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