includes/api/ApiQueryBacklinks.php

   1 <?php
   2 /**
   3  * API for MediaWiki 1.8+
   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         private $params, $rootTitle, $contID, $redirID, $redirect;
  43         private $bl_ns, $bl_from, $bl_table, $bl_code, $bl_title, $bl_sort, $bl_fields, $hasNS;
  44         private $pageMap, $resultArr;
  45
  46         // output element name, database column field prefix, database table
  47         private $backlinksSettings = array(
  48                 'backlinks' => array(
  49                         'code' => 'bl',
  50                         'prefix' => 'pl',
  51                         'linktbl' => 'pagelinks'
  52                 ),
  53                 'embeddedin' => array(
  54                         'code' => 'ei',
  55                         'prefix' => 'tl',
  56                         'linktbl' => 'templatelinks'
  57                 ),
  58                 'imageusage' => array(
  59                         'code' => 'iu',
  60                         'prefix' => 'il',
  61                         'linktbl' => 'imagelinks'
  62                 )
  63         );
  64
  65         public function __construct( $query, $moduleName ) {
  66                 extract( $this->backlinksSettings[$moduleName] );
  67                 $this->resultArr = array();
  68
  69                 parent::__construct( $query, $moduleName, $code );
  70                 $this->bl_ns = $prefix . '_namespace';
  71                 $this->bl_from = $prefix . '_from';
  72                 $this->bl_table = $linktbl;
  73                 $this->bl_code = $code;
  74
  75                 $this->hasNS = $moduleName !== 'imageusage';
  76                 if ( $this->hasNS ) {
  77                         $this->bl_title = $prefix . '_title';
  78                         $this->bl_sort = "{$this->bl_ns}, {$this->bl_title}, {$this->bl_from}";
  79                         $this->bl_fields = array(
  80                                 $this->bl_ns,
  81                                 $this->bl_title
  82                         );
  83                 } else {
  84                         $this->bl_title = $prefix . '_to';
  85                         $this->bl_sort = "{$this->bl_title}, {$this->bl_from}";
  86                         $this->bl_fields = array(
  87                                 $this->bl_title
  88                         );
  89                 }
  90         }
  91
  92         public function execute() {
  93                 $this->run();
  94         }
  95
  96         public function getCacheMode( $params ) {
  97                 return 'public';
  98         }
  99
 100         public function executeGenerator( $resultPageSet ) {
 101                 $this->run( $resultPageSet );
 102         }
 103
 104         private function prepareFirstQuery( $resultPageSet = null ) {
 105                 /* SELECT page_id, page_title, page_namespace, page_is_redirect
 106                  * FROM pagelinks, page WHERE pl_from=page_id
 107                  * AND pl_title='Foo' AND pl_namespace=0
 108                  * LIMIT 11 ORDER BY pl_from
 109                  */
 110                 $this->addTables( array( $this->bl_table, 'page' ) );
 111                 $this->addWhere( "{$this->bl_from}=page_id" );
 112                 if ( is_null( $resultPageSet ) ) {
 113                         $this->addFields( array( 'page_id', 'page_title', 'page_namespace' ) );
 114                 } else {
 115                         $this->addFields( $resultPageSet->getPageTableFields() );
 116                 }
 117
 118                 $this->addFields( 'page_is_redirect' );
 119                 $this->addWhereFld( $this->bl_title, $this->rootTitle->getDBkey() );
 120
 121                 if ( $this->hasNS ) {
 122                         $this->addWhereFld( $this->bl_ns, $this->rootTitle->getNamespace() );
 123                 }
 124                 $this->addWhereFld( 'page_namespace', $this->params['namespace'] );
 125
 126                 if ( !is_null( $this->contID ) ) {
 127                         $this->addWhere( "{$this->bl_from}>={$this->contID}" );
 128                 }
 129
 130                 if ( $this->params['filterredir'] == 'redirects' ) {
 131                         $this->addWhereFld( 'page_is_redirect', 1 );
 132                 } elseif ( $this->params['filterredir'] == 'nonredirects' && !$this->redirect ) {
 133                         // bug 22245 - Check for !redirect, as filtering nonredirects, when getting what links to them is contradictory
 134                         $this->addWhereFld( 'page_is_redirect', 0 );
 135                 }
 136
 137                 $this->addOption( 'LIMIT', $this->params['limit'] + 1 );
 138                 $this->addOption( 'ORDER BY', $this->bl_from );
 139                 $this->addOption( 'STRAIGHT_JOIN' );
 140         }
 141
 142         private function prepareSecondQuery( $resultPageSet = null ) {
 143                 /* SELECT page_id, page_title, page_namespace, page_is_redirect, pl_title, pl_namespace
 144                    FROM pagelinks, page WHERE pl_from=page_id
 145                    AND (pl_title='Foo' AND pl_namespace=0) OR (pl_title='Bar' AND pl_namespace=1)
 146                    ORDER BY pl_namespace, pl_title, pl_from LIMIT 11
 147                  */
 148                 $db = $this->getDB();
 149                 $this->addTables( array( 'page', $this->bl_table ) );
 150                 $this->addWhere( "{$this->bl_from}=page_id" );
 151
 152                 if ( is_null( $resultPageSet ) ) {
 153                         $this->addFields( array( 'page_id', 'page_title', 'page_namespace', 'page_is_redirect' ) );
 154                 } else {
 155                         $this->addFields( $resultPageSet->getPageTableFields() );
 156                 }
 157
 158                 $this->addFields( $this->bl_title );
 159                 if ( $this->hasNS ) {
 160                         $this->addFields( $this->bl_ns );
 161                 }
 162
 163                 // We can't use LinkBatch here because $this->hasNS may be false
 164                 $titleWhere = array();
 165                 foreach ( $this->redirTitles as $t ) {
 166                         $titleWhere[] = "{$this->bl_title} = " . $db->addQuotes( $t->getDBkey() ) .
 167                                         ( $this->hasNS ? " AND {$this->bl_ns} = '{$t->getNamespace()}'" : '' );
 168                 }
 169                 $this->addWhere( $db->makeList( $titleWhere, LIST_OR ) );
 170                 $this->addWhereFld( 'page_namespace', $this->params['namespace'] );
 171
 172                 if ( !is_null( $this->redirID ) ) {
 173                         $first = $this->redirTitles[0];
 174                         $title = $db->strencode( $first->getDBkey() );
 175                         $ns = $first->getNamespace();
 176                         $from = $this->redirID;
 177                         if ( $this->hasNS ) {
 178                                 $this->addWhere( "{$this->bl_ns} > $ns OR " .
 179                                                 "({$this->bl_ns} = $ns AND " .
 180                                                 "({$this->bl_title} > '$title' OR " .
 181                                                 "({$this->bl_title} = '$title' AND " .
 182                                                 "{$this->bl_from} >= $from)))" );
 183                         } else {
 184                                 $this->addWhere( "{$this->bl_title} > '$title' OR " .
 185                                                 "({$this->bl_title} = '$title' AND " .
 186                                                 "{$this->bl_from} >= $from)" );
 187                         }
 188                 }
 189                 if ( $this->params['filterredir'] == 'redirects' ) {
 190                         $this->addWhereFld( 'page_is_redirect', 1 );
 191                 } elseif ( $this->params['filterredir'] == 'nonredirects' ) {
 192                         $this->addWhereFld( 'page_is_redirect', 0 );
 193                 }
 194
 195                 $this->addOption( 'LIMIT', $this->params['limit'] + 1 );
 196                 $this->addOption( 'ORDER BY', $this->bl_sort );
 197                 $this->addOption( 'USE INDEX', array( 'page' => 'PRIMARY' ) );
 198         }
 199
 200         private function run( $resultPageSet = null ) {
 201                 $this->params = $this->extractRequestParams( false );
 202                 $this->redirect = isset( $this->params['redirect'] ) && $this->params['redirect'];
 203                 $userMax = ( $this->redirect ? ApiBase::LIMIT_BIG1 / 2 : ApiBase::LIMIT_BIG1 );
 204                 $botMax  = ( $this->redirect ? ApiBase::LIMIT_BIG2 / 2 : ApiBase::LIMIT_BIG2 );
 205                 if ( $this->params['limit'] == 'max' ) {
 206                         $this->params['limit'] = $this->getMain()->canApiHighLimits() ? $botMax : $userMax;
 207                         $this->getResult()->setParsedLimit( $this->getModuleName(), $this->params['limit'] );
 208                 }
 209
 210                 $this->processContinue();
 211                 $this->prepareFirstQuery( $resultPageSet );
 212
 213                 $res = $this->select( __METHOD__ . '::firstQuery' );
 214
 215                 $count = 0;
 216                 $this->pageMap = array(); // Maps ns and title to pageid
 217                 $this->continueStr = null;
 218                 $this->redirTitles = array();
 219                 foreach ( $res as $row ) {
 220                         if ( ++ $count > $this->params['limit'] ) {
 221                                 // We've reached the one extra which shows that there are additional pages to be had. Stop here...
 222                                 // Continue string preserved in case the redirect query doesn't pass the limit
 223                                 $this->continueStr = $this->getContinueStr( $row->page_id );
 224                                 break;
 225                         }
 226
 227                         if ( is_null( $resultPageSet ) ) {
 228                                 $this->extractRowInfo( $row );
 229                         } else {
 230                                 $this->pageMap[$row->page_namespace][$row->page_title] = $row->page_id;
 231                                 if ( $row->page_is_redirect ) {
 232                                         $this->redirTitles[] = Title::makeTitle( $row->page_namespace, $row->page_title );
 233                                 }
 234
 235                                 $resultPageSet->processDbRow( $row );
 236                         }
 237                 }
 238
 239                 if ( $this->redirect && count( $this->redirTitles ) ) {
 240                         $this->resetQueryParams();
 241                         $this->prepareSecondQuery( $resultPageSet );
 242                         $res = $this->select( __METHOD__ . '::secondQuery' );
 243                         $count = 0;
 244                         foreach ( $res as $row ) {
 245                                 if ( ++$count > $this->params['limit'] ) {
 246                                         // We've reached the one extra which shows that there are additional pages to be had. Stop here...
 247                                         // We need to keep the parent page of this redir in
 248                                         if ( $this->hasNS ) {
 249                                                 $parentID = $this->pageMap[$row-> { $this->bl_ns } ][$row-> { $this->bl_title } ];
 250                                         } else {
 251                                                 $parentID = $this->pageMap[NS_IMAGE][$row-> { $this->bl_title } ];
 252                                         }
 253                                         $this->continueStr = $this->getContinueRedirStr( $parentID, $row->page_id );
 254                                         break;
 255                                 }
 256
 257                                 if ( is_null( $resultPageSet ) ) {
 258                                         $this->extractRedirRowInfo( $row );
 259                                 } else {
 260                                         $resultPageSet->processDbRow( $row );
 261                                 }
 262                         }
 263                 }
 264                 if ( is_null( $resultPageSet ) ) {
 265                         // Try to add the result data in one go and pray that it fits
 266                         $fit = $this->getResult()->addValue( 'query', $this->getModuleName(), array_values( $this->resultArr ) );
 267                         if ( !$fit ) {
 268                                 // It didn't fit. Add elements one by one until the
 269                                 // result is full.
 270                                 foreach ( $this->resultArr as $pageID => $arr ) {
 271                                         // Add the basic entry without redirlinks first
 272                                         $fit = $this->getResult()->addValue(
 273                                                 array( 'query', $this->getModuleName() ),
 274                                                 null, array_diff_key( $arr, array( 'redirlinks' => '' ) ) );
 275                                         if ( !$fit ) {
 276                                                 $this->continueStr = $this->getContinueStr( $pageID );
 277                                                 break;
 278                                         }
 279
 280                                         $hasRedirs = false;
 281                                         foreach ( (array)@$arr['redirlinks'] as $key => $redir ) {
 282                                                 $fit = $this->getResult()->addValue(
 283                                                         array( 'query', $this->getModuleName(), $pageID, 'redirlinks' ),
 284                                                         $key, $redir );
 285                                                 if ( !$fit ) {
 286                                                         $this->continueStr = $this->getContinueRedirStr( $pageID, $redir['pageid'] );
 287                                                         break;
 288                                                 }
 289                                                 $hasRedirs = true;
 290                                         }
 291                                         if ( $hasRedirs ) {
 292                                                 $this->getResult()->setIndexedTagName_internal(
 293                                                         array( 'query', $this->getModuleName(), $pageID, 'redirlinks' ),
 294                                                         $this->bl_code );
 295                                         }
 296                                         if ( !$fit ) {
 297                                                 break;
 298                                         }
 299                                 }
 300                         }
 301
 302                         $this->getResult()->setIndexedTagName_internal(
 303                                 array( 'query', $this->getModuleName() ),
 304                                 $this->bl_code
 305                         );
 306                 }
 307                 if ( !is_null( $this->continueStr ) ) {
 308                         $this->setContinueEnumParameter( 'continue', $this->continueStr );
 309                 }
 310         }
 311
 312         private function extractRowInfo( $row ) {
 313                 $this->pageMap[$row->page_namespace][$row->page_title] = $row->page_id;
 314                 $t = Title::makeTitle( $row->page_namespace, $row->page_title );
 315                 $a = array( 'pageid' => intval( $row->page_id ) );
 316                 ApiQueryBase::addTitleInfo( $a, $t );
 317                 if ( $row->page_is_redirect ) {
 318                         $a['redirect'] = '';
 319                         $this->redirTitles[] = $t;
 320                 }
 321                 // Put all the results in an array first
 322                 $this->resultArr[$a['pageid']] = $a;
 323         }
 324
 325         private function extractRedirRowInfo( $row ) {
 326                 $a['pageid'] = intval( $row->page_id );
 327                 ApiQueryBase::addTitleInfo( $a, Title::makeTitle( $row->page_namespace, $row->page_title ) );
 328                 if ( $row->page_is_redirect ) {
 329                         $a['redirect'] = '';
 330                 }
 331                 $ns = $this->hasNS ? $row-> { $this->bl_ns } : NS_FILE;
 332                 $parentID = $this->pageMap[$ns][$row-> { $this->bl_title } ];
 333                 // Put all the results in an array first
 334                 $this->resultArr[$parentID]['redirlinks'][] = $a;
 335                 $this->getResult()->setIndexedTagName( $this->resultArr[$parentID]['redirlinks'], $this->bl_code );
 336         }
 337
 338         protected function processContinue() {
 339                 if ( !is_null( $this->params['continue'] ) ) {
 340                         $this->parseContinueParam();
 341                 } else {
 342                         if ( $this->params['title'] !== '' ) {
 343                                 $title = Title::newFromText( $this->params['title'] );
 344                                 if ( !$title ) {
 345                                         $this->dieUsageMsg( array( 'invalidtitle', $this->params['title'] ) );
 346                                 } else {
 347                                         $this->rootTitle = $title;
 348                                 }
 349                         }
 350                 }
 351
 352                 // only image titles are allowed for the root in imageinfo mode
 353                 if ( !$this->hasNS && $this->rootTitle->getNamespace() !== NS_FILE ) {
 354                         $this->dieUsage( "The title for {$this->getModuleName()} query must be an image", 'bad_image_title' );
 355                 }
 356         }
 357
 358         protected function parseContinueParam() {
 359                 $continueList = explode( '|', $this->params['continue'] );
 360                 // expected format:
 361                 // ns | key | id1 [| id2]
 362                 // ns+key: root title
 363                 // id1: first-level page ID to continue from
 364                 // id2: second-level page ID to continue from
 365
 366                 // null stuff out now so we know what's set and what isn't
 367                 $this->rootTitle = $this->contID = $this->redirID = null;
 368                 $rootNs = intval( $continueList[0] );
 369                 if ( $rootNs === 0 && $continueList[0] !== '0' ) {
 370                         // Illegal continue parameter
 371                         $this->dieUsage( 'Invalid continue param. You should pass the original value returned by the previous query', '_badcontinue' );
 372                 }
 373                 $this->rootTitle = Title::makeTitleSafe( $rootNs, $continueList[1] );
 374
 375                 if ( !$this->rootTitle ) {
 376                         $this->dieUsage( 'Invalid continue param. You should pass the original value returned by the previous query', '_badcontinue' );
 377                 }
 378                 $contID = intval( $continueList[2] );
 379
 380                 if ( $contID === 0 && $continueList[2] !== '0' ) {
 381                         $this->dieUsage( 'Invalid continue param. You should pass the original value returned by the previous query', '_badcontinue' );
 382                 }
 383                 $this->contID = $contID;
 384                 $redirID = intval( @$continueList[3] );
 385
 386                 if ( $redirID === 0 && @$continueList[3] !== '0' ) {
 387                         // This one isn't required
 388                         return;
 389                 }
 390                 $this->redirID = $redirID;
 391
 392         }
 393
 394         protected function getContinueStr( $lastPageID ) {
 395                 return $this->rootTitle->getNamespace() .
 396                 '|' . $this->rootTitle->getDBkey() .
 397                 '|' . $lastPageID;
 398         }
 399
 400         protected function getContinueRedirStr( $lastPageID, $lastRedirID ) {
 401                 return $this->getContinueStr( $lastPageID ) . '|' . $lastRedirID;
 402         }
 403
 404         public function getAllowedParams() {
 405                 $retval = array(
 406                         'title' => array(
 407                                 ApiBase::PARAM_TYPE => 'string',
 408                                 ApiBase::PARAM_REQUIRED => true
 409                         ),
 410                         'continue' => null,
 411                         'namespace' => array(
 412                                 ApiBase::PARAM_ISMULTI => true,
 413                                 ApiBase::PARAM_TYPE => 'namespace'
 414                         ),
 415                         'filterredir' => array(
 416                                 ApiBase::PARAM_DFLT => 'all',
 417                                 ApiBase::PARAM_TYPE => array(
 418                                         'all',
 419                                         'redirects',
 420                                         'nonredirects'
 421                                 )
 422                         ),
 423                         'limit' => array(
 424                                 ApiBase::PARAM_DFLT => 10,
 425                                 ApiBase::PARAM_TYPE => 'limit',
 426                                 ApiBase::PARAM_MIN => 1,
 427                                 ApiBase::PARAM_MAX => ApiBase::LIMIT_BIG1,
 428                                 ApiBase::PARAM_MAX2 => ApiBase::LIMIT_BIG2
 429                         )
 430                 );
 431                 if ( $this->getModuleName() == 'embeddedin' ) {
 432                         return $retval;
 433                 }
 434                 $retval['redirect'] = false;
 435                 return $retval;
 436         }
 437
 438         public function getParamDescription() {
 439                 $retval = array(
 440                         'title' => 'Title to search',
 441                         'continue' => 'When more results are available, use this to continue',
 442                         'namespace' => 'The namespace to enumerate',
 443                 );
 444                 if ( $this->getModuleName() != 'embeddedin' ) {
 445                         return array_merge( $retval, array(
 446                                 'redirect' => 'If linking page is a redirect, find all pages that link to that redirect as well. Maximum limit is halved.',
 447                                 'filterredir' => "How to filter for redirects. If set to nonredirects when {$this->bl_code}redirect is enabled, this is only applied to the second level",
 448                                 '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)."
 449                         ) );
 450                 }
 451                 return array_merge( $retval, array(
 452                         'filterredir' => 'How to filter for redirects',
 453                         'limit' => 'How many total pages to return'
 454                 ) );
 455         }
 456
 457         public function getDescription() {
 458                 switch ( $this->getModuleName() ) {
 459                         case 'backlinks':
 460                                 return 'Find all pages that link to the given page';
 461                         case 'embeddedin':
 462                                 return 'Find all pages that embed (transclude) the given title';
 463                         case 'imageusage':
 464                                 return 'Find all pages that use the given image title.';
 465                         default:
 466                                 ApiBase::dieDebug( __METHOD__, 'Unknown module name' );
 467                 }
 468         }
 469
 470         public function getPossibleErrors() {
 471                 return array_merge( parent::getPossibleErrors(), array(
 472                         array( 'invalidtitle', 'title' ),
 473                         array( 'code' => 'bad_image_title', 'info' => "The title for {$this->getModuleName()} query must be an image" ),
 474                         array( 'code' => '_badcontinue', 'info' => 'Invalid continue param. You should pass the original value returned by the previous query' ),
 475                 ) );
 476         }
 477
 478         protected function getExamples() {
 479                 static $examples = array(
 480                         'backlinks' => array(
 481                                 'api.php?action=query&list=backlinks&bltitle=Main%20Page',
 482                                 'api.php?action=query&generator=backlinks&gbltitle=Main%20Page&prop=info'
 483                         ),
 484                         'embeddedin' => array(
 485                                 'api.php?action=query&list=embeddedin&eititle=Template:Stub',
 486                                 'api.php?action=query&generator=embeddedin&geititle=Template:Stub&prop=info'
 487                         ),
 488                         'imageusage' => array(
 489                                 'api.php?action=query&list=imageusage&iutitle=File:Albert%20Einstein%20Head.jpg',
 490                                 'api.php?action=query&generator=imageusage&giutitle=File:Albert%20Einstein%20Head.jpg&prop=info'
 491                         )
 492                 );
 493
 494                 return $examples[$this->getModuleName()];
 495         }
 496
 497         public function getVersion() {
 498                 return __CLASS__ . ': $Id$';
 499         }
 500 }