includes/api/ApiQuery.php

   1 <?php
   2 /**
   3  * Copyright © 2006 Yuri Astrakhan "<Firstname><Lastname>@gmail.com"
   4  *
   5  * This program is free software; you can redistribute it and/or modify
   6  * it under the terms of the GNU General Public License as published by
   7  * the Free Software Foundation; either version 2 of the License, or
   8  * (at your option) any later version.
   9  *
  10  * This program is distributed in the hope that it will be useful,
  11  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  12  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
  13  * GNU General Public License for more details.
  14  *
  15  * You should have received a copy of the GNU General Public License along
  16  * with this program; if not, write to the Free Software Foundation, Inc.,
  17  * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
  18  * http://www.gnu.org/copyleft/gpl.html
  19  *
  20  * @file
  21  */
  22
  23 use Wikimedia\Rdbms\IDatabase;
  24
  25 /**
  26  * This is the main query class. It behaves similar to ApiMain: based on the
  27  * parameters given, it will create a list of titles to work on (an ApiPageSet
  28  * object), instantiate and execute various property/list/meta modules, and
  29  * assemble all resulting data into a single ApiResult object.
  30  *
  31  * In generator mode, a generator will be executed first to populate a second
  32  * ApiPageSet object, and that object will be used for all subsequent modules.
  33  *
  34  * @ingroup API
  35  */
  36 class ApiQuery extends ApiBase {
  37
  38         /**
  39          * List of Api Query prop modules
  40          * @var array
  41          */
  42         private static $QueryPropModules = [
  43                 'categories' => ApiQueryCategories::class,
  44                 'categoryinfo' => ApiQueryCategoryInfo::class,
  45                 'contributors' => ApiQueryContributors::class,
  46                 'deletedrevisions' => ApiQueryDeletedRevisions::class,
  47                 'duplicatefiles' => ApiQueryDuplicateFiles::class,
  48                 'extlinks' => ApiQueryExternalLinks::class,
  49                 'fileusage' => ApiQueryBacklinksprop::class,
  50                 'images' => ApiQueryImages::class,
  51                 'imageinfo' => ApiQueryImageInfo::class,
  52                 'info' => ApiQueryInfo::class,
  53                 'links' => ApiQueryLinks::class,
  54                 'linkshere' => ApiQueryBacklinksprop::class,
  55                 'iwlinks' => ApiQueryIWLinks::class,
  56                 'langlinks' => ApiQueryLangLinks::class,
  57                 'pageprops' => ApiQueryPageProps::class,
  58                 'redirects' => ApiQueryBacklinksprop::class,
  59                 'revisions' => ApiQueryRevisions::class,
  60                 'stashimageinfo' => ApiQueryStashImageInfo::class,
  61                 'templates' => ApiQueryLinks::class,
  62                 'transcludedin' => ApiQueryBacklinksprop::class,
  63         ];
  64
  65         /**
  66          * List of Api Query list modules
  67          * @var array
  68          */
  69         private static $QueryListModules = [
  70                 'allcategories' => ApiQueryAllCategories::class,
  71                 'alldeletedrevisions' => ApiQueryAllDeletedRevisions::class,
  72                 'allfileusages' => ApiQueryAllLinks::class,
  73                 'allimages' => ApiQueryAllImages::class,
  74                 'alllinks' => ApiQueryAllLinks::class,
  75                 'allpages' => ApiQueryAllPages::class,
  76                 'allredirects' => ApiQueryAllLinks::class,
  77                 'allrevisions' => ApiQueryAllRevisions::class,
  78                 'mystashedfiles' => ApiQueryMyStashedFiles::class,
  79                 'alltransclusions' => ApiQueryAllLinks::class,
  80                 'allusers' => ApiQueryAllUsers::class,
  81                 'backlinks' => ApiQueryBacklinks::class,
  82                 'blocks' => ApiQueryBlocks::class,
  83                 'categorymembers' => ApiQueryCategoryMembers::class,
  84                 'deletedrevs' => ApiQueryDeletedrevs::class,
  85                 'embeddedin' => ApiQueryBacklinks::class,
  86                 'exturlusage' => ApiQueryExtLinksUsage::class,
  87                 'filearchive' => ApiQueryFilearchive::class,
  88                 'imageusage' => ApiQueryBacklinks::class,
  89                 'iwbacklinks' => ApiQueryIWBacklinks::class,
  90                 'langbacklinks' => ApiQueryLangBacklinks::class,
  91                 'logevents' => ApiQueryLogEvents::class,
  92                 'pageswithprop' => ApiQueryPagesWithProp::class,
  93                 'pagepropnames' => ApiQueryPagePropNames::class,
  94                 'prefixsearch' => ApiQueryPrefixSearch::class,
  95                 'protectedtitles' => ApiQueryProtectedTitles::class,
  96                 'querypage' => ApiQueryQueryPage::class,
  97                 'random' => ApiQueryRandom::class,
  98                 'recentchanges' => ApiQueryRecentChanges::class,
  99                 'search' => ApiQuerySearch::class,
 100                 'tags' => ApiQueryTags::class,
 101                 'usercontribs' => ApiQueryContributions::class,
 102                 'users' => ApiQueryUsers::class,
 103                 'watchlist' => ApiQueryWatchlist::class,
 104                 'watchlistraw' => ApiQueryWatchlistRaw::class,
 105         ];
 106
 107         /**
 108          * List of Api Query meta modules
 109          * @var array
 110          */
 111         private static $QueryMetaModules = [
 112                 'allmessages' => ApiQueryAllMessages::class,
 113                 'authmanagerinfo' => ApiQueryAuthManagerInfo::class,
 114                 'siteinfo' => ApiQuerySiteinfo::class,
 115                 'userinfo' => ApiQueryUserInfo::class,
 116                 'filerepoinfo' => ApiQueryFileRepoInfo::class,
 117                 'tokens' => ApiQueryTokens::class,
 118         ];
 119
 120         /**
 121          * @var ApiPageSet
 122          */
 123         private $mPageSet;
 124
 125         private $mParams;
 126         private $mNamedDB = [];
 127         private $mModuleMgr;
 128
 129         /**
 130          * @param ApiMain $main
 131          * @param string $action
 132          */
 133         public function __construct( ApiMain $main, $action ) {
 134                 parent::__construct( $main, $action );
 135
 136                 $this->mModuleMgr = new ApiModuleManager( $this );
 137
 138                 // Allow custom modules to be added in LocalSettings.php
 139                 $config = $this->getConfig();
 140                 $this->mModuleMgr->addModules( self::$QueryPropModules, 'prop' );
 141                 $this->mModuleMgr->addModules( $config->get( 'APIPropModules' ), 'prop' );
 142                 $this->mModuleMgr->addModules( self::$QueryListModules, 'list' );
 143                 $this->mModuleMgr->addModules( $config->get( 'APIListModules' ), 'list' );
 144                 $this->mModuleMgr->addModules( self::$QueryMetaModules, 'meta' );
 145                 $this->mModuleMgr->addModules( $config->get( 'APIMetaModules' ), 'meta' );
 146
 147                 Hooks::run( 'ApiQuery::moduleManager', [ $this->mModuleMgr ] );
 148
 149                 // Create PageSet that will process titles/pageids/revids/generator
 150                 $this->mPageSet = new ApiPageSet( $this );
 151         }
 152
 153         /**
 154          * Overrides to return this instance's module manager.
 155          * @return ApiModuleManager
 156          */
 157         public function getModuleManager() {
 158                 return $this->mModuleMgr;
 159         }
 160
 161         /**
 162          * Get the query database connection with the given name.
 163          * If no such connection has been requested before, it will be created.
 164          * Subsequent calls with the same $name will return the same connection
 165          * as the first, regardless of the values of $db and $groups
 166          * @param string $name Name to assign to the database connection
 167          * @param int $db One of the DB_* constants
 168          * @param string|string[] $groups Query groups
 169          * @return IDatabase
 170          */
 171         public function getNamedDB( $name, $db, $groups ) {
 172                 if ( !array_key_exists( $name, $this->mNamedDB ) ) {
 173                         $this->mNamedDB[$name] = wfGetDB( $db, $groups );
 174                 }
 175
 176                 return $this->mNamedDB[$name];
 177         }
 178
 179         /**
 180          * Gets the set of pages the user has requested (or generated)
 181          * @return ApiPageSet
 182          */
 183         public function getPageSet() {
 184                 return $this->mPageSet;
 185         }
 186
 187         /**
 188          * @return ApiFormatRaw|null
 189          */
 190         public function getCustomPrinter() {
 191                 // If &exportnowrap is set, use the raw formatter
 192                 if ( $this->getParameter( 'export' ) &&
 193                         $this->getParameter( 'exportnowrap' )
 194                 ) {
 195                         return new ApiFormatRaw( $this->getMain(),
 196                                 $this->getMain()->createPrinterByName( 'xml' ) );
 197                 } else {
 198                         return null;
 199                 }
 200         }
 201
 202         /**
 203          * Query execution happens in the following steps:
 204          * #1 Create a PageSet object with any pages requested by the user
 205          * #2 If using a generator, execute it to get a new ApiPageSet object
 206          * #3 Instantiate all requested modules.
 207          *    This way the PageSet object will know what shared data is required,
 208          *    and minimize DB calls.
 209          * #4 Output all normalization and redirect resolution information
 210          * #5 Execute all requested modules
 211          */
 212         public function execute() {
 213                 $this->mParams = $this->extractRequestParams();
 214
 215                 // Instantiate requested modules
 216                 $allModules = [];
 217                 $this->instantiateModules( $allModules, 'prop' );
 218                 $propModules = array_keys( $allModules );
 219                 $this->instantiateModules( $allModules, 'list' );
 220                 $this->instantiateModules( $allModules, 'meta' );
 221
 222                 // Filter modules based on continue parameter
 223                 $continuationManager = new ApiContinuationManager( $this, $allModules, $propModules );
 224                 $this->setContinuationManager( $continuationManager );
 225                 $modules = $continuationManager->getRunModules();
 226
 227                 if ( !$continuationManager->isGeneratorDone() ) {
 228                         // Query modules may optimize data requests through the $this->getPageSet()
 229                         // object by adding extra fields from the page table.
 230                         foreach ( $modules as $module ) {
 231                                 $module->requestExtraData( $this->mPageSet );
 232                         }
 233                         // Populate page/revision information
 234                         $this->mPageSet->execute();
 235                         // Record page information (title, namespace, if exists, etc)
 236                         $this->outputGeneralPageInfo();
 237                 } else {
 238                         $this->mPageSet->executeDryRun();
 239                 }
 240
 241                 $cacheMode = $this->mPageSet->getCacheMode();
 242
 243                 // Execute all unfinished modules
 244                 /** @var ApiQueryBase $module */
 245                 foreach ( $modules as $module ) {
 246                         $params = $module->extractRequestParams();
 247                         $cacheMode = $this->mergeCacheMode(
 248                                 $cacheMode, $module->getCacheMode( $params ) );
 249                         $module->execute();
 250                         Hooks::run( 'APIQueryAfterExecute', [ &$module ] );
 251                 }
 252
 253                 // Set the cache mode
 254                 $this->getMain()->setCacheMode( $cacheMode );
 255
 256                 // Write the continuation data into the result
 257                 $this->setContinuationManager( null );
 258                 if ( $this->mParams['rawcontinue'] ) {
 259                         $data = $continuationManager->getRawNonContinuation();
 260                         if ( $data ) {
 261                                 $this->getResult()->addValue( null, 'query-noncontinue', $data,
 262                                         ApiResult::ADD_ON_TOP | ApiResult::NO_SIZE_CHECK );
 263                         }
 264                         $data = $continuationManager->getRawContinuation();
 265                         if ( $data ) {
 266                                 $this->getResult()->addValue( null, 'query-continue', $data,
 267                                         ApiResult::ADD_ON_TOP | ApiResult::NO_SIZE_CHECK );
 268                         }
 269                 } else {
 270                         $continuationManager->setContinuationIntoResult( $this->getResult() );
 271                 }
 272         }
 273
 274         /**
 275          * Update a cache mode string, applying the cache mode of a new module to it.
 276          * The cache mode may increase in the level of privacy, but public modules
 277          * added to private data do not decrease the level of privacy.
 278          *
 279          * @param string $cacheMode
 280          * @param string $modCacheMode
 281          * @return string
 282          */
 283         protected function mergeCacheMode( $cacheMode, $modCacheMode ) {
 284                 if ( $modCacheMode === 'anon-public-user-private' ) {
 285                         if ( $cacheMode !== 'private' ) {
 286                                 $cacheMode = 'anon-public-user-private';
 287                         }
 288                 } elseif ( $modCacheMode === 'public' ) {
 289                         // do nothing, if it's public already it will stay public
 290                 } else { // private
 291                         $cacheMode = 'private';
 292                 }
 293
 294                 return $cacheMode;
 295         }
 296
 297         /**
 298          * Create instances of all modules requested by the client
 299          * @param array $modules To append instantiated modules to
 300          * @param string $param Parameter name to read modules from
 301          */
 302         private function instantiateModules( &$modules, $param ) {
 303                 $wasPosted = $this->getRequest()->wasPosted();
 304                 if ( isset( $this->mParams[$param] ) ) {
 305                         foreach ( $this->mParams[$param] as $moduleName ) {
 306                                 $instance = $this->mModuleMgr->getModule( $moduleName, $param );
 307                                 if ( $instance === null ) {
 308                                         ApiBase::dieDebug( __METHOD__, 'Error instantiating module' );
 309                                 }
 310                                 if ( !$wasPosted && $instance->mustBePosted() ) {
 311                                         $this->dieWithErrorOrDebug( [ 'apierror-mustbeposted', $moduleName ] );
 312                                 }
 313                                 // Ignore duplicates. TODO 2.0: die()?
 314                                 if ( !array_key_exists( $moduleName, $modules ) ) {
 315                                         $modules[$moduleName] = $instance;
 316                                 }
 317                         }
 318                 }
 319         }
 320
 321         /**
 322          * Appends an element for each page in the current pageSet with the
 323          * most general information (id, title), plus any title normalizations
 324          * and missing or invalid title/pageids/revids.
 325          */
 326         private function outputGeneralPageInfo() {
 327                 $pageSet = $this->getPageSet();
 328                 $result = $this->getResult();
 329
 330                 // We can't really handle max-result-size failure here, but we need to
 331                 // check anyway in case someone set the limit stupidly low.
 332                 $fit = true;
 333
 334                 $values = $pageSet->getNormalizedTitlesAsResult( $result );
 335                 if ( $values ) {
 336                         $fit = $fit && $result->addValue( 'query', 'normalized', $values );
 337                 }
 338                 $values = $pageSet->getConvertedTitlesAsResult( $result );
 339                 if ( $values ) {
 340                         $fit = $fit && $result->addValue( 'query', 'converted', $values );
 341                 }
 342                 $values = $pageSet->getInterwikiTitlesAsResult( $result, $this->mParams['iwurl'] );
 343                 if ( $values ) {
 344                         $fit = $fit && $result->addValue( 'query', 'interwiki', $values );
 345                 }
 346                 $values = $pageSet->getRedirectTitlesAsResult( $result );
 347                 if ( $values ) {
 348                         $fit = $fit && $result->addValue( 'query', 'redirects', $values );
 349                 }
 350                 $values = $pageSet->getMissingRevisionIDsAsResult( $result );
 351                 if ( $values ) {
 352                         $fit = $fit && $result->addValue( 'query', 'badrevids', $values );
 353                 }
 354
 355                 // Page elements
 356                 $pages = [];
 357
 358                 // Report any missing titles
 359                 foreach ( $pageSet->getMissingTitles() as $fakeId => $title ) {
 360                         $vals = [];
 361                         ApiQueryBase::addTitleInfo( $vals, $title );
 362                         $vals['missing'] = true;
 363                         if ( $title->isKnown() ) {
 364                                 $vals['known'] = true;
 365                         }
 366                         $pages[$fakeId] = $vals;
 367                 }
 368                 // Report any invalid titles
 369                 foreach ( $pageSet->getInvalidTitlesAndReasons() as $fakeId => $data ) {
 370                         $pages[$fakeId] = $data + [ 'invalid' => true ];
 371                 }
 372                 // Report any missing page ids
 373                 foreach ( $pageSet->getMissingPageIDs() as $pageid ) {
 374                         $pages[$pageid] = [
 375                                 'pageid' => $pageid,
 376                                 'missing' => true,
 377                         ];
 378                 }
 379                 // Report special pages
 380                 /** @var Title $title */
 381                 foreach ( $pageSet->getSpecialTitles() as $fakeId => $title ) {
 382                         $vals = [];
 383                         ApiQueryBase::addTitleInfo( $vals, $title );
 384                         $vals['special'] = true;
 385                         if ( !$title->isKnown() ) {
 386                                 $vals['missing'] = true;
 387                         }
 388                         $pages[$fakeId] = $vals;
 389                 }
 390
 391                 // Output general page information for found titles
 392                 foreach ( $pageSet->getGoodTitles() as $pageid => $title ) {
 393                         $vals = [];
 394                         $vals['pageid'] = $pageid;
 395                         ApiQueryBase::addTitleInfo( $vals, $title );
 396                         $pages[$pageid] = $vals;
 397                 }
 398
 399                 if ( count( $pages ) ) {
 400                         $pageSet->populateGeneratorData( $pages );
 401                         ApiResult::setArrayType( $pages, 'BCarray' );
 402
 403                         if ( $this->mParams['indexpageids'] ) {
 404                                 $pageIDs = array_keys( ApiResult::stripMetadataNonRecursive( $pages ) );
 405                                 // json treats all map keys as strings - converting to match
 406                                 $pageIDs = array_map( 'strval', $pageIDs );
 407                                 ApiResult::setIndexedTagName( $pageIDs, 'id' );
 408                                 $fit = $fit && $result->addValue( 'query', 'pageids', $pageIDs );
 409                         }
 410
 411                         ApiResult::setIndexedTagName( $pages, 'page' );
 412                         $fit = $fit && $result->addValue( 'query', 'pages', $pages );
 413                 }
 414
 415                 if ( !$fit ) {
 416                         $this->dieWithError( 'apierror-badconfig-resulttoosmall', 'badconfig' );
 417                 }
 418
 419                 if ( $this->mParams['export'] ) {
 420                         $this->doExport( $pageSet, $result );
 421                 }
 422         }
 423
 424         /**
 425          * @param ApiPageSet $pageSet Pages to be exported
 426          * @param ApiResult $result Result to output to
 427          */
 428         private function doExport( $pageSet, $result ) {
 429                 $exportTitles = [];
 430                 $titles = $pageSet->getGoodTitles();
 431                 if ( count( $titles ) ) {
 432                         $user = $this->getUser();
 433                         /** @var Title $title */
 434                         foreach ( $titles as $title ) {
 435                                 if ( $title->userCan( 'read', $user ) ) {
 436                                         $exportTitles[] = $title;
 437                                 }
 438                         }
 439                 }
 440
 441                 $exporter = new WikiExporter( $this->getDB() );
 442                 $sink = new DumpStringOutput;
 443                 $exporter->setOutputSink( $sink );
 444                 $exporter->openStream();
 445                 foreach ( $exportTitles as $title ) {
 446                         $exporter->pageByTitle( $title );
 447                 }
 448                 $exporter->closeStream();
 449
 450                 // Don't check the size of exported stuff
 451                 // It's not continuable, so it would cause more
 452                 // problems than it'd solve
 453                 if ( $this->mParams['exportnowrap'] ) {
 454                         $result->reset();
 455                         // Raw formatter will handle this
 456                         $result->addValue( null, 'text', $sink, ApiResult::NO_SIZE_CHECK );
 457                         $result->addValue( null, 'mime', 'text/xml', ApiResult::NO_SIZE_CHECK );
 458                         $result->addValue( null, 'filename', 'export.xml', ApiResult::NO_SIZE_CHECK );
 459                 } else {
 460                         $result->addValue( 'query', 'export', $sink, ApiResult::NO_SIZE_CHECK );
 461                         $result->addValue( 'query', ApiResult::META_BC_SUBELEMENTS, [ 'export' ] );
 462                 }
 463         }
 464
 465         public function getAllowedParams( $flags = 0 ) {
 466                 $result = [
 467                         'prop' => [
 468                                 ApiBase::PARAM_ISMULTI => true,
 469                                 ApiBase::PARAM_TYPE => 'submodule',
 470                         ],
 471                         'list' => [
 472                                 ApiBase::PARAM_ISMULTI => true,
 473                                 ApiBase::PARAM_TYPE => 'submodule',
 474                         ],
 475                         'meta' => [
 476                                 ApiBase::PARAM_ISMULTI => true,
 477                                 ApiBase::PARAM_TYPE => 'submodule',
 478                         ],
 479                         'indexpageids' => false,
 480                         'export' => false,
 481                         'exportnowrap' => false,
 482                         'iwurl' => false,
 483                         'continue' => [
 484                                 ApiBase::PARAM_HELP_MSG => 'api-help-param-continue',
 485                         ],
 486                         'rawcontinue' => false,
 487                 ];
 488                 if ( $flags ) {
 489                         $result += $this->getPageSet()->getFinalParams( $flags );
 490                 }
 491
 492                 return $result;
 493         }
 494
 495         public function isReadMode() {
 496                 // We need to make an exception for certain meta modules that should be
 497                 // accessible even without the 'read' right. Restrict the exception as
 498                 // much as possible: no other modules allowed, and no pageset
 499                 // parameters either. We do allow the 'rawcontinue' and 'indexpageids'
 500                 // parameters since frameworks might add these unconditionally and they
 501                 // can't expose anything here.
 502                 $this->mParams = $this->extractRequestParams();
 503                 $params = array_filter(
 504                         array_diff_key(
 505                                 $this->mParams + $this->getPageSet()->extractRequestParams(),
 506                                 [ 'rawcontinue' => 1, 'indexpageids' => 1 ]
 507                         )
 508                 );
 509                 if ( array_keys( $params ) !== [ 'meta' ] ) {
 510                         return true;
 511                 }
 512
 513                 // Ask each module if it requires read mode. Any true => this returns
 514                 // true.
 515                 $modules = [];
 516                 $this->instantiateModules( $modules, 'meta' );
 517                 foreach ( $modules as $module ) {
 518                         if ( $module->isReadMode() ) {
 519                                 return true;
 520                         }
 521                 }
 522
 523                 return false;
 524         }
 525
 526         protected function getExamplesMessages() {
 527                 return [
 528                         'action=query&prop=revisions&meta=siteinfo&' .
 529                                 'titles=Main%20Page&rvprop=user|comment&continue='
 530                                 => 'apihelp-query-example-revisions',
 531                         'action=query&generator=allpages&gapprefix=API/&prop=revisions&continue='
 532                                 => 'apihelp-query-example-allpages',
 533                 ];
 534         }
 535
 536         public function getHelpUrls() {
 537                 return [
 538                         'https://www.mediawiki.org/wiki/Special:MyLanguage/API:Query',
 539                         'https://www.mediawiki.org/wiki/Special:MyLanguage/API:Meta',
 540                         'https://www.mediawiki.org/wiki/Special:MyLanguage/API:Properties',
 541                         'https://www.mediawiki.org/wiki/Special:MyLanguage/API:Lists',
 542                 ];
 543         }
 544 }