includes/api/ApiQueryBacklinks.php

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