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 = [
  45                 'categories' => 'ApiQueryCategories',
  46                 'categoryinfo' => 'ApiQueryCategoryInfo',
  47                 'contributors' => 'ApiQueryContributors',
  48                 'deletedrevisions' => 'ApiQueryDeletedRevisions',
  49                 'duplicatefiles' => 'ApiQueryDuplicateFiles',
  50                 'extlinks' => 'ApiQueryExternalLinks',
  51                 'fileusage' => 'ApiQueryBacklinksprop',
  52                 'images' => 'ApiQueryImages',
  53                 'imageinfo' => 'ApiQueryImageInfo',
  54                 'info' => 'ApiQueryInfo',
  55                 'links' => 'ApiQueryLinks',
  56                 'linkshere' => 'ApiQueryBacklinksprop',
  57                 'iwlinks' => 'ApiQueryIWLinks',
  58                 'langlinks' => 'ApiQueryLangLinks',
  59                 'pageprops' => 'ApiQueryPageProps',
  60                 'redirects' => 'ApiQueryBacklinksprop',
  61                 'revisions' => 'ApiQueryRevisions',
  62                 'stashimageinfo' => 'ApiQueryStashImageInfo',
  63                 'templates' => 'ApiQueryLinks',
  64                 'transcludedin' => 'ApiQueryBacklinksprop',
  65         ];
  66
  67         /**
  68          * List of Api Query list modules
  69          * @var array
  70          */
  71         private static $QueryListModules = [
  72                 'allcategories' => 'ApiQueryAllCategories',
  73                 'alldeletedrevisions' => 'ApiQueryAllDeletedRevisions',
  74                 'allfileusages' => 'ApiQueryAllLinks',
  75                 'allimages' => 'ApiQueryAllImages',
  76                 'alllinks' => 'ApiQueryAllLinks',
  77                 'allpages' => 'ApiQueryAllPages',
  78                 'allredirects' => 'ApiQueryAllLinks',
  79                 'allrevisions' => 'ApiQueryAllRevisions',
  80                 'mystashedfiles' => 'ApiQueryMyStashedFiles',
  81                 'alltransclusions' => 'ApiQueryAllLinks',
  82                 'allusers' => 'ApiQueryAllUsers',
  83                 'backlinks' => 'ApiQueryBacklinks',
  84                 'blocks' => 'ApiQueryBlocks',
  85                 'categorymembers' => 'ApiQueryCategoryMembers',
  86                 'deletedrevs' => 'ApiQueryDeletedrevs',
  87                 'embeddedin' => 'ApiQueryBacklinks',
  88                 'exturlusage' => 'ApiQueryExtLinksUsage',
  89                 'filearchive' => 'ApiQueryFilearchive',
  90                 'imageusage' => 'ApiQueryBacklinks',
  91                 'iwbacklinks' => 'ApiQueryIWBacklinks',
  92                 'langbacklinks' => 'ApiQueryLangBacklinks',
  93                 'logevents' => 'ApiQueryLogEvents',
  94                 'pageswithprop' => 'ApiQueryPagesWithProp',
  95                 'pagepropnames' => 'ApiQueryPagePropNames',
  96                 'prefixsearch' => 'ApiQueryPrefixSearch',
  97                 'protectedtitles' => 'ApiQueryProtectedTitles',
  98                 'querypage' => 'ApiQueryQueryPage',
  99                 'random' => 'ApiQueryRandom',
 100                 'recentchanges' => 'ApiQueryRecentChanges',
 101                 'search' => 'ApiQuerySearch',
 102                 'tags' => 'ApiQueryTags',
 103                 'usercontribs' => 'ApiQueryContributions',
 104                 'users' => 'ApiQueryUsers',
 105                 'watchlist' => 'ApiQueryWatchlist',
 106                 'watchlistraw' => 'ApiQueryWatchlistRaw',
 107         ];
 108
 109         /**
 110          * List of Api Query meta modules
 111          * @var array
 112          */
 113         private static $QueryMetaModules = [
 114                 'allmessages' => 'ApiQueryAllMessages',
 115                 'authmanagerinfo' => 'ApiQueryAuthManagerInfo',
 116                 'siteinfo' => 'ApiQuerySiteinfo',
 117                 'userinfo' => 'ApiQueryUserInfo',
 118                 'filerepoinfo' => 'ApiQueryFileRepoInfo',
 119                 'tokens' => 'ApiQueryTokens',
 120         ];
 121
 122         /**
 123          * @var ApiPageSet
 124          */
 125         private $mPageSet;
 126
 127         private $mParams;
 128         private $mNamedDB = [];
 129         private $mModuleMgr;
 130
 131         /**
 132          * @param ApiMain $main
 133          * @param string $action
 134          */
 135         public function __construct( ApiMain $main, $action ) {
 136                 parent::__construct( $main, $action );
 137
 138                 $this->mModuleMgr = new ApiModuleManager( $this );
 139
 140                 // Allow custom modules to be added in LocalSettings.php
 141                 $config = $this->getConfig();
 142                 $this->mModuleMgr->addModules( self::$QueryPropModules, 'prop' );
 143                 $this->mModuleMgr->addModules( $config->get( 'APIPropModules' ), 'prop' );
 144                 $this->mModuleMgr->addModules( self::$QueryListModules, 'list' );
 145                 $this->mModuleMgr->addModules( $config->get( 'APIListModules' ), 'list' );
 146                 $this->mModuleMgr->addModules( self::$QueryMetaModules, 'meta' );
 147                 $this->mModuleMgr->addModules( $config->get( 'APIMetaModules' ), 'meta' );
 148
 149                 Hooks::run( 'ApiQuery::moduleManager', [ $this->mModuleMgr ] );
 150
 151                 // Create PageSet that will process titles/pageids/revids/generator
 152                 $this->mPageSet = new ApiPageSet( $this );
 153         }
 154
 155         /**
 156          * Overrides to return this instance's module manager.
 157          * @return ApiModuleManager
 158          */
 159         public function getModuleManager() {
 160                 return $this->mModuleMgr;
 161         }
 162
 163         /**
 164          * Get the query database connection with the given name.
 165          * If no such connection has been requested before, it will be created.
 166          * Subsequent calls with the same $name will return the same connection
 167          * as the first, regardless of the values of $db and $groups
 168          * @param string $name Name to assign to the database connection
 169          * @param int $db One of the DB_* constants
 170          * @param array $groups Query groups
 171          * @return Database
 172          */
 173         public function getNamedDB( $name, $db, $groups ) {
 174                 if ( !array_key_exists( $name, $this->mNamedDB ) ) {
 175                         $this->mNamedDB[$name] = wfGetDB( $db, $groups );
 176                 }
 177
 178                 return $this->mNamedDB[$name];
 179         }
 180
 181         /**
 182          * Gets the set of pages the user has requested (or generated)
 183          * @return ApiPageSet
 184          */
 185         public function getPageSet() {
 186                 return $this->mPageSet;
 187         }
 188
 189         /**
 190          * @return ApiFormatRaw|null
 191          */
 192         public function getCustomPrinter() {
 193                 // If &exportnowrap is set, use the raw formatter
 194                 if ( $this->getParameter( 'export' ) &&
 195                         $this->getParameter( 'exportnowrap' )
 196                 ) {
 197                         return new ApiFormatRaw( $this->getMain(),
 198                                 $this->getMain()->createPrinterByName( 'xml' ) );
 199                 } else {
 200                         return null;
 201                 }
 202         }
 203
 204         /**
 205          * Query execution happens in the following steps:
 206          * #1 Create a PageSet object with any pages requested by the user
 207          * #2 If using a generator, execute it to get a new ApiPageSet object
 208          * #3 Instantiate all requested modules.
 209          *    This way the PageSet object will know what shared data is required,
 210          *    and minimize DB calls.
 211          * #4 Output all normalization and redirect resolution information
 212          * #5 Execute all requested modules
 213          */
 214         public function execute() {
 215                 $this->mParams = $this->extractRequestParams();
 216
 217                 // Instantiate requested modules
 218                 $allModules = [];
 219                 $this->instantiateModules( $allModules, 'prop' );
 220                 $propModules = array_keys( $allModules );
 221                 $this->instantiateModules( $allModules, 'list' );
 222                 $this->instantiateModules( $allModules, 'meta' );
 223
 224                 // Filter modules based on continue parameter
 225                 $continuationManager = new ApiContinuationManager( $this, $allModules, $propModules );
 226                 $this->setContinuationManager( $continuationManager );
 227                 $modules = $continuationManager->getRunModules();
 228
 229                 if ( !$continuationManager->isGeneratorDone() ) {
 230                         // Query modules may optimize data requests through the $this->getPageSet()
 231                         // object by adding extra fields from the page table.
 232                         foreach ( $modules as $module ) {
 233                                 $module->requestExtraData( $this->mPageSet );
 234                         }
 235                         // Populate page/revision information
 236                         $this->mPageSet->execute();
 237                         // Record page information (title, namespace, if exists, etc)
 238                         $this->outputGeneralPageInfo();
 239                 } else {
 240                         $this->mPageSet->executeDryRun();
 241                 }
 242
 243                 $cacheMode = $this->mPageSet->getCacheMode();
 244
 245                 // Execute all unfinished modules
 246                 /** @var $module ApiQueryBase */
 247                 foreach ( $modules as $module ) {
 248                         $params = $module->extractRequestParams();
 249                         $cacheMode = $this->mergeCacheMode(
 250                                 $cacheMode, $module->getCacheMode( $params ) );
 251                         $module->execute();
 252                         Hooks::run( 'APIQueryAfterExecute', [ &$module ] );
 253                 }
 254
 255                 // Set the cache mode
 256                 $this->getMain()->setCacheMode( $cacheMode );
 257
 258                 // Write the continuation data into the result
 259                 $this->setContinuationManager( null );
 260                 if ( $this->mParams['rawcontinue'] ) {
 261                         $data = $continuationManager->getRawNonContinuation();
 262                         if ( $data ) {
 263                                 $this->getResult()->addValue( null, 'query-noncontinue', $data,
 264                                         ApiResult::ADD_ON_TOP | ApiResult::NO_SIZE_CHECK );
 265                         }
 266                         $data = $continuationManager->getRawContinuation();
 267                         if ( $data ) {
 268                                 $this->getResult()->addValue( null, 'query-continue', $data,
 269                                         ApiResult::ADD_ON_TOP | ApiResult::NO_SIZE_CHECK );
 270                         }
 271                 } else {
 272                         $continuationManager->setContinuationIntoResult( $this->getResult() );
 273                 }
 274         }
 275
 276         /**
 277          * Update a cache mode string, applying the cache mode of a new module to it.
 278          * The cache mode may increase in the level of privacy, but public modules
 279          * added to private data do not decrease the level of privacy.
 280          *
 281          * @param string $cacheMode
 282          * @param string $modCacheMode
 283          * @return string
 284          */
 285         protected function mergeCacheMode( $cacheMode, $modCacheMode ) {
 286                 if ( $modCacheMode === 'anon-public-user-private' ) {
 287                         if ( $cacheMode !== 'private' ) {
 288                                 $cacheMode = 'anon-public-user-private';
 289                         }
 290                 } elseif ( $modCacheMode === 'public' ) {
 291                         // do nothing, if it's public already it will stay public
 292                 } else { // private
 293                         $cacheMode = 'private';
 294                 }
 295
 296                 return $cacheMode;
 297         }
 298
 299         /**
 300          * Create instances of all modules requested by the client
 301          * @param array $modules To append instantiated modules to
 302          * @param string $param Parameter name to read modules from
 303          */
 304         private function instantiateModules( &$modules, $param ) {
 305                 $wasPosted = $this->getRequest()->wasPosted();
 306                 if ( isset( $this->mParams[$param] ) ) {
 307                         foreach ( $this->mParams[$param] as $moduleName ) {
 308                                 $instance = $this->mModuleMgr->getModule( $moduleName, $param );
 309                                 if ( $instance === null ) {
 310                                         ApiBase::dieDebug( __METHOD__, 'Error instantiating module' );
 311                                 }
 312                                 if ( !$wasPosted && $instance->mustBePosted() ) {
 313                                         $this->dieWithErrorOrDebug( [ 'apierror-mustbeposted', $moduleName ] );
 314                                 }
 315                                 // Ignore duplicates. TODO 2.0: die()?
 316                                 if ( !array_key_exists( $moduleName, $modules ) ) {
 317                                         $modules[$moduleName] = $instance;
 318                                 }
 319                         }
 320                 }
 321         }
 322
 323         /**
 324          * Appends an element for each page in the current pageSet with the
 325          * most general information (id, title), plus any title normalizations
 326          * and missing or invalid title/pageids/revids.
 327          */
 328         private function outputGeneralPageInfo() {
 329                 $pageSet = $this->getPageSet();
 330                 $result = $this->getResult();
 331
 332                 // We can't really handle max-result-size failure here, but we need to
 333                 // check anyway in case someone set the limit stupidly low.
 334                 $fit = true;
 335
 336                 $values = $pageSet->getNormalizedTitlesAsResult( $result );
 337                 if ( $values ) {
 338                         $fit = $fit && $result->addValue( 'query', 'normalized', $values );
 339                 }
 340                 $values = $pageSet->getConvertedTitlesAsResult( $result );
 341                 if ( $values ) {
 342                         $fit = $fit && $result->addValue( 'query', 'converted', $values );
 343                 }
 344                 $values = $pageSet->getInterwikiTitlesAsResult( $result, $this->mParams['iwurl'] );
 345                 if ( $values ) {
 346                         $fit = $fit && $result->addValue( 'query', 'interwiki', $values );
 347                 }
 348                 $values = $pageSet->getRedirectTitlesAsResult( $result );
 349                 if ( $values ) {
 350                         $fit = $fit && $result->addValue( 'query', 'redirects', $values );
 351                 }
 352                 $values = $pageSet->getMissingRevisionIDsAsResult( $result );
 353                 if ( $values ) {
 354                         $fit = $fit && $result->addValue( 'query', 'badrevids', $values );
 355                 }
 356
 357                 // Page elements
 358                 $pages = [];
 359
 360                 // Report any missing titles
 361                 foreach ( $pageSet->getMissingTitles() as $fakeId => $title ) {
 362                         $vals = [];
 363                         ApiQueryBase::addTitleInfo( $vals, $title );
 364                         $vals['missing'] = true;
 365                         if ( $title->isKnown() ) {
 366                                 $vals['known'] = true;
 367                         }
 368                         $pages[$fakeId] = $vals;
 369                 }
 370                 // Report any invalid titles
 371                 foreach ( $pageSet->getInvalidTitlesAndReasons() as $fakeId => $data ) {
 372                         $pages[$fakeId] = $data + [ 'invalid' => true ];
 373                 }
 374                 // Report any missing page ids
 375                 foreach ( $pageSet->getMissingPageIDs() as $pageid ) {
 376                         $pages[$pageid] = [
 377                                 'pageid' => $pageid,
 378                                 'missing' => true,
 379                         ];
 380                 }
 381                 // Report special pages
 382                 /** @var $title Title */
 383                 foreach ( $pageSet->getSpecialTitles() as $fakeId => $title ) {
 384                         $vals = [];
 385                         ApiQueryBase::addTitleInfo( $vals, $title );
 386                         $vals['special'] = true;
 387                         if ( !$title->isKnown() ) {
 388                                 $vals['missing'] = true;
 389                         }
 390                         $pages[$fakeId] = $vals;
 391                 }
 392
 393                 // Output general page information for found titles
 394                 foreach ( $pageSet->getGoodTitles() as $pageid => $title ) {
 395                         $vals = [];
 396                         $vals['pageid'] = $pageid;
 397                         ApiQueryBase::addTitleInfo( $vals, $title );
 398                         $pages[$pageid] = $vals;
 399                 }
 400
 401                 if ( count( $pages ) ) {
 402                         $pageSet->populateGeneratorData( $pages );
 403                         ApiResult::setArrayType( $pages, 'BCarray' );
 404
 405                         if ( $this->mParams['indexpageids'] ) {
 406                                 $pageIDs = array_keys( ApiResult::stripMetadataNonRecursive( $pages ) );
 407                                 // json treats all map keys as strings - converting to match
 408                                 $pageIDs = array_map( 'strval', $pageIDs );
 409                                 ApiResult::setIndexedTagName( $pageIDs, 'id' );
 410                                 $fit = $fit && $result->addValue( 'query', 'pageids', $pageIDs );
 411                         }
 412
 413                         ApiResult::setIndexedTagName( $pages, 'page' );
 414                         $fit = $fit && $result->addValue( 'query', 'pages', $pages );
 415                 }
 416
 417                 if ( !$fit ) {
 418                         $this->dieWithError( 'apierror-badconfig-resulttoosmall', 'badconfig' );
 419                 }
 420
 421                 if ( $this->mParams['export'] ) {
 422                         $this->doExport( $pageSet, $result );
 423                 }
 424         }
 425
 426         /**
 427          * @param ApiPageSet $pageSet Pages to be exported
 428          * @param ApiResult $result Result to output to
 429          */
 430         private function doExport( $pageSet, $result ) {
 431                 $exportTitles = [];
 432                 $titles = $pageSet->getGoodTitles();
 433                 if ( count( $titles ) ) {
 434                         $user = $this->getUser();
 435                         /** @var $title Title */
 436                         foreach ( $titles as $title ) {
 437                                 if ( $title->userCan( 'read', $user ) ) {
 438                                         $exportTitles[] = $title;
 439                                 }
 440                         }
 441                 }
 442
 443                 $exporter = new WikiExporter( $this->getDB() );
 444                 $sink = new DumpStringOutput;
 445                 $exporter->setOutputSink( $sink );
 446                 $exporter->openStream();
 447                 foreach ( $exportTitles as $title ) {
 448                         $exporter->pageByTitle( $title );
 449                 }
 450                 $exporter->closeStream();
 451
 452                 // Don't check the size of exported stuff
 453                 // It's not continuable, so it would cause more
 454                 // problems than it'd solve
 455                 if ( $this->mParams['exportnowrap'] ) {
 456                         $result->reset();
 457                         // Raw formatter will handle this
 458                         $result->addValue( null, 'text', $sink, ApiResult::NO_SIZE_CHECK );
 459                         $result->addValue( null, 'mime', 'text/xml', ApiResult::NO_SIZE_CHECK );
 460                 } else {
 461                         $result->addValue( 'query', 'export', $sink, ApiResult::NO_SIZE_CHECK );
 462                         $result->addValue( 'query', ApiResult::META_BC_SUBELEMENTS, [ 'export' ] );
 463                 }
 464         }
 465
 466         public function getAllowedParams( $flags = 0 ) {
 467                 $result = [
 468                         'prop' => [
 469                                 ApiBase::PARAM_ISMULTI => true,
 470                                 ApiBase::PARAM_TYPE => 'submodule',
 471                         ],
 472                         'list' => [
 473                                 ApiBase::PARAM_ISMULTI => true,
 474                                 ApiBase::PARAM_TYPE => 'submodule',
 475                         ],
 476                         'meta' => [
 477                                 ApiBase::PARAM_ISMULTI => true,
 478                                 ApiBase::PARAM_TYPE => 'submodule',
 479                         ],
 480                         'indexpageids' => false,
 481                         'export' => false,
 482                         'exportnowrap' => false,
 483                         'iwurl' => false,
 484                         'continue' => [
 485                                 ApiBase::PARAM_HELP_MSG => 'api-help-param-continue',
 486                         ],
 487                         'rawcontinue' => false,
 488                 ];
 489                 if ( $flags ) {
 490                         $result += $this->getPageSet()->getFinalParams( $flags );
 491                 }
 492
 493                 return $result;
 494         }
 495
 496         public function isReadMode() {
 497                 // We need to make an exception for certain meta modules that should be
 498                 // accessible even without the 'read' right. Restrict the exception as
 499                 // much as possible: no other modules allowed, and no pageset
 500                 // parameters either. We do allow the 'rawcontinue' and 'indexpageids'
 501                 // parameters since frameworks might add these unconditionally and they
 502                 // can't expose anything here.
 503                 $this->mParams = $this->extractRequestParams();
 504                 $params = array_filter(
 505                         array_diff_key(
 506                                 $this->mParams + $this->getPageSet()->extractRequestParams(),
 507                                 [ 'rawcontinue' => 1, 'indexpageids' => 1 ]
 508                         )
 509                 );
 510                 if ( array_keys( $params ) !== [ 'meta' ] ) {
 511                         return true;
 512                 }
 513
 514                 // Ask each module if it requires read mode. Any true => this returns
 515                 // true.
 516                 $modules = [];
 517                 $this->instantiateModules( $modules, 'meta' );
 518                 foreach ( $modules as $module ) {
 519                         if ( $module->isReadMode() ) {
 520                                 return true;
 521                         }
 522                 }
 523
 524                 return false;
 525         }
 526
 527         protected function getExamplesMessages() {
 528                 return [
 529                         'action=query&prop=revisions&meta=siteinfo&' .
 530                                 'titles=Main%20Page&rvprop=user|comment&continue='
 531                                 => 'apihelp-query-example-revisions',
 532                         'action=query&generator=allpages&gapprefix=API/&prop=revisions&continue='
 533                                 => 'apihelp-query-example-allpages',
 534                 ];
 535         }
 536
 537         public function getHelpUrls() {
 538                 return [
 539                         'https://www.mediawiki.org/wiki/Special:MyLanguage/API:Query',
 540                         'https://www.mediawiki.org/wiki/Special:MyLanguage/API:Meta',
 541                         'https://www.mediawiki.org/wiki/Special:MyLanguage/API:Properties',
 542                         'https://www.mediawiki.org/wiki/Special:MyLanguage/API:Lists',
 543                 ];
 544         }
 545 }