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                 //
 201                 // Create PageSet
 202                 //
 203                 $this->mPageSet = new ApiPageSet($this, $this->redirects);
 204
 205                 //
 206                 // Instantiate requested modules
 207                 //
 208                 $modules = array ();
 209                 $this->InstantiateModules($modules, 'prop', $this->mQueryPropModules);
 210                 $this->InstantiateModules($modules, 'list', $this->mQueryListModules);
 211                 $this->InstantiateModules($modules, 'meta', $this->mQueryMetaModules);
 212
 213                 //
 214                 // If given, execute generator to substitute user supplied data with generated data.
 215                 //
 216                 if (isset ($this->params['generator'])) {
 217                         $this->executeGeneratorModule($this->params['generator'], $modules);
 218                 } else {
 219                         // Append custom fields and populate page/revision information
 220                         $this->addCustomFldsToPageSet($modules, $this->mPageSet);
 221                         $this->mPageSet->execute();
 222                 }
 223
 224                 //
 225                 // Record page information (title, namespace, if exists, etc)
 226                 //
 227                 $this->outputGeneralPageInfo();
 228
 229                 //
 230                 // Execute all requested modules.
 231                 //
 232                 foreach ($modules as $module) {
 233                         $module->profileIn();
 234                         $module->execute();
 235                         wfRunHooks('APIQueryAfterExecute', array(&$module));
 236                         $module->profileOut();
 237                 }
 238         }
 239
 240         /**
 241          * Query modules may optimize data requests through the $this->getPageSet() object
 242          * by adding extra fields from the page table.
 243          * This function will gather all the extra request fields from the modules.
 244          * @param $modules array of module objects
 245          * @param $pageSet ApiPageSet
 246          */
 247         private function addCustomFldsToPageSet($modules, $pageSet) {
 248                 // Query all requested modules.
 249                 foreach ($modules as $module) {
 250                         $module->requestExtraData($pageSet);
 251                 }
 252         }
 253
 254         /**
 255          * Create instances of all modules requested by the client
 256          * @param $modules array to append instatiated modules to
 257          * @param $param string Parameter name to read modules from
 258          * @param $moduleList array(modulename => classname)
 259          */
 260         private function InstantiateModules(&$modules, $param, $moduleList) {
 261                 $list = @$this->params[$param];
 262                 if (!is_null ($list))
 263                         foreach ($list as $moduleName)
 264                                 $modules[] = new $moduleList[$moduleName] ($this, $moduleName);
 265         }
 266
 267         /**
 268          * Appends an element for each page in the current pageSet with the
 269          * most general information (id, title), plus any title normalizations
 270          * and missing or invalid title/pageids/revids.
 271          */
 272         private function outputGeneralPageInfo() {
 273
 274                 $pageSet = $this->getPageSet();
 275                 $result = $this->getResult();
 276
 277                 # We don't check for a full result set here because we can't be adding
 278                 # more than 380K. The maximum revision size is in the megabyte range,
 279                 # and the maximum result size must be even higher than that.
 280
 281                 // Title normalizations
 282                 $normValues = array ();
 283                 foreach ($pageSet->getNormalizedTitles() as $rawTitleStr => $titleStr) {
 284                         $normValues[] = array (
 285                                 'from' => $rawTitleStr,
 286                                 'to' => $titleStr
 287                         );
 288                 }
 289
 290                 if (count($normValues)) {
 291                         $result->setIndexedTagName($normValues, 'n');
 292                         $result->addValue('query', 'normalized', $normValues);
 293                 }
 294
 295                 // Interwiki titles
 296                 $intrwValues = array ();
 297                 foreach ($pageSet->getInterwikiTitles() as $rawTitleStr => $interwikiStr) {
 298                         $intrwValues[] = array (
 299                                 'title' => $rawTitleStr,
 300                                 'iw' => $interwikiStr
 301                         );
 302                 }
 303
 304                 if (count($intrwValues)) {
 305                         $result->setIndexedTagName($intrwValues, 'i');
 306                         $result->addValue('query', 'interwiki', $intrwValues);
 307                 }
 308
 309                 // Show redirect information
 310                 $redirValues = array ();
 311                 foreach ($pageSet->getRedirectTitles() as $titleStrFrom => $titleStrTo) {
 312                         $redirValues[] = array (
 313                                 'from' => strval($titleStrFrom),
 314                                 'to' => $titleStrTo
 315                         );
 316                 }
 317
 318                 if (count($redirValues)) {
 319                         $result->setIndexedTagName($redirValues, 'r');
 320                         $result->addValue('query', 'redirects', $redirValues);
 321                 }
 322
 323                 //
 324                 // Missing revision elements
 325                 //
 326                 $missingRevIDs = $pageSet->getMissingRevisionIDs();
 327                 if (count($missingRevIDs)) {
 328                         $revids = array ();
 329                         foreach ($missingRevIDs as $revid) {
 330                                 $revids[$revid] = array (
 331                                         'revid' => $revid
 332                                 );
 333                         }
 334                         $result->setIndexedTagName($revids, 'rev');
 335                         $result->addValue('query', 'badrevids', $revids);
 336                 }
 337
 338                 //
 339                 // Page elements
 340                 //
 341                 $pages = array ();
 342
 343                 // Report any missing titles
 344                 foreach ($pageSet->getMissingTitles() as $fakeId => $title) {
 345                         $vals = array();
 346                         ApiQueryBase :: addTitleInfo($vals, $title);
 347                         $vals['missing'] = '';
 348                         $pages[$fakeId] = $vals;
 349                 }
 350                 // Report any invalid titles
 351                 foreach ($pageSet->getInvalidTitles() as $fakeId => $title)
 352                         $pages[$fakeId] = array('title' => $title, 'invalid' => '');
 353                 // Report any missing page ids
 354                 foreach ($pageSet->getMissingPageIDs() as $pageid) {
 355                         $pages[$pageid] = array (
 356                                 'pageid' => $pageid,
 357                                 'missing' => ''
 358                         );
 359                 }
 360
 361                 // Output general page information for found titles
 362                 foreach ($pageSet->getGoodTitles() as $pageid => $title) {
 363                         $vals = array();
 364                         $vals['pageid'] = $pageid;
 365                         ApiQueryBase :: addTitleInfo($vals, $title);
 366                         $pages[$pageid] = $vals;
 367                 }
 368
 369                 if (count($pages)) {
 370
 371                         if ($this->params['indexpageids']) {
 372                                 $pageIDs = array_keys($pages);
 373                                 // json treats all map keys as strings - converting to match
 374                                 $pageIDs = array_map('strval', $pageIDs);
 375                                 $result->setIndexedTagName($pageIDs, 'id');
 376                                 $result->addValue('query', 'pageids', $pageIDs);
 377                         }
 378
 379                         $result->setIndexedTagName($pages, 'page');
 380                         $result->addValue('query', 'pages', $pages);
 381                 }
 382                 if ($this->params['export']) {
 383                         $exporter = new WikiExporter($this->getDB());
 384                         // WikiExporter writes to stdout, so catch its
 385                         // output with an ob
 386                         ob_start();
 387                         $exporter->openStream();
 388                         foreach (@$pageSet->getGoodTitles() as $title)
 389                                 if ($title->userCanRead())
 390                                         $exporter->pageByTitle($title);
 391                         $exporter->closeStream();
 392                         $exportxml = ob_get_contents();
 393                         ob_end_clean();
 394                         // Don't check the size of exported stuff
 395                         // It's not continuable, so it would cause more
 396                         // problems than it'd solve
 397                         $result->disableSizeCheck();
 398                         if ($this->params['exportnowrap']) {
 399                                 $result->reset();
 400                                 // Raw formatter will handle this
 401                                 $result->addValue(null, 'text', $exportxml);
 402                                 $result->addValue(null, 'mime', 'text/xml');
 403                         } else {
 404                                 $r = array();
 405                                 ApiResult::setContent($r, $exportxml);
 406                                 $result->addValue('query', 'export', $r);
 407                         }
 408                         $result->enableSizeCheck();
 409                 }
 410         }
 411
 412         /**
 413          * For generator mode, execute generator, and use its output as new
 414          * ApiPageSet
 415          * @param $generatorName string Module name
 416          * @param $modules array of module objects
 417          */
 418         protected function executeGeneratorModule($generatorName, $modules) {
 419
 420                 // Find class that implements requested generator
 421                 if (isset ($this->mQueryListModules[$generatorName])) {
 422                         $className = $this->mQueryListModules[$generatorName];
 423                 } elseif (isset ($this->mQueryPropModules[$generatorName])) {
 424                         $className = $this->mQueryPropModules[$generatorName];
 425                 } else {
 426                         ApiBase :: dieDebug(__METHOD__, "Unknown generator=$generatorName");
 427                 }
 428
 429                 // Generator results
 430                 $resultPageSet = new ApiPageSet($this, $this->redirects);
 431
 432                 // Create and execute the generator
 433                 $generator = new $className ($this, $generatorName);
 434                 if (!$generator instanceof ApiQueryGeneratorBase)
 435                         $this->dieUsage("Module $generatorName cannot be used as a generator", "badgenerator");
 436
 437                 $generator->setGeneratorMode();
 438
 439                 // Add any additional fields modules may need
 440                 $generator->requestExtraData($this->mPageSet);
 441                 $this->addCustomFldsToPageSet($modules, $resultPageSet);
 442
 443                 // Populate page information with the original user input
 444                 $this->mPageSet->execute();
 445
 446                 // populate resultPageSet with the generator output
 447                 $generator->profileIn();
 448                 $generator->executeGenerator($resultPageSet);
 449                 wfRunHooks('APIQueryGeneratorAfterExecute', array(&$generator, &$resultPageSet));
 450                 $resultPageSet->finishPageSetGeneration();
 451                 $generator->profileOut();
 452
 453                 // Swap the resulting pageset back in
 454                 $this->mPageSet = $resultPageSet;
 455         }
 456
 457         public function getAllowedParams() {
 458                 return array (
 459                         'prop' => array (
 460                                 ApiBase :: PARAM_ISMULTI => true,
 461                                 ApiBase :: PARAM_TYPE => $this->mPropModuleNames
 462                         ),
 463                         'list' => array (
 464                                 ApiBase :: PARAM_ISMULTI => true,
 465                                 ApiBase :: PARAM_TYPE => $this->mListModuleNames
 466                         ),
 467                         'meta' => array (
 468                                 ApiBase :: PARAM_ISMULTI => true,
 469                                 ApiBase :: PARAM_TYPE => $this->mMetaModuleNames
 470                         ),
 471                         'generator' => array (
 472                                 ApiBase :: PARAM_TYPE => $this->mAllowedGenerators
 473                         ),
 474                         'redirects' => false,
 475                         'indexpageids' => false,
 476                         'export' => false,
 477                         'exportnowrap' => false,
 478                 );
 479         }
 480
 481         /**
 482          * Override the parent to generate help messages for all available query modules.
 483          * @return string
 484          */
 485         public function makeHelpMsg() {
 486
 487                 $msg = '';
 488
 489                 // Make sure the internal object is empty
 490                 // (just in case a sub-module decides to optimize during instantiation)
 491                 $this->mPageSet = null;
 492                 $this->mAllowedGenerators = array();    // Will be repopulated
 493
 494                 $astriks = str_repeat('--- ', 8);
 495                 $astriks2 = str_repeat('*** ', 10);
 496                 $msg .= "\n$astriks Query: Prop  $astriks\n\n";
 497                 $msg .= $this->makeHelpMsgHelper($this->mQueryPropModules, 'prop');
 498                 $msg .= "\n$astriks Query: List  $astriks\n\n";
 499                 $msg .= $this->makeHelpMsgHelper($this->mQueryListModules, 'list');
 500                 $msg .= "\n$astriks Query: Meta  $astriks\n\n";
 501                 $msg .= $this->makeHelpMsgHelper($this->mQueryMetaModules, 'meta');
 502                 $msg .= "\n\n$astriks2 Modules: continuation  $astriks2\n\n";
 503
 504                 // Perform the base call last because the $this->mAllowedGenerators
 505                 // will be updated inside makeHelpMsgHelper()
 506                 // Use parent to make default message for the query module
 507                 $msg = parent :: makeHelpMsg() . $msg;
 508
 509                 return $msg;
 510         }
 511
 512         /**
 513          * For all modules in $moduleList, generate help messages and join them together
 514          * @param $moduleList array(modulename => classname)
 515          * @param $paramName string Parameter name
 516          * @return string
 517          */
 518         private function makeHelpMsgHelper($moduleList, $paramName) {
 519
 520                 $moduleDescriptions = array ();
 521
 522                 foreach ($moduleList as $moduleName => $moduleClass) {
 523                         $module = new $moduleClass ($this, $moduleName, null);
 524
 525                         $msg = ApiMain::makeHelpMsgHeader($module, $paramName);
 526                         $msg2 = $module->makeHelpMsg();
 527                         if ($msg2 !== false)
 528                                 $msg .= $msg2;
 529                         if ($module instanceof ApiQueryGeneratorBase) {
 530                                 $this->mAllowedGenerators[] = $moduleName;
 531                                 $msg .= "Generator:\n  This module may be used as a generator\n";
 532                         }
 533                         $moduleDescriptions[] = $msg;
 534                 }
 535
 536                 return implode("\n", $moduleDescriptions);
 537         }
 538
 539         /**
 540          * Override to add extra parameters from PageSet
 541          * @return string
 542          */
 543         public function makeHelpMsgParameters() {
 544                 $psModule = new ApiPageSet($this);
 545                 return $psModule->makeHelpMsgParameters() . parent :: makeHelpMsgParameters();
 546         }
 547
 548         public function shouldCheckMaxlag() {
 549                 return true;
 550         }
 551
 552         public function getParamDescription() {
 553                 return array (
 554                         'prop' => 'Which properties to get for the titles/revisions/pageids',
 555                         'list' => 'Which lists to get',
 556                         'meta' => 'Which meta data to get about the site',
 557                         'generator' => array('Use the output of a list as the input for other prop/list/meta items',
 558                                         'NOTE: generator parameter names must be prefixed with a \'g\', see examples.'),
 559                         'redirects' => 'Automatically resolve redirects',
 560                         'indexpageids' => 'Include an additional pageids section listing all returned page IDs.',
 561                         'export' => 'Export the current revisions of all given or generated pages',
 562                         'exportnowrap' => 'Return the export XML without wrapping it in an XML result (same format as Special:Export). Can only be used with export',
 563                 );
 564         }
 565
 566         public function getDescription() {
 567                 return array (
 568                         'Query API module allows applications to get needed pieces of data from the MediaWiki databases,',
 569                         'and is loosely based on the old query.php interface.',
 570                         'All data modifications will first have to use query to acquire a token to prevent abuse from malicious sites.'
 571                 );
 572         }
 573
 574         protected function getExamples() {
 575                 return array (
 576                         'api.php?action=query&prop=revisions&meta=siteinfo&titles=Main%20Page&rvprop=user|comment',
 577                         'api.php?action=query&generator=allpages&gapprefix=API/&prop=revisions',
 578                 );
 579         }
 580
 581         public function getVersion() {
 582                 $psModule = new ApiPageSet($this);
 583                 $vers = array ();
 584                 $vers[] = __CLASS__ . ': $Id$';
 585                 $vers[] = $psModule->getVersion();
 586                 return $vers;
 587         }
 588 }