includes/api/ApiQuery.php

   1 <?php
   2
   3 /*
   4  * Created on Sep 7, 2006
   5  *
   6  * API for MediaWiki 1.8+
   7  *
   8  * Copyright (C) 2006 Yuri Astrakhan <Firstname><Lastname>@gmail.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  * 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.
  23  * http://www.gnu.org/copyleft/gpl.html
  24  */
  25
  26 if ( !defined( 'MEDIAWIKI' ) ) {
  27         // Eclipse helper - will be ignored in production
  28         require_once ( 'ApiBase.php' );
  29 }
  30
  31 /**
  32  * This is the main query class. It behaves similar to ApiMain: based on the
  33  * parameters given, it will create a list of titles to work on (an ApiPageSet
  34  * object), instantiate and execute various property/list/meta modules, and
  35  * assemble all resulting data into a single ApiResult object.
  36  *
  37  * In generator mode, a generator will be executed first to populate a second
  38  * ApiPageSet object, and that object will be used for all subsequent modules.
  39  *
  40  * @ingroup API
  41  */
  42 class ApiQuery extends ApiBase {
  43
  44         private $mPropModuleNames, $mListModuleNames, $mMetaModuleNames;
  45         private $mPageSet;
  46         private $params, $redirect;
  47
  48         private $mQueryPropModules = array (
  49                 'info' => 'ApiQueryInfo',
  50                 'revisions' => 'ApiQueryRevisions',
  51                 'links' => 'ApiQueryLinks',
  52                 'langlinks' => 'ApiQueryLangLinks',
  53                 'images' => 'ApiQueryImages',
  54                 'imageinfo' => 'ApiQueryImageInfo',
  55                 'templates' => 'ApiQueryLinks',
  56                 'categories' => 'ApiQueryCategories',
  57                 'extlinks' => 'ApiQueryExternalLinks',
  58                 'categoryinfo' => 'ApiQueryCategoryInfo',
  59                 'duplicatefiles' => 'ApiQueryDuplicateFiles',
  60         );
  61
  62         private $mQueryListModules = array (
  63                 'allimages' => 'ApiQueryAllimages',
  64                 'allpages' => 'ApiQueryAllpages',
  65                 'alllinks' => 'ApiQueryAllLinks',
  66                 'allcategories' => 'ApiQueryAllCategories',
  67                 'allusers' => 'ApiQueryAllUsers',
  68                 'backlinks' => 'ApiQueryBacklinks',
  69                 'blocks' => 'ApiQueryBlocks',
  70                 'categorymembers' => 'ApiQueryCategoryMembers',
  71                 'deletedrevs' => 'ApiQueryDeletedrevs',
  72                 'embeddedin' => 'ApiQueryBacklinks',
  73                 'imageusage' => 'ApiQueryBacklinks',
  74                 'logevents' => 'ApiQueryLogEvents',
  75                 'recentchanges' => 'ApiQueryRecentChanges',
  76                 'search' => 'ApiQuerySearch',
  77                 'tags' => 'ApiQueryTags',
  78                 'usercontribs' => 'ApiQueryContributions',
  79                 'watchlist' => 'ApiQueryWatchlist',
  80                 'watchlistraw' => 'ApiQueryWatchlistRaw',
  81                 'exturlusage' => 'ApiQueryExtLinksUsage',
  82                 'users' => 'ApiQueryUsers',
  83                 'random' => 'ApiQueryRandom',
  84                 'protectedtitles' => 'ApiQueryProtectedTitles',
  85         );
  86
  87         private $mQueryMetaModules = array (
  88                 'siteinfo' => 'ApiQuerySiteinfo',
  89                 'userinfo' => 'ApiQueryUserInfo',
  90                 'allmessages' => 'ApiQueryAllmessages',
  91         );
  92
  93         private $mSlaveDB = null;
  94         private $mNamedDB = array();
  95
  96         public function __construct( $main, $action ) {
  97                 parent :: __construct( $main, $action );
  98
  99                 // Allow custom modules to be added in LocalSettings.php
 100                 global $wgAPIPropModules, $wgAPIListModules, $wgAPIMetaModules;
 101                 self :: appendUserModules( $this->mQueryPropModules, $wgAPIPropModules );
 102                 self :: appendUserModules( $this->mQueryListModules, $wgAPIListModules );
 103                 self :: appendUserModules( $this->mQueryMetaModules, $wgAPIMetaModules );
 104
 105                 $this->mPropModuleNames = array_keys( $this->mQueryPropModules );
 106                 $this->mListModuleNames = array_keys( $this->mQueryListModules );
 107                 $this->mMetaModuleNames = array_keys( $this->mQueryMetaModules );
 108
 109                 // Allow the entire list of modules at first,
 110                 // but during module instantiation check if it can be used as a generator.
 111                 $this->mAllowedGenerators = array_merge( $this->mListModuleNames, $this->mPropModuleNames );
 112         }
 113
 114         /**
 115          * Helper function to append any add-in modules to the list
 116          * @param $modules array Module array
 117          * @param $newModules array Module array to add to $modules
 118          */
 119         private static function appendUserModules( &$modules, $newModules ) {
 120                 if ( is_array( $newModules ) ) {
 121                         foreach ( $newModules as $moduleName => $moduleClass ) {
 122                                 $modules[$moduleName] = $moduleClass;
 123                         }
 124                 }
 125         }
 126
 127         /**
 128          * Gets a default slave database connection object
 129          * @return Database
 130          */
 131         public function getDB() {
 132                 if ( !isset ( $this->mSlaveDB ) ) {
 133                         $this->profileDBIn();
 134                         $this->mSlaveDB = wfGetDB( DB_SLAVE, 'api' );
 135                         $this->profileDBOut();
 136                 }
 137                 return $this->mSlaveDB;
 138         }
 139
 140         /**
 141          * Get the query database connection with the given name.
 142          * If no such connection has been requested before, it will be created.
 143          * Subsequent calls with the same $name will return the same connection
 144          * as the first, regardless of the values of $db and $groups
 145          * @param $name string Name to assign to the database connection
 146          * @param $db int One of the DB_* constants
 147          * @param $groups array Query groups
 148          * @return Database
 149          */
 150         public function getNamedDB( $name, $db, $groups ) {
 151                 if ( !array_key_exists( $name, $this->mNamedDB ) ) {
 152                         $this->profileDBIn();
 153                         $this->mNamedDB[$name] = wfGetDB( $db, $groups );
 154                         $this->profileDBOut();
 155                 }
 156                 return $this->mNamedDB[$name];
 157         }
 158
 159         /**
 160          * Gets the set of pages the user has requested (or generated)
 161          * @return ApiPageSet
 162          */
 163         public function getPageSet() {
 164                 return $this->mPageSet;
 165         }
 166
 167         /**
 168          * Get the array mapping module names to class names
 169          * @return array(modulename => classname)
 170          */
 171         function getModules() {
 172                 return array_merge( $this->mQueryPropModules, $this->mQueryListModules, $this->mQueryMetaModules );
 173         }
 174
 175         public function getCustomPrinter() {
 176                 // If &exportnowrap is set, use the raw formatter
 177                 if ( $this->getParameter( 'export' ) &&
 178                                 $this->getParameter( 'exportnowrap' ) )
 179                         return new ApiFormatRaw( $this->getMain(),
 180                                 $this->getMain()->createPrinterByName( 'xml' ) );
 181                 else
 182                         return null;
 183         }
 184
 185         /**
 186          * Query execution happens in the following steps:
 187          * #1 Create a PageSet object with any pages requested by the user
 188          * #2 If using a generator, execute it to get a new ApiPageSet object
 189          * #3 Instantiate all requested modules.
 190          *    This way the PageSet object will know what shared data is required,
 191          *    and minimize DB calls.
 192          * #4 Output all normalization and redirect resolution information
 193          * #5 Execute all requested modules
 194          */
 195         public function execute() {
 196
 197                 $this->params = $this->extractRequestParams();
 198                 $this->redirects = $this->params['redirects'];
 199
 200                 // Create PageSet
 201                 $this->mPageSet = new ApiPageSet( $this, $this->redirects );
 202
 203                 // Instantiate requested modules
 204                 $modules = array ();
 205                 $this->InstantiateModules( $modules, 'prop', $this->mQueryPropModules );
 206                 $this->InstantiateModules( $modules, 'list', $this->mQueryListModules );
 207                 $this->InstantiateModules( $modules, 'meta', $this->mQueryMetaModules );
 208
 209                 // If given, execute generator to substitute user supplied data with generated data.
 210                 if ( isset ( $this->params['generator'] ) ) {
 211                         $this->executeGeneratorModule( $this->params['generator'], $modules );
 212                 } else {
 213                         // Append custom fields and populate page/revision information
 214                         $this->addCustomFldsToPageSet( $modules, $this->mPageSet );
 215                         $this->mPageSet->execute();
 216                 }
 217
 218                 // Record page information (title, namespace, if exists, etc)
 219                 $this->outputGeneralPageInfo();
 220
 221                 // Execute all requested modules.
 222                 foreach ( $modules as $module ) {
 223                         $module->profileIn();
 224                         $module->execute();
 225                         wfRunHooks( 'APIQueryAfterExecute', array( &$module ) );
 226                         $module->profileOut();
 227                 }
 228         }
 229
 230         /**
 231          * Query modules may optimize data requests through the $this->getPageSet() object
 232          * by adding extra fields from the page table.
 233          * This function will gather all the extra request fields from the modules.
 234          * @param $modules array of module objects
 235          * @param $pageSet ApiPageSet
 236          */
 237         private function addCustomFldsToPageSet( $modules, $pageSet ) {
 238                 // Query all requested modules.
 239                 foreach ( $modules as $module ) {
 240                         $module->requestExtraData( $pageSet );
 241                 }
 242         }
 243
 244         /**
 245          * Create instances of all modules requested by the client
 246          * @param $modules array to append instatiated modules to
 247          * @param $param string Parameter name to read modules from
 248          * @param $moduleList array(modulename => classname)
 249          */
 250         private function InstantiateModules( &$modules, $param, $moduleList ) {
 251                 $list = @$this->params[$param];
 252                 if ( !is_null ( $list ) )
 253                         foreach ( $list as $moduleName )
 254                                 $modules[] = new $moduleList[$moduleName] ( $this, $moduleName );
 255         }
 256
 257         /**
 258          * Appends an element for each page in the current pageSet with the
 259          * most general information (id, title), plus any title normalizations
 260          * and missing or invalid title/pageids/revids.
 261          */
 262         private function outputGeneralPageInfo() {
 263
 264                 $pageSet = $this->getPageSet();
 265                 $result = $this->getResult();
 266
 267                 // We don't check for a full result set here because we can't be adding
 268                 // more than 380K. The maximum revision size is in the megabyte range,
 269                 // and the maximum result size must be even higher than that.
 270
 271                 // Title normalizations
 272                 $normValues = array ();
 273                 foreach ( $pageSet->getNormalizedTitles() as $rawTitleStr => $titleStr ) {
 274                         $normValues[] = array (
 275                                 'from' => $rawTitleStr,
 276                                 'to' => $titleStr
 277                         );
 278                 }
 279
 280                 if ( count( $normValues ) ) {
 281                         $result->setIndexedTagName( $normValues, 'n' );
 282                         $result->addValue( 'query', 'normalized', $normValues );
 283                 }
 284
 285                 // Interwiki titles
 286                 $intrwValues = array ();
 287                 foreach ( $pageSet->getInterwikiTitles() as $rawTitleStr => $interwikiStr ) {
 288                         $intrwValues[] = array (
 289                                 'title' => $rawTitleStr,
 290                                 'iw' => $interwikiStr
 291                         );
 292                 }
 293
 294                 if ( count( $intrwValues ) ) {
 295                         $result->setIndexedTagName( $intrwValues, 'i' );
 296                         $result->addValue( 'query', 'interwiki', $intrwValues );
 297                 }
 298
 299                 // Show redirect information
 300                 $redirValues = array ();
 301                 foreach ( $pageSet->getRedirectTitles() as $titleStrFrom => $titleStrTo ) {
 302                         $redirValues[] = array (
 303                                 'from' => strval( $titleStrFrom ),
 304                                 'to' => $titleStrTo
 305                         );
 306                 }
 307
 308                 if ( count( $redirValues ) ) {
 309                         $result->setIndexedTagName( $redirValues, 'r' );
 310                         $result->addValue( 'query', 'redirects', $redirValues );
 311                 }
 312
 313                 //
 314                 // Missing revision elements
 315                 //
 316                 $missingRevIDs = $pageSet->getMissingRevisionIDs();
 317                 if ( count( $missingRevIDs ) ) {
 318                         $revids = array ();
 319                         foreach ( $missingRevIDs as $revid ) {
 320                                 $revids[$revid] = array (
 321                                         'revid' => $revid
 322                                 );
 323                         }
 324                         $result->setIndexedTagName( $revids, 'rev' );
 325                         $result->addValue( 'query', 'badrevids', $revids );
 326                 }
 327
 328                 //
 329                 // Page elements
 330                 //
 331                 $pages = array ();
 332
 333                 // Report any missing titles
 334                 foreach ( $pageSet->getMissingTitles() as $fakeId => $title ) {
 335                         $vals = array();
 336                         ApiQueryBase :: addTitleInfo( $vals, $title );
 337                         $vals['missing'] = '';
 338                         $pages[$fakeId] = $vals;
 339                 }
 340                 // Report any invalid titles
 341                 foreach ( $pageSet->getInvalidTitles() as $fakeId => $title )
 342                         $pages[$fakeId] = array( 'title' => $title, 'invalid' => '' );
 343                 // Report any missing page ids
 344                 foreach ( $pageSet->getMissingPageIDs() as $pageid ) {
 345                         $pages[$pageid] = array (
 346                                 'pageid' => $pageid,
 347                                 'missing' => ''
 348                         );
 349                 }
 350
 351                 // Output general page information for found titles
 352                 foreach ( $pageSet->getGoodTitles() as $pageid => $title ) {
 353                         $vals = array();
 354                         $vals['pageid'] = $pageid;
 355                         ApiQueryBase :: addTitleInfo( $vals, $title );
 356                         $pages[$pageid] = $vals;
 357                 }
 358
 359                 if ( count( $pages ) ) {
 360
 361                         if ( $this->params['indexpageids'] ) {
 362                                 $pageIDs = array_keys( $pages );
 363                                 // json treats all map keys as strings - converting to match
 364                                 $pageIDs = array_map( 'strval', $pageIDs );
 365                                 $result->setIndexedTagName( $pageIDs, 'id' );
 366                                 $result->addValue( 'query', 'pageids', $pageIDs );
 367                         }
 368
 369                         $result->setIndexedTagName( $pages, 'page' );
 370                         $result->addValue( 'query', 'pages', $pages );
 371                 }
 372                 if ( $this->params['export'] ) {
 373                         $exporter = new WikiExporter( $this->getDB() );
 374                         // WikiExporter writes to stdout, so catch its
 375                         // output with an ob
 376                         ob_start();
 377                         $exporter->openStream();
 378                         foreach ( @$pageSet->getGoodTitles() as $title )
 379                                 if ( $title->userCanRead() )
 380                                         $exporter->pageByTitle( $title );
 381                         $exporter->closeStream();
 382                         $exportxml = ob_get_contents();
 383                         ob_end_clean();
 384
 385                         // Don't check the size of exported stuff
 386                         // It's not continuable, so it would cause more
 387                         // problems than it'd solve
 388                         $result->disableSizeCheck();
 389                         if ( $this->params['exportnowrap'] ) {
 390                                 $result->reset();
 391                                 // Raw formatter will handle this
 392                                 $result->addValue( null, 'text', $exportxml );
 393                                 $result->addValue( null, 'mime', 'text/xml' );
 394                         } else {
 395                                 $r = array();
 396                                 ApiResult::setContent( $r, $exportxml );
 397                                 $result->addValue( 'query', 'export', $r );
 398                         }
 399                         $result->enableSizeCheck();
 400                 }
 401         }
 402
 403         /**
 404          * For generator mode, execute generator, and use its output as new
 405          * ApiPageSet
 406          * @param $generatorName string Module name
 407          * @param $modules array of module objects
 408          */
 409         protected function executeGeneratorModule( $generatorName, $modules ) {
 410
 411                 // Find class that implements requested generator
 412                 if ( isset ( $this->mQueryListModules[$generatorName] ) ) {
 413                         $className = $this->mQueryListModules[$generatorName];
 414                 } elseif ( isset ( $this->mQueryPropModules[$generatorName] ) ) {
 415                         $className = $this->mQueryPropModules[$generatorName];
 416                 } else {
 417                         ApiBase :: dieDebug( __METHOD__, "Unknown generator=$generatorName" );
 418                 }
 419
 420                 // Generator results
 421                 $resultPageSet = new ApiPageSet( $this, $this->redirects );
 422
 423                 // Create and execute the generator
 424                 $generator = new $className ( $this, $generatorName );
 425                 if ( !$generator instanceof ApiQueryGeneratorBase )
 426                         $this->dieUsage( "Module $generatorName cannot be used as a generator", "badgenerator" );
 427
 428                 $generator->setGeneratorMode();
 429
 430                 // Add any additional fields modules may need
 431                 $generator->requestExtraData( $this->mPageSet );
 432                 $this->addCustomFldsToPageSet( $modules, $resultPageSet );
 433
 434                 // Populate page information with the original user input
 435                 $this->mPageSet->execute();
 436
 437                 // populate resultPageSet with the generator output
 438                 $generator->profileIn();
 439                 $generator->executeGenerator( $resultPageSet );
 440                 wfRunHooks( 'APIQueryGeneratorAfterExecute', array( &$generator, &$resultPageSet ) );
 441                 $resultPageSet->finishPageSetGeneration();
 442                 $generator->profileOut();
 443
 444                 // Swap the resulting pageset back in
 445                 $this->mPageSet = $resultPageSet;
 446         }
 447
 448         public function getAllowedParams() {
 449                 return array (
 450                         'prop' => array (
 451                                 ApiBase :: PARAM_ISMULTI => true,
 452                                 ApiBase :: PARAM_TYPE => $this->mPropModuleNames
 453                         ),
 454                         'list' => array (
 455                                 ApiBase :: PARAM_ISMULTI => true,
 456                                 ApiBase :: PARAM_TYPE => $this->mListModuleNames
 457                         ),
 458                         'meta' => array (
 459                                 ApiBase :: PARAM_ISMULTI => true,
 460                                 ApiBase :: PARAM_TYPE => $this->mMetaModuleNames
 461                         ),
 462                         'generator' => array (
 463                                 ApiBase :: PARAM_TYPE => $this->mAllowedGenerators
 464                         ),
 465                         'redirects' => false,
 466                         'indexpageids' => false,
 467                         'export' => false,
 468                         'exportnowrap' => false,
 469                 );
 470         }
 471
 472         /**
 473          * Override the parent to generate help messages for all available query modules.
 474          * @return string
 475          */
 476         public function makeHelpMsg() {
 477
 478                 $msg = '';
 479
 480                 // Make sure the internal object is empty
 481                 // (just in case a sub-module decides to optimize during instantiation)
 482                 $this->mPageSet = null;
 483                 $this->mAllowedGenerators = array();    // Will be repopulated
 484
 485                 $astriks = str_repeat( '--- ', 8 );
 486                 $astriks2 = str_repeat( '*** ', 10 );
 487                 $msg .= "\n$astriks Query: Prop  $astriks\n\n";
 488                 $msg .= $this->makeHelpMsgHelper( $this->mQueryPropModules, 'prop' );
 489                 $msg .= "\n$astriks Query: List  $astriks\n\n";
 490                 $msg .= $this->makeHelpMsgHelper( $this->mQueryListModules, 'list' );
 491                 $msg .= "\n$astriks Query: Meta  $astriks\n\n";
 492                 $msg .= $this->makeHelpMsgHelper( $this->mQueryMetaModules, 'meta' );
 493                 $msg .= "\n\n$astriks2 Modules: continuation  $astriks2\n\n";
 494
 495                 // Perform the base call last because the $this->mAllowedGenerators
 496                 // will be updated inside makeHelpMsgHelper()
 497                 // Use parent to make default message for the query module
 498                 $msg = parent :: makeHelpMsg() . $msg;
 499
 500                 return $msg;
 501         }
 502
 503         /**
 504          * For all modules in $moduleList, generate help messages and join them together
 505          * @param $moduleList array(modulename => classname)
 506          * @param $paramName string Parameter name
 507          * @return string
 508          */
 509         private function makeHelpMsgHelper( $moduleList, $paramName ) {
 510
 511                 $moduleDescriptions = array ();
 512
 513                 foreach ( $moduleList as $moduleName => $moduleClass ) {
 514                         $module = new $moduleClass ( $this, $moduleName, null );
 515
 516                         $msg = ApiMain::makeHelpMsgHeader( $module, $paramName );
 517                         $msg2 = $module->makeHelpMsg();
 518                         if ( $msg2 !== false )
 519                                 $msg .= $msg2;
 520                         if ( $module instanceof ApiQueryGeneratorBase ) {
 521                                 $this->mAllowedGenerators[] = $moduleName;
 522                                 $msg .= "Generator:\n  This module may be used as a generator\n";
 523                         }
 524                         $moduleDescriptions[] = $msg;
 525                 }
 526
 527                 return implode( "\n", $moduleDescriptions );
 528         }
 529
 530         /**
 531          * Override to add extra parameters from PageSet
 532          * @return string
 533          */
 534         public function makeHelpMsgParameters() {
 535                 $psModule = new ApiPageSet( $this );
 536                 return $psModule->makeHelpMsgParameters() . parent :: makeHelpMsgParameters();
 537         }
 538
 539         public function shouldCheckMaxlag() {
 540                 return true;
 541         }
 542
 543         public function getParamDescription() {
 544                 return array (
 545                         'prop' => 'Which properties to get for the titles/revisions/pageids',
 546                         'list' => 'Which lists to get',
 547                         'meta' => 'Which meta data to get about the site',
 548                         'generator' => array( 'Use the output of a list as the input for other prop/list/meta items',
 549                                         'NOTE: generator parameter names must be prefixed with a \'g\', see examples.' ),
 550                         'redirects' => 'Automatically resolve redirects',
 551                         'indexpageids' => 'Include an additional pageids section listing all returned page IDs.',
 552                         'export' => 'Export the current revisions of all given or generated pages',
 553                         'exportnowrap' => 'Return the export XML without wrapping it in an XML result (same format as Special:Export). Can only be used with export',
 554                 );
 555         }
 556
 557         public function getDescription() {
 558                 return array (
 559                         'Query API module allows applications to get needed pieces of data from the MediaWiki databases,',
 560                         'and is loosely based on the old query.php interface.',
 561                         'All data modifications will first have to use query to acquire a token to prevent abuse from malicious sites.'
 562                 );
 563         }
 564
 565         protected function getExamples() {
 566                 return array (
 567                         'api.php?action=query&prop=revisions&meta=siteinfo&titles=Main%20Page&rvprop=user|comment',
 568                         'api.php?action=query&generator=allpages&gapprefix=API/&prop=revisions',
 569                 );
 570         }
 571
 572         public function getVersion() {
 573                 $psModule = new ApiPageSet( $this );
 574                 $vers = array ();
 575                 $vers[] = __CLASS__ . ': $Id$';
 576                 $vers[] = $psModule->getVersion();
 577                 return $vers;
 578         }
 579 }