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