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