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