modules = array();
 239                 $this->instantiateModules( $modules, 'prop' );
 240                 $this->instantiateModules( $modules, 'list' );
 241                 $this->instantiateModules( $modules, 'meta' );
 242
 243                 // Query modules may optimize data requests through the $this->getPageSet()
 244                 // object by adding extra fields from the page table.
 245                 // This function will gather all the extra request fields from the modules.
 246                 foreach ( $modules as $module ) {
 247                         if ( !$this->getRequest()->wasPosted() && $module->mustBePosted() ) {
 248                                 $this->dieUsageMsgOrDebug( array( 'mustbeposted', $module->getModuleName() ) );
 249                         }
 250
 251                         $module->requestExtraData( $this->mPageSet );
 252                 }
 253
 254                 // Populate page/revision information
 255                 $this->mPageSet->execute();
 256                 $cacheMode = $this->mPageSet->getCacheMode();
 257
 258                 // Record page information (title, namespace, if exists, etc)
 259                 $this->outputGeneralPageInfo();
 260
 261                 // Execute all requested modules.
 262                 /**
 263                  * @var $module ApiQueryBase
 264                  */
 265                 foreach ( $modules as $module ) {
 266                         $params = $module->extractRequestParams();
 267                         $cacheMode = $this->mergeCacheMode(
 268                                 $cacheMode, $module->getCacheMode( $params ) );
 269                         $module->profileIn();
 270                         $module->execute();
 271                         wfRunHooks( 'APIQueryAfterExecute', array( &$module ) );
 272                         $module->profileOut();
 273                 }
 274
 275                 // Set the cache mode
 276                 $this->getMain()->setCacheMode( $cacheMode );
 277         }
 278
 279         /**
 280          * Update a cache mode string, applying the cache mode of a new module to it.
 281          * The cache mode may increase in the level of privacy, but public modules
 282          * added to private data do not decrease the level of privacy.
 283          *
 284          * @param $cacheMode string
 285          * @param $modCacheMode string
 286          * @return string
 287          */
 288         protected function mergeCacheMode( $cacheMode, $modCacheMode ) {
 289                 if ( $modCacheMode === 'anon-public-user-private' ) {
 290                         if ( $cacheMode !== 'private' ) {
 291                                 $cacheMode = 'anon-public-user-private';
 292                         }
 293                 } elseif ( $modCacheMode === 'public' ) {
 294                         // do nothing, if it's public already it will stay public
 295                 } else { // private
 296                         $cacheMode = 'private';
 297                 }
 298                 return $cacheMode;
 299         }
 300
 301         /**
 302          * Create instances of all modules requested by the client
 303          * @param $modules Array to append instantiated modules to
 304          * @param $param string Parameter name to read modules from
 305          */
 306         private function instantiateModules( &$modules, $param ) {
 307                 if ( isset( $this->params[$param] ) ) {
 308                         foreach ( $this->params[$param] as $moduleName ) {
 309                                 $modules[] = $this->mModuleMgr->getModule( $moduleName );
 310                         }
 311                 }
 312         }
 313
 314         /**
 315          * Appends an element for each page in the current pageSet with the
 316          * most general information (id, title), plus any title normalizations
 317          * and missing or invalid title/pageids/revids.
 318          */
 319         private function outputGeneralPageInfo() {
 320                 $pageSet = $this->getPageSet();
 321                 $result = $this->getResult();
 322
 323                 // We don't check for a full result set here because we can't be adding
 324                 // more than 380K. The maximum revision size is in the megabyte range,
 325                 // and the maximum result size must be even higher than that.
 326
 327                 $values = $pageSet->getNormalizedTitlesAsResult( $result );
 328                 if ( $values ) {
 329                         $result->addValue( 'query', 'normalized', $values );
 330                 }
 331                 $values = $pageSet->getConvertedTitlesAsResult( $result );
 332                 if ( $values ) {
 333                         $result->addValue( 'query', 'converted', $values );
 334                 }
 335                 $values = $pageSet->getInterwikiTitlesAsResult( $result, $this->iwUrl );
 336                 if ( $values ) {
 337                         $result->addValue( 'query', 'interwiki', $values );
 338                 }
 339                 $values = $pageSet->getRedirectTitlesAsResult( $result );
 340                 if ( $values ) {
 341                         $result->addValue( 'query', 'redirects', $values );
 342                 }
 343                 $values = $pageSet->getMissingRevisionIDsAsResult( $result );
 344                 if ( $values ) {
 345                         $result->addValue( 'query', 'badrevids', $values );
 346                 }
 347
 348                 // Page elements
 349                 $pages = array();
 350
 351                 // Report any missing titles
 352                 foreach ( $pageSet->getMissingTitles() as $fakeId => $title ) {
 353                         $vals = array();
 354                         ApiQueryBase::addTitleInfo( $vals, $title );
 355                         $vals['missing'] = '';
 356                         $pages[$fakeId] = $vals;
 357                 }
 358                 // Report any invalid titles
 359                 foreach ( $pageSet->getInvalidTitles() as $fakeId => $title ) {
 360                         $pages[$fakeId] = array( 'title' => $title, 'invalid' => '' );
 361                 }
 362                 // Report any missing page ids
 363                 foreach ( $pageSet->getMissingPageIDs() as $pageid ) {
 364                         $pages[$pageid] = array(
 365                                 'pageid' => $pageid,
 366                                 'missing' => ''
 367                         );
 368                 }
 369                 // Report special pages
 370                 foreach ( $pageSet->getSpecialTitles() as $fakeId => $title ) {
 371                         $vals = array();
 372                         ApiQueryBase::addTitleInfo( $vals, $title );
 373                         $vals['special'] = '';
 374                         if ( $title->isSpecialPage() &&
 375                                         !SpecialPageFactory::exists( $title->getDbKey() ) ) {
 376                                 $vals['missing'] = '';
 377                         } elseif ( $title->getNamespace() == NS_MEDIA &&
 378                                         !wfFindFile( $title ) ) {
 379                                 $vals['missing'] = '';
 380                         }
 381                         $pages[$fakeId] = $vals;
 382                 }
 383
 384                 // Output general page information for found titles
 385                 foreach ( $pageSet->getGoodTitles() as $pageid => $title ) {
 386                         $vals = array();
 387                         $vals['pageid'] = $pageid;
 388                         ApiQueryBase::addTitleInfo( $vals, $title );
 389                         $pages[$pageid] = $vals;
 390                 }
 391
 392                 if ( count( $pages ) ) {
 393                         if ( $this->params['indexpageids'] ) {
 394                                 $pageIDs = array_keys( $pages );
 395                                 // json treats all map keys as strings - converting to match
 396                                 $pageIDs = array_map( 'strval', $pageIDs );
 397                                 $result->setIndexedTagName( $pageIDs, 'id' );
 398                                 $result->addValue( 'query', 'pageids', $pageIDs );
 399                         }
 400
 401                         $result->setIndexedTagName( $pages, 'page' );
 402                         $result->addValue( 'query', 'pages', $pages );
 403                 }
 404                 if ( $this->params['export'] ) {
 405                         $this->doExport( $pageSet, $result );
 406                 }
 407         }
 408
 409         /**
 410          * @param $pageSet ApiPageSet Pages to be exported
 411          * @param $result ApiResult Result to output to
 412          */
 413         private function doExport( $pageSet, $result ) {
 414                 $exportTitles = array();
 415                 $titles = $pageSet->getGoodTitles();
 416                 if ( count( $titles ) ) {
 417                         $user = $this->getUser();
 418                         foreach ( $titles as $title ) {
 419                                 if ( $title->userCan( 'read', $user ) ) {
 420                                         $exportTitles[] = $title;
 421                                 }
 422                         }
 423                 }
 424
 425                 $exporter = new WikiExporter( $this->getDB() );
 426                 // WikiExporter writes to stdout, so catch its
 427                 // output with an ob
 428                 ob_start();
 429                 $exporter->openStream();
 430                 foreach ( $exportTitles as $title ) {
 431                         $exporter->pageByTitle( $title );
 432                 }
 433                 $exporter->closeStream();
 434                 $exportxml = ob_get_contents();
 435                 ob_end_clean();
 436
 437                 // Don't check the size of exported stuff
 438                 // It's not continuable, so it would cause more
 439                 // problems than it'd solve
 440                 $result->disableSizeCheck();
 441                 if ( $this->params['exportnowrap'] ) {
 442                         $result->reset();
 443                         // Raw formatter will handle this
 444                         $result->addValue( null, 'text', $exportxml );
 445                         $result->addValue( null, 'mime', 'text/xml' );
 446                 } else {
 447                         $r = array();
 448                         ApiResult::setContent( $r, $exportxml );
 449                         $result->addValue( 'query', 'export', $r );
 450                 }
 451                 $result->enableSizeCheck();
 452         }
 453
 454         public function getAllowedParams( $flags = 0 ) {
 455                 $result = array(
 456                         'prop' => array(
 457                                 ApiBase::PARAM_ISMULTI => true,
 458                                 ApiBase::PARAM_TYPE => $this->mModuleMgr->getNames( 'prop' )
 459                         ),
 460                         'list' => array(
 461                                 ApiBase::PARAM_ISMULTI => true,
 462                                 ApiBase::PARAM_TYPE => $this->mModuleMgr->getNames( 'list' )
 463                         ),
 464                         'meta' => array(
 465                                 ApiBase::PARAM_ISMULTI => true,
 466                                 ApiBase::PARAM_TYPE => $this->mModuleMgr->getNames( 'meta' )
 467                         ),
 468                         'indexpageids' => false,
 469                         'export' => false,
 470                         'exportnowrap' => false,
 471                         'iwurl' => false,
 472                 );
 473                 if ( $flags ) {
 474                         $result += $this->getPageSet()->getFinalParams( $flags );
 475                 }
 476                 return $result;
 477         }
 478
 479         /**
 480          * Override the parent to generate help messages for all available query modules.
 481          * @return string
 482          */
 483         public function makeHelpMsg() {
 484
 485                 // Use parent to make default message for the query module
 486                 $msg = parent::makeHelpMsg();
 487
 488                 $querySeparator = str_repeat( '--- ', 12 );
 489                 $moduleSeparator = str_repeat( '*** ', 14 );
 490                 $msg .= "\n$querySeparator Query: Prop  $querySeparator\n\n";
 491                 $msg .= $this->makeHelpMsgHelper( 'prop' );
 492                 $msg .= "\n$querySeparator Query: List  $querySeparator\n\n";
 493                 $msg .= $this->makeHelpMsgHelper( 'list' );
 494                 $msg .= "\n$querySeparator Query: Meta  $querySeparator\n\n";
 495                 $msg .= $this->makeHelpMsgHelper( 'meta' );
 496                 $msg .= "\n\n$moduleSeparator Modules: continuation  $moduleSeparator\n\n";
 497
 498                 return $msg;
 499         }
 500
 501         /**
 502          * For all modules of a given group, generate help messages and join them together
 503          * @param $group string Module group
 504          * @return string
 505          */
 506         private function makeHelpMsgHelper( $group ) {
 507                 $moduleDescriptions = array();
 508
 509                 $moduleNames = $this->mModuleMgr->getNames( $group );
 510                 sort( $moduleNames );
 511                 foreach ( $moduleNames as $name ) {
 512                         /**
 513                          * @var $module ApiQueryBase
 514                          */
 515                         $module = $this->mModuleMgr->getModule( $name );
 516
 517                         $msg = ApiMain::makeHelpMsgHeader( $module, $group );
 518                         $msg2 = $module->makeHelpMsg();
 519                         if ( $msg2 !== false ) {
 520                                 $msg .= $msg2;
 521                         }
 522                         if ( $module instanceof ApiQueryGeneratorBase ) {
 523                                 $msg .= "Generator:\n  This module may be used as a generator\n";
 524                         }
 525                         $moduleDescriptions[] = $msg;
 526                 }
 527
 528                 return implode( "\n", $moduleDescriptions );
 529         }
 530
 531         public function shouldCheckMaxlag() {
 532                 return true;
 533         }
 534
 535         public function getParamDescription() {
 536                 return $this->getPageSet()->getParamDescription() + array(
 537                         'prop' => 'Which properties to get for the titles/revisions/pageids. Module help is available below',
 538                         'list' => 'Which lists to get. Module help is available below',
 539                         'meta' => 'Which metadata to get about the site. Module help is available below',
 540                         'indexpageids' => 'Include an additional pageids section listing all returned page IDs',
 541                         'export' => 'Export the current revisions of all given or generated pages',
 542                         'exportnowrap' => 'Return the export XML without wrapping it in an XML result (same format as Special:Export). Can only be used with export',
 543                         'iwurl' => 'Whether to get the full URL if the title is an interwiki link',
 544                 );
 545         }
 546
 547         public function getDescription() {
 548                 return array(
 549                         'Query API module allows applications to get needed pieces of data from the MediaWiki databases,',
 550                         'and is loosely based on the old query.php interface.',
 551                         'All data modifications will first have to use query to acquire a token to prevent abuse from malicious sites'
 552                 );
 553         }
 554
 555         public function getPossibleErrors() {
 556                 return array_merge(
 557                         parent::getPossibleErrors(),
 558                         $this->getPageSet()->getPossibleErrors()
 559                 );
 560         }
 561
 562         public function getExamples() {
 563                 return array(
 564                         'api.php?action=query&prop=revisions&meta=siteinfo&titles=Main%20Page&rvprop=user|comment',
 565                         'api.php?action=query&generator=allpages&gapprefix=API/&prop=revisions',
 566                 );
 567         }
 568
 569         public function getHelpUrls() {
 570                 return array(
 571                         'https://www.mediawiki.org/wiki/API:Meta',
 572                         'https://www.mediawiki.org/wiki/API:Properties',
 573                         'https://www.mediawiki.org/wiki/API:Lists',
 574                 );
 575         }
 576 }