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 /**
  28  * This is a three-in-one module to query:
  29  *   * backlinks  - links pointing to the given page,
  30  *   * embeddedin - what pages transclude the given page within themselves,
  31  *   * imageusage - what pages use the given image
  32  *
  33  * @ingroup API
  34  */
  35 class ApiQueryBacklinks extends ApiQueryGeneratorBase {
  36
  37         /**
  38          * @var Title
  39          */
  40         private $rootTitle;
  41
  42         private $params, $contID, $redirID, $redirect;
  43         private $bl_ns, $bl_from, $bl_table, $bl_code, $bl_title, $bl_sort, $bl_fields, $hasNS;
  44
  45         /**
  46          * Maps ns and title to pageid
  47          *
  48          * @var array
  49          */
  50         private $pageMap = array();
  51         private $resultArr;
  52
  53         private $redirTitles = array();
  54         private $continueStr = null;
  55
  56         // output element name, database column field prefix, database table
  57         private $backlinksSettings = array(
  58                 'backlinks' => array(
  59                         'code' => 'bl',
  60                         'prefix' => 'pl',
  61                         'linktbl' => 'pagelinks',
  62                         'helpurl' => 'https://www.mediawiki.org/wiki/API:Backlinks',
  63                 ),
  64                 'embeddedin' => array(
  65                         'code' => 'ei',
  66                         'prefix' => 'tl',
  67                         'linktbl' => 'templatelinks',
  68                         'helpurl' => 'https://www.mediawiki.org/wiki/API:Embeddedin',
  69                 ),
  70                 'imageusage' => array(
  71                         'code' => 'iu',
  72                         'prefix' => 'il',
  73                         'linktbl' => 'imagelinks',
  74                         'helpurl' => 'https://www.mediawiki.org/wiki/API:Imageusage',
  75                 )
  76         );
  77
  78         public function __construct( $query, $moduleName ) {
  79                 $settings = $this->backlinksSettings[$moduleName];
  80                 $prefix = $settings['prefix'];
  81                 $code = $settings['code'];
  82                 $this->resultArr = array();
  83
  84                 parent::__construct( $query, $moduleName, $code );
  85                 $this->bl_ns = $prefix . '_namespace';
  86                 $this->bl_from = $prefix . '_from';
  87                 $this->bl_table = $settings['linktbl'];
  88                 $this->bl_code = $code;
  89                 $this->helpUrl = $settings['helpurl'];
  90
  91                 $this->hasNS = $moduleName !== 'imageusage';
  92                 if ( $this->hasNS ) {
  93                         $this->bl_title = $prefix . '_title';
  94                         $this->bl_sort = "{$this->bl_ns}, {$this->bl_title}, {$this->bl_from}";
  95                         $this->bl_fields = array(
  96                                 $this->bl_ns,
  97                                 $this->bl_title
  98                         );
  99                 } else {
 100                         $this->bl_title = $prefix . '_to';
 101                         $this->bl_sort = "{$this->bl_title}, {$this->bl_from}";
 102                         $this->bl_fields = array(
 103                                 $this->bl_title
 104                         );
 105                 }
 106         }
 107
 108         public function execute() {
 109                 $this->run();
 110         }
 111
 112         public function getCacheMode( $params ) {
 113                 return 'public';
 114         }
 115
 116         public function executeGenerator( $resultPageSet ) {
 117                 $this->run( $resultPageSet );
 118         }
 119
 120         /**
 121          * @param $resultPageSet ApiPageSet
 122          * @return void
 123          */
 124         private function prepareFirstQuery( $resultPageSet = null ) {
 125                 /* SELECT page_id, page_title, page_namespace, page_is_redirect
 126                  * FROM pagelinks, page WHERE pl_from=page_id
 127                  * AND pl_title='Foo' AND pl_namespace=0
 128                  * LIMIT 11 ORDER BY pl_from
 129                  */
 130                 $this->addTables( array( $this->bl_table, 'page' ) );
 131                 $this->addWhere( "{$this->bl_from}=page_id" );
 132                 if ( is_null( $resultPageSet ) ) {
 133                         $this->addFields( array( 'page_id', 'page_title', 'page_namespace' ) );
 134                 } else {
 135                         $this->addFields( $resultPageSet->getPageTableFields() );
 136                 }
 137
 138                 $this->addFields( 'page_is_redirect' );
 139                 $this->addWhereFld( $this->bl_title, $this->rootTitle->getDBkey() );
 140
 141                 if ( $this->hasNS ) {
 142                         $this->addWhereFld( $this->bl_ns, $this->rootTitle->getNamespace() );
 143                 }
 144                 $this->addWhereFld( 'page_namespace', $this->params['namespace'] );
 145
 146                 if ( !is_null( $this->contID ) ) {
 147                         $this->addWhere( "{$this->bl_from}>={$this->contID}" );
 148                 }
 149
 150                 if ( $this->params['filterredir'] == 'redirects' ) {
 151                         $this->addWhereFld( 'page_is_redirect', 1 );
 152                 } elseif ( $this->params['filterredir'] == 'nonredirects' && !$this->redirect ) {
 153                         // bug 22245 - Check for !redirect, as filtering nonredirects, when getting what links to them is contradictory
 154                         $this->addWhereFld( 'page_is_redirect', 0 );
 155                 }
 156
 157                 $this->addOption( 'LIMIT', $this->params['limit'] + 1 );
 158                 $this->addOption( 'ORDER BY', $this->bl_from );
 159                 $this->addOption( 'STRAIGHT_JOIN' );
 160         }
 161
 162         /**
 163          * @param $resultPageSet ApiPageSet
 164          * @return void
 165          */
 166         private function prepareSecondQuery( $resultPageSet = null ) {
 167                 /* SELECT page_id, page_title, page_namespace, page_is_redirect, pl_title, pl_namespace
 168                    FROM pagelinks, page WHERE pl_from=page_id
 169                    AND (pl_title='Foo' AND pl_namespace=0) OR (pl_title='Bar' AND pl_namespace=1)
 170                    ORDER BY pl_namespace, pl_title, pl_from LIMIT 11
 171                  */
 172                 $db = $this->getDB();
 173                 $this->addTables( array( 'page', $this->bl_table ) );
 174                 $this->addWhere( "{$this->bl_from}=page_id" );
 175
 176                 if ( is_null( $resultPageSet ) ) {
 177                         $this->addFields( array( 'page_id', 'page_title', 'page_namespace', 'page_is_redirect' ) );
 178                 } else {
 179                         $this->addFields( $resultPageSet->getPageTableFields() );
 180                 }
 181
 182                 $this->addFields( $this->bl_title );
 183                 if ( $this->hasNS ) {
 184                         $this->addFields( $this->bl_ns );
 185                 }
 186
 187                 // We can't use LinkBatch here because $this->hasNS may be false
 188                 $titleWhere = array();
 189                 foreach ( $this->redirTitles as $t ) {
 190                         $titleWhere[] = "{$this->bl_title} = " . $db->addQuotes( $t->getDBkey() ) .
 191                                         ( $this->hasNS ? " AND {$this->bl_ns} = {$t->getNamespace()}" : '' );
 192                 }
 193                 $this->addWhere( $db->makeList( $titleWhere, LIST_OR ) );
 194                 $this->addWhereFld( 'page_namespace', $this->params['namespace'] );
 195
 196                 if ( !is_null( $this->redirID ) ) {
 197                         $first = $this->redirTitles[0];
 198                         $title = $db->addQuotes( $first->getDBkey() );
 199                         $ns = $first->getNamespace();
 200                         $from = $this->redirID;
 201                         if ( $this->hasNS ) {
 202                                 $this->addWhere( "{$this->bl_ns} > $ns OR " .
 203                                                 "({$this->bl_ns} = $ns AND " .
 204                                                 "({$this->bl_title} > $title OR " .
 205                                                 "({$this->bl_title} = $title AND " .
 206                                                 "{$this->bl_from} >= $from)))" );
 207                         } else {
 208                                 $this->addWhere( "{$this->bl_title} > $title OR " .
 209                                                 "({$this->bl_title} = $title AND " .
 210                                                 "{$this->bl_from} >= $from)" );
 211                         }
 212                 }
 213                 if ( $this->params['filterredir'] == 'redirects' ) {
 214                         $this->addWhereFld( 'page_is_redirect', 1 );
 215                 } elseif ( $this->params['filterredir'] == 'nonredirects' ) {
 216                         $this->addWhereFld( 'page_is_redirect', 0 );
 217                 }
 218
 219                 $this->addOption( 'LIMIT', $this->params['limit'] + 1 );
 220                 $this->addOption( 'ORDER BY', $this->bl_sort );
 221                 $this->addOption( 'USE INDEX', array( 'page' => 'PRIMARY' ) );
 222         }
 223
 224         /**
 225          * @param $resultPageSet ApiPageSet
 226          * @return void
 227          */
 228         private function run( $resultPageSet = null ) {
 229                 $this->params = $this->extractRequestParams( false );
 230                 $this->redirect = isset( $this->params['redirect'] ) && $this->params['redirect'];
 231                 $userMax = ( $this->redirect ? ApiBase::LIMIT_BIG1 / 2 : ApiBase::LIMIT_BIG1 );
 232                 $botMax  = ( $this->redirect ? ApiBase::LIMIT_BIG2 / 2 : ApiBase::LIMIT_BIG2 );
 233
 234                 $result = $this->getResult();
 235
 236                 if ( $this->params['limit'] == 'max' ) {
 237                         $this->params['limit'] = $this->getMain()->canApiHighLimits() ? $botMax : $userMax;
 238                         $result->setParsedLimit( $this->getModuleName(), $this->params['limit'] );
 239                 }
 240
 241                 $this->processContinue();
 242                 $this->prepareFirstQuery( $resultPageSet );
 243
 244                 $res = $this->select( __METHOD__ . '::firstQuery' );
 245
 246                 $count = 0;
 247
 248                 foreach ( $res as $row ) {
 249                         if ( ++ $count > $this->params['limit'] ) {
 250                                 // We've reached the one extra which shows that there are additional pages to be had. Stop here...
 251                                 // Continue string preserved in case the redirect query doesn't pass the limit
 252                                 $this->continueStr = $this->getContinueStr( $row->page_id );
 253                                 break;
 254                         }
 255
 256                         if ( is_null( $resultPageSet ) ) {
 257                                 $this->extractRowInfo( $row );
 258                         } else {
 259                                 $this->pageMap[$row->page_namespace][$row->page_title] = $row->page_id;
 260                                 if ( $row->page_is_redirect ) {
 261                                         $this->redirTitles[] = Title::makeTitle( $row->page_namespace, $row->page_title );
 262                                 }
 263
 264                                 $resultPageSet->processDbRow( $row );
 265                         }
 266                 }
 267
 268                 if ( $this->redirect && count( $this->redirTitles ) ) {
 269                         $this->resetQueryParams();
 270                         $this->prepareSecondQuery( $resultPageSet );
 271                         $res = $this->select( __METHOD__ . '::secondQuery' );
 272                         $count = 0;
 273                         foreach ( $res as $row ) {
 274                                 if ( ++$count > $this->params['limit'] ) {
 275                                         // We've reached the one extra which shows that there are additional pages to be had. Stop here...
 276                                         // We need to keep the parent page of this redir in
 277                                         if ( $this->hasNS ) {
 278                                                 $parentID = $this->pageMap[$row-> { $this->bl_ns } ][$row-> { $this->bl_title } ];
 279                                         } else {
 280                                                 $parentID = $this->pageMap[NS_FILE][$row-> { $this->bl_title } ];
 281                                         }
 282                                         $this->continueStr = $this->getContinueRedirStr( $parentID, $row->page_id );
 283                                         break;
 284                                 }
 285
 286                                 if ( is_null( $resultPageSet ) ) {
 287                                         $this->extractRedirRowInfo( $row );
 288                                 } else {
 289                                         $resultPageSet->processDbRow( $row );
 290                                 }
 291                         }
 292                 }
 293                 if ( is_null( $resultPageSet ) ) {
 294                         // Try to add the result data in one go and pray that it fits
 295                         $fit = $result->addValue( 'query', $this->getModuleName(), array_values( $this->resultArr ) );
 296                         if ( !$fit ) {
 297                                 // It didn't fit. Add elements one by one until the
 298                                 // result is full.
 299                                 foreach ( $this->resultArr as $pageID => $arr ) {
 300                                         // Add the basic entry without redirlinks first
 301                                         $fit = $result->addValue(
 302                                                 array( 'query', $this->getModuleName() ),
 303                                                 null, array_diff_key( $arr, array( 'redirlinks' => '' ) ) );
 304                                         if ( !$fit ) {
 305                                                 $this->continueStr = $this->getContinueStr( $pageID );
 306                                                 break;
 307                                         }
 308
 309                                         $hasRedirs = false;
 310                                         $redirLinks = isset( $arr['redirlinks'] ) ? $arr['redirlinks'] : array();
 311                                         foreach ( (array)$redirLinks as $key => $redir ) {
 312                                                 $fit = $result->addValue(
 313                                                         array( 'query', $this->getModuleName(), $pageID, 'redirlinks' ),
 314                                                         $key, $redir );
 315                                                 if ( !$fit ) {
 316                                                         $this->continueStr = $this->getContinueRedirStr( $pageID, $redir['pageid'] );
 317                                                         break;
 318                                                 }
 319                                                 $hasRedirs = true;
 320                                         }
 321                                         if ( $hasRedirs ) {
 322                                                 $result->setIndexedTagName_internal(
 323                                                         array( 'query', $this->getModuleName(), $pageID, 'redirlinks' ),
 324                                                         $this->bl_code );
 325                                         }
 326                                         if ( !$fit ) {
 327                                                 break;
 328                                         }
 329                                 }
 330                         }
 331
 332                         $result->setIndexedTagName_internal(
 333                                 array( 'query', $this->getModuleName() ),
 334                                 $this->bl_code
 335                         );
 336                 }
 337                 if ( !is_null( $this->continueStr ) ) {
 338                         $this->setContinueEnumParameter( 'continue', $this->continueStr );
 339                 }
 340         }
 341
 342         private function extractRowInfo( $row ) {
 343                 $this->pageMap[$row->page_namespace][$row->page_title] = $row->page_id;
 344                 $t = Title::makeTitle( $row->page_namespace, $row->page_title );
 345                 $a = array( 'pageid' => intval( $row->page_id ) );
 346                 ApiQueryBase::addTitleInfo( $a, $t );
 347                 if ( $row->page_is_redirect ) {
 348                         $a['redirect'] = '';
 349                         $this->redirTitles[] = $t;
 350                 }
 351                 // Put all the results in an array first
 352                 $this->resultArr[$a['pageid']] = $a;
 353         }
 354
 355         private function extractRedirRowInfo( $row ) {
 356                 $a['pageid'] = intval( $row->page_id );
 357                 ApiQueryBase::addTitleInfo( $a, Title::makeTitle( $row->page_namespace, $row->page_title ) );
 358                 if ( $row->page_is_redirect ) {
 359                         $a['redirect'] = '';
 360                 }
 361                 $ns = $this->hasNS ? $row-> { $this->bl_ns } : NS_FILE;
 362                 $parentID = $this->pageMap[$ns][$row-> { $this->bl_title } ];
 363                 // Put all the results in an array first
 364                 $this->resultArr[$parentID]['redirlinks'][] = $a;
 365                 $this->getResult()->setIndexedTagName( $this->resultArr[$parentID]['redirlinks'], $this->bl_code );
 366         }
 367
 368         protected function processContinue() {
 369                 if ( !is_null( $this->params['continue'] ) ) {
 370                         $this->parseContinueParam();
 371                 } else {
 372                         $this->rootTitle = $this->getTitleOrPageId( $this->params )->getTitle();
 373                 }
 374
 375                 // only image titles are allowed for the root in imageinfo mode
 376                 if ( !$this->hasNS && $this->rootTitle->getNamespace() !== NS_FILE ) {
 377                         $this->dieUsage( "The title for {$this->getModuleName()} query must be an image", 'bad_image_title' );
 378                 }
 379         }
 380
 381         protected function parseContinueParam() {
 382                 $continueList = explode( '|', $this->params['continue'] );
 383                 // expected format:
 384                 // ns | key | id1 [| id2]
 385                 // ns+key: root title
 386                 // id1: first-level page ID to continue from
 387                 // id2: second-level page ID to continue from
 388
 389                 // null stuff out now so we know what's set and what isn't
 390                 $this->rootTitle = $this->contID = $this->redirID = null;
 391                 $rootNs = intval( $continueList[0] );
 392                 if ( $rootNs === 0 && $continueList[0] !== '0' ) {
 393                         // Illegal continue parameter
 394                         $this->dieUsage( 'Invalid continue param. You should pass the original value returned by the previous query', '_badcontinue' );
 395                 }
 396                 $this->rootTitle = Title::makeTitleSafe( $rootNs, $continueList[1] );
 397
 398                 if ( !$this->rootTitle ) {
 399                         $this->dieUsage( 'Invalid continue param. You should pass the original value returned by the previous query', '_badcontinue' );
 400                 }
 401                 $contID = intval( $continueList[2] );
 402
 403                 if ( $contID === 0 && $continueList[2] !== '0' ) {
 404                         $this->dieUsage( 'Invalid continue param. You should pass the original value returned by the previous query', '_badcontinue' );
 405                 }
 406                 $this->contID = $contID;
 407                 $id2 = isset( $continueList[3] ) ? $continueList[3] : null;
 408                 $redirID = intval( $id2 );
 409
 410                 if ( $redirID === 0 && $id2 !== '0' ) {
 411                         // This one isn't required
 412                         return;
 413                 }
 414                 $this->redirID = $redirID;
 415
 416         }
 417
 418         protected function getContinueStr( $lastPageID ) {
 419                 return $this->rootTitle->getNamespace() .
 420                 '|' . $this->rootTitle->getDBkey() .
 421                 '|' . $lastPageID;
 422         }
 423
 424         protected function getContinueRedirStr( $lastPageID, $lastRedirID ) {
 425                 return $this->getContinueStr( $lastPageID ) . '|' . $lastRedirID;
 426         }
 427
 428         public function getAllowedParams() {
 429                 $retval = array(
 430                         'title' => array(
 431                                 ApiBase::PARAM_TYPE => 'string',
 432                         ),
 433                         'pageid' => array(
 434                                 ApiBase::PARAM_TYPE => 'integer',
 435                         ),
 436                         'continue' => null,
 437                         'namespace' => array(
 438                                 ApiBase::PARAM_ISMULTI => true,
 439                                 ApiBase::PARAM_TYPE => 'namespace'
 440                         ),
 441                         'filterredir' => array(
 442                                 ApiBase::PARAM_DFLT => 'all',
 443                                 ApiBase::PARAM_TYPE => array(
 444                                         'all',
 445                                         'redirects',
 446                                         'nonredirects'
 447                                 )
 448                         ),
 449                         'limit' => array(
 450                                 ApiBase::PARAM_DFLT => 10,
 451                                 ApiBase::PARAM_TYPE => 'limit',
 452                                 ApiBase::PARAM_MIN => 1,
 453                                 ApiBase::PARAM_MAX => ApiBase::LIMIT_BIG1,
 454                                 ApiBase::PARAM_MAX2 => ApiBase::LIMIT_BIG2
 455                         )
 456                 );
 457                 if ( $this->getModuleName() == 'embeddedin' ) {
 458                         return $retval;
 459                 }
 460                 $retval['redirect'] = false;
 461                 return $retval;
 462         }
 463
 464         public function getParamDescription() {
 465                 $retval = array(
 466                         'title' => "Title to search. Cannot be used together with {$this->bl_code}pageid",
 467                         'pageid' => "Pageid to search. Cannot be used together with {$this->bl_code}title",
 468                         'continue' => 'When more results are available, use this to continue',
 469                         'namespace' => 'The namespace to enumerate',
 470                 );
 471                 if ( $this->getModuleName() != 'embeddedin' ) {
 472                         return array_merge( $retval, array(
 473                                 'redirect' => 'If linking page is a redirect, find all pages that link to that redirect as well. Maximum limit is halved.',
 474                                 'filterredir' => "How to filter for redirects. If set to nonredirects when {$this->bl_code}redirect is enabled, this is only applied to the second level",
 475                                 '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)."
 476                         ) );
 477                 }
 478                 return array_merge( $retval, array(
 479                         'filterredir' => 'How to filter for redirects',
 480                         'limit' => 'How many total pages to return'
 481                 ) );
 482         }
 483
 484         public function getResultProperties() {
 485                 return array(
 486                         '' => array(
 487                                 'pageid' => 'integer',
 488                                 'ns' => 'namespace',
 489                                 'title' => 'string',
 490                                 'redirect' => 'boolean'
 491                         )
 492                 );
 493         }
 494
 495         public function getDescription() {
 496                 switch ( $this->getModuleName() ) {
 497                         case 'backlinks':
 498                                 return 'Find all pages that link to the given page';
 499                         case 'embeddedin':
 500                                 return 'Find all pages that embed (transclude) the given title';
 501                         case 'imageusage':
 502                                 return 'Find all pages that use the given image title.';
 503                         default:
 504                                 ApiBase::dieDebug( __METHOD__, 'Unknown module name' );
 505                 }
 506         }
 507
 508         public function getPossibleErrors() {
 509                 return array_merge( parent::getPossibleErrors(),
 510                         $this->getTitleOrPageIdErrorMessage(),
 511                         array(
 512                                 array( 'code' => 'bad_image_title', 'info' => "The title for {$this->getModuleName()} query must be an image" ),
 513                                 array( 'code' => '_badcontinue', 'info' => 'Invalid continue param. You should pass the original value returned by the previous query' ),
 514                         )
 515                 );
 516         }
 517
 518         public function getExamples() {
 519                 static $examples = array(
 520                         'backlinks' => array(
 521                                 'api.php?action=query&list=backlinks&bltitle=Main%20Page',
 522                                 'api.php?action=query&generator=backlinks&gbltitle=Main%20Page&prop=info'
 523                         ),
 524                         'embeddedin' => array(
 525                                 'api.php?action=query&list=embeddedin&eititle=Template:Stub',
 526                                 'api.php?action=query&generator=embeddedin&geititle=Template:Stub&prop=info'
 527                         ),
 528                         'imageusage' => array(
 529                                 'api.php?action=query&list=imageusage&iutitle=File:Albert%20Einstein%20Head.jpg',
 530                                 'api.php?action=query&generator=imageusage&giutitle=File:Albert%20Einstein%20Head.jpg&prop=info'
 531                         )
 532                 );
 533
 534                 return $examples[$this->getModuleName()];
 535         }
 536
 537         public function getHelpUrls() {
 538                 return $this->helpUrl;
 539         }
 540
 541         public function getVersion() {
 542                 return __CLASS__ . ': $Id$';
 543         }
 544 }