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         #'brokenredirects' => 'ApiQueryBrokenRedirects',
  71                 'categorymembers' => 'ApiQueryCategoryMembers',
  72         #'doubleredirects' => 'ApiQueryDoubleRedirects',
  73                 'deletedrevs' => 'ApiQueryDeletedrevs',
  74                 'embeddedin' => 'ApiQueryBacklinks',
  75                 'imageusage' => 'ApiQueryBacklinks',
  76                 'logevents' => 'ApiQueryLogEvents',
  77                 'recentchanges' => 'ApiQueryRecentChanges',
  78                 'search' => 'ApiQuerySearch',
  79                 'usercontribs' => 'ApiQueryContributions',
  80                 'watchlist' => 'ApiQueryWatchlist',
  81                 'watchlistraw' => 'ApiQueryWatchlistRaw',
  82                 'exturlusage' => 'ApiQueryExtLinksUsage',
  83                 'users' => 'ApiQueryUsers',
  84                 'random' => 'ApiQueryRandom',
  85                 'protectedtitles' => 'ApiQueryProtectedTitles',
  86         );
  87
  88         private $mQueryMetaModules = array (
  89                 'siteinfo' => 'ApiQuerySiteinfo',
  90                 'userinfo' => 'ApiQueryUserInfo',
  91                 'allmessages' => 'ApiQueryAllmessages',
  92         );
  93
  94         private $mSlaveDB = null;
  95         private $mNamedDB = array();
  96
  97         public function __construct($main, $action) {
  98                 parent :: __construct($main, $action);
  99
 100                 // Allow custom modules to be added in LocalSettings.php
 101                 global $wgAPIPropModules, $wgAPIListModules, $wgAPIMetaModules;
 102                 self :: appendUserModules($this->mQueryPropModules, $wgAPIPropModules);
 103                 self :: appendUserModules($this->mQueryListModules, $wgAPIListModules);
 104                 self :: appendUserModules($this->mQueryMetaModules, $wgAPIMetaModules);
 105
 106                 $this->mPropModuleNames = array_keys($this->mQueryPropModules);
 107                 $this->mListModuleNames = array_keys($this->mQueryListModules);
 108                 $this->mMetaModuleNames = array_keys($this->mQueryMetaModules);
 109
 110                 // Allow the entire list of modules at first,
 111                 // but during module instantiation check if it can be used as a generator.
 112                 $this->mAllowedGenerators = array_merge($this->mListModuleNames, $this->mPropModuleNames);
 113         }
 114
 115         /**
 116          * Helper function to append any add-in modules to the list
 117          * @param $modules array Module array
 118          * @param $newModules array Module array to add to $modules
 119          */
 120         private static function appendUserModules(&$modules, $newModules) {
 121                 if (is_array( $newModules )) {
 122                         foreach ( $newModules as $moduleName => $moduleClass) {
 123                                 $modules[$moduleName] = $moduleClass;
 124                         }
 125                 }
 126         }
 127
 128         /**
 129          * Gets a default slave database connection object
 130          * @return Database
 131          */
 132         public function getDB() {
 133                 if (!isset ($this->mSlaveDB)) {
 134                         $this->profileDBIn();
 135                         $this->mSlaveDB = wfGetDB(DB_SLAVE,'api');
 136                         $this->profileDBOut();
 137                 }
 138                 return $this->mSlaveDB;
 139         }
 140
 141         /**
 142          * Get the query database connection with the given name.
 143          * If no such connection has been requested before, it will be created.
 144          * Subsequent calls with the same $name will return the same connection
 145          * as the first, regardless of the values of $db and $groups
 146          * @param $name string Name to assign to the database connection
 147          * @param $db int One of the DB_* constants
 148          * @param $groups array Query groups
 149          * @return Database
 150          */
 151         public function getNamedDB($name, $db, $groups) {
 152                 if (!array_key_exists($name, $this->mNamedDB)) {
 153                         $this->profileDBIn();
 154                         $this->mNamedDB[$name] = wfGetDB($db, $groups);
 155                         $this->profileDBOut();
 156                 }
 157                 return $this->mNamedDB[$name];
 158         }
 159
 160         /**
 161          * Gets the set of pages the user has requested (or generated)
 162          * @return ApiPageSet
 163          */
 164         public function getPageSet() {
 165                 return $this->mPageSet;
 166         }
 167
 168         /**
 169          * Get the array mapping module names to class names
 170          * @return array(modulename => classname)
 171          */
 172         function getModules() {
 173                 return array_merge($this->mQueryPropModules, $this->mQueryListModules, $this->mQueryMetaModules);
 174         }
 175
 176         public function getCustomPrinter() {
 177                 // If &exportnowrap is set, use the raw formatter
 178                 if ($this->getParameter('export') &&
 179                                 $this->getParameter('exportnowrap'))
 180                         return new ApiFormatRaw($this->getMain(),
 181                                 $this->getMain()->createPrinterByName('xml'));
 182                 else
 183                         return null;
 184         }
 185
 186         /**
 187          * Query execution happens in the following steps:
 188          * #1 Create a PageSet object with any pages requested by the user
 189          * #2 If using a generator, execute it to get a new ApiPageSet object
 190          * #3 Instantiate all requested modules.
 191          *    This way the PageSet object will know what shared data is required,
 192          *    and minimize DB calls.
 193          * #4 Output all normalization and redirect resolution information
 194          * #5 Execute all requested modules
 195          */
 196         public function execute() {
 197
 198                 $this->params = $this->extractRequestParams();
 199                 $this->redirects = $this->params['redirects'];
 200
 201                 //
 202                 // Create PageSet
 203                 //
 204                 $this->mPageSet = new ApiPageSet($this, $this->redirects);
 205
 206                 //
 207                 // Instantiate requested modules
 208                 //
 209                 $modules = array ();
 210                 $this->InstantiateModules($modules, 'prop', $this->mQueryPropModules);
 211                 $this->InstantiateModules($modules, 'list', $this->mQueryListModules);
 212                 $this->InstantiateModules($modules, 'meta', $this->mQueryMetaModules);
 213
 214                 //
 215                 // If given, execute generator to substitute user supplied data with generated data.
 216                 //
 217                 if (isset ($this->params['generator'])) {
 218                         $this->executeGeneratorModule($this->params['generator'], $modules);
 219                 } else {
 220                         // Append custom fields and populate page/revision information
 221                         $this->addCustomFldsToPageSet($modules, $this->mPageSet);
 222                         $this->mPageSet->execute();
 223                 }
 224
 225                 //
 226                 // Record page information (title, namespace, if exists, etc)
 227                 //
 228                 $this->outputGeneralPageInfo();
 229
 230                 //
 231                 // Execute all requested modules.
 232                 //
 233                 foreach ($modules as $module) {
 234                         $module->profileIn();
 235                         $module->execute();
 236                         wfRunHooks('APIQueryAfterExecute', array(&$module));
 237                         $module->profileOut();
 238                 }
 239         }
 240
 241         /**
 242          * Query modules may optimize data requests through the $this->getPageSet() object
 243          * by adding extra fields from the page table.
 244          * This function will gather all the extra request fields from the modules.
 245          * @param $modules array of module objects
 246          * @param $pageSet ApiPageSet
 247          */
 248         private function addCustomFldsToPageSet($modules, $pageSet) {
 249                 // Query all requested modules.
 250                 foreach ($modules as $module) {
 251                         $module->requestExtraData($pageSet);
 252                 }
 253         }
 254
 255         /**
 256          * Create instances of all modules requested by the client
 257          * @param $modules array to append instatiated modules to
 258          * @param $param string Parameter name to read modules from
 259          * @param $moduleList array(modulename => classname)
 260          */
 261         private function InstantiateModules(&$modules, $param, $moduleList) {
 262                 $list = @$this->params[$param];
 263                 if (!is_null ($list))
 264                         foreach ($list as $moduleName)
 265                                 $modules[] = new $moduleList[$moduleName] ($this, $moduleName);
 266         }
 267
 268         /**
 269          * Appends an element for each page in the current pageSet with the
 270          * most general information (id, title), plus any title normalizations
 271          * and missing or invalid title/pageids/revids.
 272          */
 273         private function outputGeneralPageInfo() {
 274
 275                 $pageSet = $this->getPageSet();
 276                 $result = $this->getResult();
 277
 278                 # We don't check for a full result set here because we can't be adding
 279                 # more than 380K. The maximum revision size is in the megabyte range,
 280                 # and the maximum result size must be even higher than that.
 281
 282                 // Title normalizations
 283                 $normValues = array ();
 284                 foreach ($pageSet->getNormalizedTitles() as $rawTitleStr => $titleStr) {
 285                         $normValues[] = array (
 286                                 'from' => $rawTitleStr,
 287                                 'to' => $titleStr
 288                         );
 289                 }
 290
 291                 if (count($normValues)) {
 292                         $result->setIndexedTagName($normValues, 'n');
 293                         $result->addValue('query', 'normalized', $normValues);
 294                 }
 295
 296                 // Interwiki titles
 297                 $intrwValues = array ();
 298                 foreach ($pageSet->getInterwikiTitles() as $rawTitleStr => $interwikiStr) {
 299                         $intrwValues[] = array (
 300                                 'title' => $rawTitleStr,
 301                                 'iw' => $interwikiStr
 302                         );
 303                 }
 304
 305                 if (count($intrwValues)) {
 306                         $result->setIndexedTagName($intrwValues, 'i');
 307                         $result->addValue('query', 'interwiki', $intrwValues);
 308                 }
 309
 310                 // Show redirect information
 311                 $redirValues = array ();
 312                 foreach ($pageSet->getRedirectTitles() as $titleStrFrom => $titleStrTo) {
 313                         $redirValues[] = array (
 314                                 'from' => strval($titleStrFrom),
 315                                 'to' => $titleStrTo
 316                         );
 317                 }
 318
 319                 if (count($redirValues)) {
 320                         $result->setIndexedTagName($redirValues, 'r');
 321                         $result->addValue('query', 'redirects', $redirValues);
 322                 }
 323
 324                 //
 325                 // Missing revision elements
 326                 //
 327                 $missingRevIDs = $pageSet->getMissingRevisionIDs();
 328                 if (count($missingRevIDs)) {
 329                         $revids = array ();
 330                         foreach ($missingRevIDs as $revid) {
 331                                 $revids[$revid] = array (
 332                                         'revid' => $revid
 333                                 );
 334                         }
 335                         $result->setIndexedTagName($revids, 'rev');
 336                         $result->addValue('query', 'badrevids', $revids);
 337                 }
 338
 339                 //
 340                 // Page elements
 341                 //
 342                 $pages = array ();
 343
 344                 // Report any missing titles
 345                 foreach ($pageSet->getMissingTitles() as $fakeId => $title) {
 346                         $vals = array();
 347                         ApiQueryBase :: addTitleInfo($vals, $title);
 348                         $vals['missing'] = '';
 349                         $pages[$fakeId] = $vals;
 350                 }
 351                 // Report any invalid titles
 352                 foreach ($pageSet->getInvalidTitles() as $fakeId => $title)
 353                         $pages[$fakeId] = array('title' => $title, 'invalid' => '');
 354                 // Report any missing page ids
 355                 foreach ($pageSet->getMissingPageIDs() as $pageid) {
 356                         $pages[$pageid] = array (
 357                                 'pageid' => $pageid,
 358                                 'missing' => ''
 359                         );
 360                 }
 361
 362                 // Output general page information for found titles
 363                 foreach ($pageSet->getGoodTitles() as $pageid => $title) {
 364                         $vals = array();
 365                         $vals['pageid'] = $pageid;
 366                         ApiQueryBase :: addTitleInfo($vals, $title);
 367                         $pages[$pageid] = $vals;
 368                 }
 369
 370                 if (count($pages)) {
 371
 372                         if ($this->params['indexpageids']) {
 373                                 $pageIDs = array_keys($pages);
 374                                 // json treats all map keys as strings - converting to match
 375                                 $pageIDs = array_map('strval', $pageIDs);
 376                                 $result->setIndexedTagName($pageIDs, 'id');
 377                                 $result->addValue('query', 'pageids', $pageIDs);
 378                         }
 379
 380                         $result->setIndexedTagName($pages, 'page');
 381                         $result->addValue('query', 'pages', $pages);
 382                 }
 383                 if ($this->params['export']) {
 384                         $exporter = new WikiExporter($this->getDB());
 385                         // WikiExporter writes to stdout, so catch its
 386                         // output with an ob
 387                         ob_start();
 388                         $exporter->openStream();
 389                         foreach (@$pageSet->getGoodTitles() as $title)
 390                                 if ($title->userCanRead())
 391                                         $exporter->pageByTitle($title);
 392                         $exporter->closeStream();
 393                         $exportxml = ob_get_contents();
 394                         ob_end_clean();
 395                         // Don't check the size of exported stuff
 396                         // It's not continuable, so it would cause more
 397                         // problems than it'd solve
 398                         $result->disableSizeCheck();
 399                         if ($this->params['exportnowrap']) {
 400                                 $result->reset();
 401                                 // Raw formatter will handle this
 402                                 $result->addValue(null, 'text', $exportxml);
 403                                 $result->addValue(null, 'mime', 'text/xml');
 404                         } else {
 405                                 $r = array();
 406                                 ApiResult::setContent($r, $exportxml);
 407                                 $result->addValue('query', 'export', $r);
 408                         }
 409                         $result->enableSizeCheck();
 410                 }
 411         }
 412
 413         /**
 414          * For generator mode, execute generator, and use its output as new
 415          * ApiPageSet
 416          * @param $generatorName string Module name
 417          * @param $modules array of module objects
 418          */
 419         protected function executeGeneratorModule($generatorName, $modules) {
 420
 421                 // Find class that implements requested generator
 422                 if (isset ($this->mQueryListModules[$generatorName])) {
 423                         $className = $this->mQueryListModules[$generatorName];
 424                 } elseif (isset ($this->mQueryPropModules[$generatorName])) {
 425                         $className = $this->mQueryPropModules[$generatorName];
 426                 } else {
 427                         ApiBase :: dieDebug(__METHOD__, "Unknown generator=$generatorName");
 428                 }
 429
 430                 // Generator results
 431                 $resultPageSet = new ApiPageSet($this, $this->redirects);
 432
 433                 // Create and execute the generator
 434                 $generator = new $className ($this, $generatorName);
 435                 if (!$generator instanceof ApiQueryGeneratorBase)
 436                         $this->dieUsage("Module $generatorName cannot be used as a generator", "badgenerator");
 437
 438                 $generator->setGeneratorMode();
 439
 440                 // Add any additional fields modules may need
 441                 $generator->requestExtraData($this->mPageSet);
 442                 $this->addCustomFldsToPageSet($modules, $resultPageSet);
 443
 444                 // Populate page information with the original user input
 445                 $this->mPageSet->execute();
 446
 447                 // populate resultPageSet with the generator output
 448                 $generator->profileIn();
 449                 $generator->executeGenerator($resultPageSet);
 450                 wfRunHooks('APIQueryGeneratorAfterExecute', array(&$generator, &$resultPageSet));
 451                 $resultPageSet->finishPageSetGeneration();
 452                 $generator->profileOut();
 453
 454                 // Swap the resulting pageset back in
 455                 $this->mPageSet = $resultPageSet;
 456         }
 457
 458         public function getAllowedParams() {
 459                 return array (
 460                         'prop' => array (
 461                                 ApiBase :: PARAM_ISMULTI => true,
 462                                 ApiBase :: PARAM_TYPE => $this->mPropModuleNames
 463                         ),
 464                         'list' => array (
 465                                 ApiBase :: PARAM_ISMULTI => true,
 466                                 ApiBase :: PARAM_TYPE => $this->mListModuleNames
 467                         ),
 468                         'meta' => array (
 469                                 ApiBase :: PARAM_ISMULTI => true,
 470                                 ApiBase :: PARAM_TYPE => $this->mMetaModuleNames
 471                         ),
 472                         'generator' => array (
 473                                 ApiBase :: PARAM_TYPE => $this->mAllowedGenerators
 474                         ),
 475                         'redirects' => false,
 476                         'indexpageids' => false,
 477                         'export' => false,
 478                         'exportnowrap' => false,
 479                 );
 480         }
 481
 482         /**
 483          * Override the parent to generate help messages for all available query modules.
 484          * @return string
 485          */
 486         public function makeHelpMsg() {
 487
 488                 $msg = '';
 489
 490                 // Make sure the internal object is empty
 491                 // (just in case a sub-module decides to optimize during instantiation)
 492                 $this->mPageSet = null;
 493                 $this->mAllowedGenerators = array();    // Will be repopulated
 494
 495                 $astriks = str_repeat('--- ', 8);
 496                 $astriks2 = str_repeat('*** ', 10);
 497                 $msg .= "\n$astriks Query: Prop  $astriks\n\n";
 498                 $msg .= $this->makeHelpMsgHelper($this->mQueryPropModules, 'prop');
 499                 $msg .= "\n$astriks Query: List  $astriks\n\n";
 500                 $msg .= $this->makeHelpMsgHelper($this->mQueryListModules, 'list');
 501                 $msg .= "\n$astriks Query: Meta  $astriks\n\n";
 502                 $msg .= $this->makeHelpMsgHelper($this->mQueryMetaModules, 'meta');
 503                 $msg .= "\n\n$astriks2 Modules: continuation  $astriks2\n\n";
 504
 505                 // Perform the base call last because the $this->mAllowedGenerators
 506                 // will be updated inside makeHelpMsgHelper()
 507                 // Use parent to make default message for the query module
 508                 $msg = parent :: makeHelpMsg() . $msg;
 509
 510                 return $msg;
 511         }
 512
 513         /**
 514          * For all modules in $moduleList, generate help messages and join them together
 515          * @param $moduleList array(modulename => classname)
 516          * @param $paramName string Parameter name
 517          * @return string
 518          */
 519         private function makeHelpMsgHelper($moduleList, $paramName) {
 520
 521                 $moduleDescriptions = array ();
 522
 523                 foreach ($moduleList as $moduleName => $moduleClass) {
 524                         $module = new $moduleClass ($this, $moduleName, null);
 525
 526                         $msg = ApiMain::makeHelpMsgHeader($module, $paramName);
 527                         $msg2 = $module->makeHelpMsg();
 528                         if ($msg2 !== false)
 529                                 $msg .= $msg2;
 530                         if ($module instanceof ApiQueryGeneratorBase) {
 531                                 $this->mAllowedGenerators[] = $moduleName;
 532                                 $msg .= "Generator:\n  This module may be used as a generator\n";
 533                         }
 534                         $moduleDescriptions[] = $msg;
 535                 }
 536
 537                 return implode("\n", $moduleDescriptions);
 538         }
 539
 540         /**
 541          * Override to add extra parameters from PageSet
 542          * @return string
 543          */
 544         public function makeHelpMsgParameters() {
 545                 $psModule = new ApiPageSet($this);
 546                 return $psModule->makeHelpMsgParameters() . parent :: makeHelpMsgParameters();
 547         }
 548
 549         public function shouldCheckMaxlag() {
 550                 return true;
 551         }
 552
 553         public function getParamDescription() {
 554                 return array (
 555                         'prop' => 'Which properties to get for the titles/revisions/pageids',
 556                         'list' => 'Which lists to get',
 557                         'meta' => 'Which meta data to get about the site',
 558                         'generator' => array('Use the output of a list as the input for other prop/list/meta items',
 559                                         'NOTE: generator parameter names must be prefixed with a \'g\', see examples.'),
 560                         'redirects' => 'Automatically resolve redirects',
 561                         'indexpageids' => 'Include an additional pageids section listing all returned page IDs.',
 562                         'export' => 'Export the current revisions of all given or generated pages',
 563                         'exportnowrap' => 'Return the export XML without wrapping it in an XML result (same format as Special:Export). Can only be used with export',
 564                 );
 565         }
 566
 567         public function getDescription() {
 568                 return array (
 569                         'Query API module allows applications to get needed pieces of data from the MediaWiki databases,',
 570                         'and is loosely based on the old query.php interface.',
 571                         'All data modifications will first have to use query to acquire a token to prevent abuse from malicious sites.'
 572                 );
 573         }
 574
 575         protected function getExamples() {
 576                 return array (
 577                         'api.php?action=query&prop=revisions&meta=siteinfo&titles=Main%20Page&rvprop=user|comment',
 578                         'api.php?action=query&generator=allpages&gapprefix=API/&prop=revisions',
 579                 );
 580         }
 581
 582         public function getVersion() {
 583                 $psModule = new ApiPageSet($this);
 584                 $vers = array ();
 585                 $vers[] = __CLASS__ . ': $Id$';
 586                 $vers[] = $psModule->getVersion();
 587                 return $vers;
 588         }
 589 }