[Summary view] [Print] [Text view]

   1  <?php
   2  /**
   3   *
   4   *
   5   * Created on Oct 16, 2006
   6   *
   7   * Copyright © 2006 Yuri Astrakhan "<Firstname><Lastname>@gmail.com"
   8   *
   9   * This program is free software; you can redistribute it and/or modify
  10   * it under the terms of the GNU General Public License as published by
  11   * the Free Software Foundation; either version 2 of the License, or
  12   * (at your option) any later version.
  13   *
  14   * This program is distributed in the hope that it will be useful,
  15   * but WITHOUT ANY WARRANTY; without even the implied warranty of
  16   * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
  17   * GNU General Public License for more details.
  18   *
  19   * You should have received a copy of the GNU General Public License along
  20   * with this program; if not, write to the Free Software Foundation, Inc.,
  21   * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
  22   * http://www.gnu.org/copyleft/gpl.html
  23   *
  24   * @file
  25   */
  26  
  27  /**
  28   * This is a three-in-one module to query:
  29   *   * backlinks  - links pointing to the given page,
  30   *   * embeddedin - what pages transclude the given page within themselves,
  31   *   * imageusage - what pages use the given image
  32   *
  33   * @ingroup API
  34   */
  35  class ApiQueryBacklinks extends ApiQueryGeneratorBase {
  36  
  37      /**
  38       * @var Title
  39       */
  40      private $rootTitle;
  41  
  42      private $params, $contID, $redirID, $redirect;
  43      private $bl_ns, $bl_from, $bl_table, $bl_code, $bl_title, $bl_fields, $hasNS;
  44  
  45      /**
  46       * Maps ns and title to pageid
  47       *
  48       * @var array
  49       */
  50      private $pageMap = array();
  51      private $resultArr;
  52  
  53      private $redirTitles = array();
  54      private $continueStr = null;
  55  
  56      // output element name, database column field prefix, database table
  57      private $backlinksSettings = array(
  58          'backlinks' => array(
  59              'code' => 'bl',
  60              'prefix' => 'pl',
  61              'linktbl' => 'pagelinks',
  62              'helpurl' => 'https://www.mediawiki.org/wiki/API:Backlinks',
  63          ),
  64          'embeddedin' => array(
  65              'code' => 'ei',
  66              'prefix' => 'tl',
  67              'linktbl' => 'templatelinks',
  68              'helpurl' => 'https://www.mediawiki.org/wiki/API:Embeddedin',
  69          ),
  70          'imageusage' => array(
  71              'code' => 'iu',
  72              'prefix' => 'il',
  73              'linktbl' => 'imagelinks',
  74              'helpurl' => 'https://www.mediawiki.org/wiki/API:Imageusage',
  75          )
  76      );
  77  
  78  	public function __construct( ApiQuery $query, $moduleName ) {
  79          $settings = $this->backlinksSettings[$moduleName];
  80          $prefix = $settings['prefix'];
  81          $code = $settings['code'];
  82          $this->resultArr = array();
  83  
  84          parent::__construct( $query, $moduleName, $code );
  85          $this->bl_ns = $prefix . '_namespace';
  86          $this->bl_from = $prefix . '_from';
  87          $this->bl_table = $settings['linktbl'];
  88          $this->bl_code = $code;
  89          $this->helpUrl = $settings['helpurl'];
  90  
  91          $this->hasNS = $moduleName !== 'imageusage';
  92          if ( $this->hasNS ) {
  93              $this->bl_title = $prefix . '_title';
  94              $this->bl_fields = array(
  95                  $this->bl_ns,
  96                  $this->bl_title
  97              );
  98          } else {
  99              $this->bl_title = $prefix . '_to';
 100              $this->bl_fields = array(
 101                  $this->bl_title
 102              );
 103          }
 104      }
 105  
 106  	public function execute() {
 107          $this->run();
 108      }
 109  
 110  	public function getCacheMode( $params ) {
 111          return 'public';
 112      }
 113  
 114  	public function executeGenerator( $resultPageSet ) {
 115          $this->run( $resultPageSet );
 116      }
 117  
 118      /**
 119       * @param ApiPageSet $resultPageSet
 120       * @return void
 121       */
 122  	private function prepareFirstQuery( $resultPageSet = null ) {
 123          /* SELECT page_id, page_title, page_namespace, page_is_redirect
 124           * FROM pagelinks, page WHERE pl_from=page_id
 125           * AND pl_title='Foo' AND pl_namespace=0
 126           * LIMIT 11 ORDER BY pl_from
 127           */
 128          $this->addTables( array( $this->bl_table, 'page' ) );
 129          $this->addWhere( "{$this->bl_from}=page_id" );
 130          if ( is_null( $resultPageSet ) ) {
 131              $this->addFields( array( 'page_id', 'page_title', 'page_namespace' ) );
 132          } else {
 133              $this->addFields( $resultPageSet->getPageTableFields() );
 134          }
 135  
 136          $this->addFields( 'page_is_redirect' );
 137          $this->addWhereFld( $this->bl_title, $this->rootTitle->getDBkey() );
 138  
 139          if ( $this->hasNS ) {
 140              $this->addWhereFld( $this->bl_ns, $this->rootTitle->getNamespace() );
 141          }
 142          $this->addWhereFld( 'page_namespace', $this->params['namespace'] );
 143  
 144          if ( !is_null( $this->contID ) ) {
 145              $op = $this->params['dir'] == 'descending' ? '<' : '>';
 146              $this->addWhere( "{$this->bl_from}$op={$this->contID}" );
 147          }
 148  
 149          if ( $this->params['filterredir'] == 'redirects' ) {
 150              $this->addWhereFld( 'page_is_redirect', 1 );
 151          } elseif ( $this->params['filterredir'] == 'nonredirects' && !$this->redirect ) {
 152              // bug 22245 - Check for !redirect, as filtering nonredirects, when
 153              // getting what links to them is contradictory
 154              $this->addWhereFld( 'page_is_redirect', 0 );
 155          }
 156  
 157          $this->addOption( 'LIMIT', $this->params['limit'] + 1 );
 158          $sort = ( $this->params['dir'] == 'descending' ? ' DESC' : '' );
 159          $this->addOption( 'ORDER BY', $this->bl_from . $sort );
 160          $this->addOption( 'STRAIGHT_JOIN' );
 161      }
 162  
 163      /**
 164       * @param ApiPageSet $resultPageSet
 165       * @return void
 166       */
 167  	private function prepareSecondQuery( $resultPageSet = null ) {
 168          /* SELECT page_id, page_title, page_namespace, page_is_redirect, pl_title, pl_namespace
 169             FROM pagelinks, page WHERE pl_from=page_id
 170             AND (pl_title='Foo' AND pl_namespace=0) OR (pl_title='Bar' AND pl_namespace=1)
 171             ORDER BY pl_namespace, pl_title, pl_from LIMIT 11
 172           */
 173          $db = $this->getDB();
 174          $this->addTables( array( 'page', $this->bl_table ) );
 175          $this->addWhere( "{$this->bl_from}=page_id" );
 176  
 177          if ( is_null( $resultPageSet ) ) {
 178              $this->addFields( array( 'page_id', 'page_title', 'page_namespace', 'page_is_redirect' ) );
 179          } else {
 180              $this->addFields( $resultPageSet->getPageTableFields() );
 181          }
 182  
 183          $this->addFields( $this->bl_title );
 184          if ( $this->hasNS ) {
 185              $this->addFields( $this->bl_ns );
 186          }
 187  
 188          // We can't use LinkBatch here because $this->hasNS may be false
 189          $titleWhere = array();
 190          $allRedirNs = array();
 191          $allRedirDBkey = array();
 192          /** @var $t Title */
 193          foreach ( $this->redirTitles as $t ) {
 194              $redirNs = $t->getNamespace();
 195              $redirDBkey = $t->getDBkey();
 196              $titleWhere[] = "{$this->bl_title} = " . $db->addQuotes( $redirDBkey ) .
 197                  ( $this->hasNS ? " AND {$this->bl_ns} = {$redirNs}" : '' );
 198              $allRedirNs[] = $redirNs;
 199              $allRedirDBkey[] = $redirDBkey;
 200          }
 201          $this->addWhere( $db->makeList( $titleWhere, LIST_OR ) );
 202          $this->addWhereFld( 'page_namespace', $this->params['namespace'] );
 203  
 204          if ( !is_null( $this->redirID ) ) {
 205              $op = $this->params['dir'] == 'descending' ? '<' : '>';
 206              /** @var $first Title */
 207              $first = $this->redirTitles[0];
 208              $title = $db->addQuotes( $first->getDBkey() );
 209              $ns = $first->getNamespace();
 210              $from = $this->redirID;
 211              if ( $this->hasNS ) {
 212                  $this->addWhere( "{$this->bl_ns} $op $ns OR " .
 213                      "({$this->bl_ns} = $ns AND " .
 214                      "({$this->bl_title} $op $title OR " .
 215                      "({$this->bl_title} = $title AND " .
 216                      "{$this->bl_from} $op= $from)))" );
 217              } else {
 218                  $this->addWhere( "{$this->bl_title} $op $title OR " .
 219                      "({$this->bl_title} = $title AND " .
 220                      "{$this->bl_from} $op= $from)" );
 221              }
 222          }
 223          if ( $this->params['filterredir'] == 'redirects' ) {
 224              $this->addWhereFld( 'page_is_redirect', 1 );
 225          } elseif ( $this->params['filterredir'] == 'nonredirects' ) {
 226              $this->addWhereFld( 'page_is_redirect', 0 );
 227          }
 228  
 229          $this->addOption( 'LIMIT', $this->params['limit'] + 1 );
 230          $orderBy = array();
 231          $sort = ( $this->params['dir'] == 'descending' ? ' DESC' : '' );
 232          // Don't order by namespace/title if it's constant in the WHERE clause
 233          if ( $this->hasNS && count( array_unique( $allRedirNs ) ) != 1 ) {
 234              $orderBy[] = $this->bl_ns . $sort;
 235          }
 236          if ( count( array_unique( $allRedirDBkey ) ) != 1 ) {
 237              $orderBy[] = $this->bl_title . $sort;
 238          }
 239          $orderBy[] = $this->bl_from . $sort;
 240          $this->addOption( 'ORDER BY', $orderBy );
 241          $this->addOption( 'USE INDEX', array( 'page' => 'PRIMARY' ) );
 242      }
 243  
 244      /**
 245       * @param ApiPageSet $resultPageSet
 246       * @return void
 247       */
 248  	private function run( $resultPageSet = null ) {
 249          $this->params = $this->extractRequestParams( false );
 250          $this->redirect = isset( $this->params['redirect'] ) && $this->params['redirect'];
 251          $userMax = ( $this->redirect ? ApiBase::LIMIT_BIG1 / 2 : ApiBase::LIMIT_BIG1 );
 252          $botMax = ( $this->redirect ? ApiBase::LIMIT_BIG2 / 2 : ApiBase::LIMIT_BIG2 );
 253  
 254          $result = $this->getResult();
 255  
 256          if ( $this->params['limit'] == 'max' ) {
 257              $this->params['limit'] = $this->getMain()->canApiHighLimits() ? $botMax : $userMax;
 258              $result->setParsedLimit( $this->getModuleName(), $this->params['limit'] );
 259          } else {
 260              $this->params['limit'] = intval( $this->params['limit'] );
 261              $this->validateLimit( 'limit', $this->params['limit'], 1, $userMax, $botMax );
 262          }
 263  
 264          $this->processContinue();
 265          $this->prepareFirstQuery( $resultPageSet );
 266  
 267          $res = $this->select( __METHOD__ . '::firstQuery' );
 268  
 269          $count = 0;
 270  
 271          foreach ( $res as $row ) {
 272              if ( ++$count > $this->params['limit'] ) {
 273                  // We've reached the one extra which shows that there are
 274                  // additional pages to be had. Stop here...
 275                  // Continue string preserved in case the redirect query doesn't pass the limit
 276                  $this->continueStr = $this->getContinueStr( $row->page_id );
 277                  break;
 278              }
 279  
 280              if ( is_null( $resultPageSet ) ) {
 281                  $this->extractRowInfo( $row );
 282              } else {
 283                  $this->pageMap[$row->page_namespace][$row->page_title] = $row->page_id;
 284                  if ( $row->page_is_redirect ) {
 285                      $this->redirTitles[] = Title::makeTitle( $row->page_namespace, $row->page_title );
 286                  }
 287  
 288                  $resultPageSet->processDbRow( $row );
 289              }
 290          }
 291  
 292          if ( $this->redirect && count( $this->redirTitles ) ) {
 293              $this->resetQueryParams();
 294              $this->prepareSecondQuery( $resultPageSet );
 295              $res = $this->select( __METHOD__ . '::secondQuery' );
 296              $count = 0;
 297              foreach ( $res as $row ) {
 298                  if ( ++$count > $this->params['limit'] ) {
 299                      // We've reached the one extra which shows that there are
 300                      // additional pages to be had. Stop here...
 301                      // We need to keep the parent page of this redir in
 302                      if ( $this->hasNS ) {
 303                          $parentID = $this->pageMap[$row->{$this->bl_ns}][$row->{$this->bl_title}];
 304                      } else {
 305                          $parentID = $this->pageMap[NS_FILE][$row->{$this->bl_title}];
 306                      }
 307                      $this->continueStr = $this->getContinueRedirStr( $parentID, $row->page_id );
 308                      break;
 309                  }
 310  
 311                  if ( is_null( $resultPageSet ) ) {
 312                      $this->extractRedirRowInfo( $row );
 313                  } else {
 314                      $resultPageSet->processDbRow( $row );
 315                  }
 316              }
 317          }
 318          if ( is_null( $resultPageSet ) ) {
 319              // Try to add the result data in one go and pray that it fits
 320              $fit = $result->addValue( 'query', $this->getModuleName(), array_values( $this->resultArr ) );
 321              if ( !$fit ) {
 322                  // It didn't fit. Add elements one by one until the
 323                  // result is full.
 324                  foreach ( $this->resultArr as $pageID => $arr ) {
 325                      // Add the basic entry without redirlinks first
 326                      $fit = $result->addValue(
 327                          array( 'query', $this->getModuleName() ),
 328                          null, array_diff_key( $arr, array( 'redirlinks' => '' ) ) );
 329                      if ( !$fit ) {
 330                          $this->continueStr = $this->getContinueStr( $pageID );
 331                          break;
 332                      }
 333  
 334                      $hasRedirs = false;
 335                      $redirLinks = isset( $arr['redirlinks'] ) ? $arr['redirlinks'] : array();
 336                      foreach ( (array)$redirLinks as $key => $redir ) {
 337                          $fit = $result->addValue(
 338                              array( 'query', $this->getModuleName(), $pageID, 'redirlinks' ),
 339                              $key, $redir );
 340                          if ( !$fit ) {
 341                              $this->continueStr = $this->getContinueRedirStr( $pageID, $redir['pageid'] );
 342                              break;
 343                          }
 344                          $hasRedirs = true;
 345                      }
 346                      if ( $hasRedirs ) {
 347                          $result->setIndexedTagName_internal(
 348                              array( 'query', $this->getModuleName(), $pageID, 'redirlinks' ),
 349                              $this->bl_code );
 350                      }
 351                      if ( !$fit ) {
 352                          break;
 353                      }
 354                  }
 355              }
 356  
 357              $result->setIndexedTagName_internal(
 358                  array( 'query', $this->getModuleName() ),
 359                  $this->bl_code
 360              );
 361          }
 362          if ( !is_null( $this->continueStr ) ) {
 363              $this->setContinueEnumParameter( 'continue', $this->continueStr );
 364          }
 365      }
 366  
 367  	private function extractRowInfo( $row ) {
 368          $this->pageMap[$row->page_namespace][$row->page_title] = $row->page_id;
 369          $t = Title::makeTitle( $row->page_namespace, $row->page_title );
 370          $a = array( 'pageid' => intval( $row->page_id ) );
 371          ApiQueryBase::addTitleInfo( $a, $t );
 372          if ( $row->page_is_redirect ) {
 373              $a['redirect'] = '';
 374              $this->redirTitles[] = $t;
 375          }
 376          // Put all the results in an array first
 377          $this->resultArr[$a['pageid']] = $a;
 378      }
 379  
 380  	private function extractRedirRowInfo( $row ) {
 381          $a['pageid'] = intval( $row->page_id );
 382          ApiQueryBase::addTitleInfo( $a, Title::makeTitle( $row->page_namespace, $row->page_title ) );
 383          if ( $row->page_is_redirect ) {
 384              $a['redirect'] = '';
 385          }
 386          $ns = $this->hasNS ? $row->{$this->bl_ns} : NS_FILE;
 387          $parentID = $this->pageMap[$ns][$row->{$this->bl_title}];
 388          // Put all the results in an array first
 389          $this->resultArr[$parentID]['redirlinks'][] = $a;
 390          $this->getResult()->setIndexedTagName(
 391              $this->resultArr[$parentID]['redirlinks'],
 392              $this->bl_code
 393          );
 394      }
 395  
 396  	protected function processContinue() {
 397          if ( !is_null( $this->params['continue'] ) ) {
 398              $this->parseContinueParam();
 399          } else {
 400              $this->rootTitle = $this->getTitleOrPageId( $this->params )->getTitle();
 401          }
 402  
 403          // only image titles are allowed for the root in imageinfo mode
 404          if ( !$this->hasNS && $this->rootTitle->getNamespace() !== NS_FILE ) {
 405              $this->dieUsage(
 406                  "The title for {$this->getModuleName()} query must be an image",
 407                  'bad_image_title'
 408              );
 409          }
 410      }
 411  
 412  	protected function parseContinueParam() {
 413          $continueList = explode( '|', $this->params['continue'] );
 414          // expected format:
 415          // ns | key | id1 [| id2]
 416          // ns+key: root title
 417          // id1: first-level page ID to continue from
 418          // id2: second-level page ID to continue from
 419  
 420          // null stuff out now so we know what's set and what isn't
 421          $this->rootTitle = $this->contID = $this->redirID = null;
 422          $rootNs = intval( $continueList[0] );
 423          $this->dieContinueUsageIf( $rootNs === 0 && $continueList[0] !== '0' );
 424  
 425          $this->rootTitle = Title::makeTitleSafe( $rootNs, $continueList[1] );
 426          $this->dieContinueUsageIf( !$this->rootTitle );
 427  
 428          $contID = intval( $continueList[2] );
 429          $this->dieContinueUsageIf( $contID === 0 && $continueList[2] !== '0' );
 430  
 431          $this->contID = $contID;
 432          $id2 = isset( $continueList[3] ) ? $continueList[3] : null;
 433          $redirID = intval( $id2 );
 434  
 435          if ( $redirID === 0 && $id2 !== '0' ) {
 436              // This one isn't required
 437              return;
 438          }
 439          $this->redirID = $redirID;
 440      }
 441  
 442  	protected function getContinueStr( $lastPageID ) {
 443          return $this->rootTitle->getNamespace() .
 444          '|' . $this->rootTitle->getDBkey() .
 445          '|' . $lastPageID;
 446      }
 447  
 448  	protected function getContinueRedirStr( $lastPageID, $lastRedirID ) {
 449          return $this->getContinueStr( $lastPageID ) . '|' . $lastRedirID;
 450      }
 451  
 452  	public function getAllowedParams() {
 453          $retval = array(
 454              'title' => array(
 455                  ApiBase::PARAM_TYPE => 'string',
 456              ),
 457              'pageid' => array(
 458                  ApiBase::PARAM_TYPE => 'integer',
 459              ),
 460              'continue' => null,
 461              'namespace' => array(
 462                  ApiBase::PARAM_ISMULTI => true,
 463                  ApiBase::PARAM_TYPE => 'namespace'
 464              ),
 465              'dir' => array(
 466                  ApiBase::PARAM_DFLT => 'ascending',
 467                  ApiBase::PARAM_TYPE => array(
 468                      'ascending',
 469                      'descending'
 470                  )
 471              ),
 472              'filterredir' => array(
 473                  ApiBase::PARAM_DFLT => 'all',
 474                  ApiBase::PARAM_TYPE => array(
 475                      'all',
 476                      'redirects',
 477                      'nonredirects'
 478                  )
 479              ),
 480              'limit' => array(
 481                  ApiBase::PARAM_DFLT => 10,
 482                  ApiBase::PARAM_TYPE => 'limit',
 483                  ApiBase::PARAM_MIN => 1,
 484                  ApiBase::PARAM_MAX => ApiBase::LIMIT_BIG1,
 485                  ApiBase::PARAM_MAX2 => ApiBase::LIMIT_BIG2
 486              )
 487          );
 488          if ( $this->getModuleName() == 'embeddedin' ) {
 489              return $retval;
 490          }
 491          $retval['redirect'] = false;
 492  
 493          return $retval;
 494      }
 495  
 496  	public function getParamDescription() {
 497          $retval = array(
 498              'title' => "Title to search. Cannot be used together with {$this->bl_code}pageid",
 499              'pageid' => "Pageid to search. Cannot be used together with {$this->bl_code}title",
 500              'continue' => 'When more results are available, use this to continue',
 501              'namespace' => 'The namespace to enumerate',
 502              'dir' => 'The direction in which to list',
 503          );
 504          if ( $this->getModuleName() != 'embeddedin' ) {
 505              return array_merge( $retval, array(
 506                  'redirect' => 'If linking page is a redirect, find all pages ' .
 507                      'that link to that redirect as well. Maximum limit is halved.',
 508                  'filterredir' => 'How to filter for redirects. If set to ' .
 509                      "nonredirects when {$this->bl_code}redirect is enabled, " .
 510                      'this is only applied to the second level',
 511                  'limit' => 'How many total pages to return. If ' .
 512                      "{$this->bl_code}redirect is enabled, limit applies to each " .
 513                      'level separately (which means you may get up to 2 * limit results).'
 514              ) );
 515          }
 516  
 517          return array_merge( $retval, array(
 518              'filterredir' => 'How to filter for redirects',
 519              'limit' => 'How many total pages to return'
 520          ) );
 521      }
 522  
 523  	public function getDescription() {
 524          switch ( $this->getModuleName() ) {
 525              case 'backlinks':
 526                  return 'Find all pages that link to the given page.';
 527              case 'embeddedin':
 528                  return 'Find all pages that embed (transclude) the given title.';
 529              case 'imageusage':
 530                  return 'Find all pages that use the given image title.';
 531              default:
 532                  ApiBase::dieDebug( __METHOD__, 'Unknown module name.' );
 533          }
 534      }
 535  
 536  	public function getExamples() {
 537          static $examples = array(
 538              'backlinks' => array(
 539                  'api.php?action=query&list=backlinks&bltitle=Main%20Page',
 540                  'api.php?action=query&generator=backlinks&gbltitle=Main%20Page&prop=info'
 541              ),
 542              'embeddedin' => array(
 543                  'api.php?action=query&list=embeddedin&eititle=Template:Stub',
 544                  'api.php?action=query&generator=embeddedin&geititle=Template:Stub&prop=info'
 545              ),
 546              'imageusage' => array(
 547                  'api.php?action=query&list=imageusage&iutitle=File:Albert%20Einstein%20Head.jpg',
 548                  'api.php?action=query&generator=imageusage&giutitle=File:Albert%20Einstein%20Head.jpg&prop=info'
 549              )
 550          );
 551  
 552          return $examples[$this->getModuleName()];
 553      }
 554  
 555  	public function getHelpUrls() {
 556          return $this->helpUrl;
 557      }
 558  }