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_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_fields = array(
  95                                 $this->bl_ns,
  96                                 $this->bl_title
  97                         );
  98                 } else {
  99                         $this->bl_title = $prefix . '_to';
 100                         $this->bl_fields = array(
 101                                 $this->bl_title
 102                         );
 103                 }
 104         }
 105
 106         public function execute() {
 107                 $this->run();
 108         }
 109
 110         public function getCacheMode( $params ) {
 111                 return 'public';
 112         }
 113
 114         public function executeGenerator( $resultPageSet ) {
 115                 $this->run( $resultPageSet );
 116         }
 117
 118         /**
 119          * @param $resultPageSet ApiPageSet
 120          * @return void
 121          */
 122         private function prepareFirstQuery( $resultPageSet = null ) {
 123                 /* SELECT page_id, page_title, page_namespace, page_is_redirect
 124                  * FROM pagelinks, page WHERE pl_from=page_id
 125                  * AND pl_title='Foo' AND pl_namespace=0
 126                  * LIMIT 11 ORDER BY pl_from
 127                  */
 128                 $this->addTables( array( $this->bl_table, 'page' ) );
 129                 $this->addWhere( "{$this->bl_from}=page_id" );
 130                 if ( is_null( $resultPageSet ) ) {
 131                         $this->addFields( array( 'page_id', 'page_title', 'page_namespace' ) );
 132                 } else {
 133                         $this->addFields( $resultPageSet->getPageTableFields() );
 134                 }
 135
 136                 $this->addFields( 'page_is_redirect' );
 137                 $this->addWhereFld( $this->bl_title, $this->rootTitle->getDBkey() );
 138
 139                 if ( $this->hasNS ) {
 140                         $this->addWhereFld( $this->bl_ns, $this->rootTitle->getNamespace() );
 141                 }
 142                 $this->addWhereFld( 'page_namespace', $this->params['namespace'] );
 143
 144                 if ( !is_null( $this->contID ) ) {
 145                         $op = $this->params['dir'] == 'descending' ? '<' : '>';
 146                         $this->addWhere( "{$this->bl_from}$op={$this->contID}" );
 147                 }
 148
 149                 if ( $this->params['filterredir'] == 'redirects' ) {
 150                         $this->addWhereFld( 'page_is_redirect', 1 );
 151                 } elseif ( $this->params['filterredir'] == 'nonredirects' && !$this->redirect ) {
 152                         // bug 22245 - Check for !redirect, as filtering nonredirects, when getting what links to them is contradictory
 153                         $this->addWhereFld( 'page_is_redirect', 0 );
 154                 }
 155
 156                 $this->addOption( 'LIMIT', $this->params['limit'] + 1 );
 157                 $sort = ( $this->params['dir'] == 'descending' ? ' DESC' : '' );
 158                 $this->addOption( 'ORDER BY', $this->bl_from . $sort );
 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                 $allRedirNs = array();
 190                 $allRedirDBkey = array();
 191                 /** @var $t Title */
 192                 foreach ( $this->redirTitles as $t ) {
 193                         $redirNs = $t->getNamespace();
 194                         $redirDBkey = $t->getDBkey();
 195                         $titleWhere[] = "{$this->bl_title} = " . $db->addQuotes( $redirDBkey ) .
 196                                         ( $this->hasNS ? " AND {$this->bl_ns} = {$redirNs}" : '' );
 197                         $allRedirNs[] = $redirNs;
 198                         $allRedirDBkey[] = $redirDBkey;
 199                 }
 200                 $this->addWhere( $db->makeList( $titleWhere, LIST_OR ) );
 201                 $this->addWhereFld( 'page_namespace', $this->params['namespace'] );
 202
 203                 if ( !is_null( $this->redirID ) ) {
 204                         $op = $this->params['dir'] == 'descending' ? '<' : '>';
 205                         /** @var $first Title */
 206                         $first = $this->redirTitles[0];
 207                         $title = $db->addQuotes( $first->getDBkey() );
 208                         $ns = $first->getNamespace();
 209                         $from = $this->redirID;
 210                         if ( $this->hasNS ) {
 211                                 $this->addWhere( "{$this->bl_ns} $op $ns OR " .
 212                                                 "({$this->bl_ns} = $ns AND " .
 213                                                 "({$this->bl_title} $op $title OR " .
 214                                                 "({$this->bl_title} = $title AND " .
 215                                                 "{$this->bl_from} $op= $from)))" );
 216                         } else {
 217                                 $this->addWhere( "{$this->bl_title} $op $title OR " .
 218                                                 "({$this->bl_title} = $title AND " .
 219                                                 "{$this->bl_from} $op= $from)" );
 220                         }
 221                 }
 222                 if ( $this->params['filterredir'] == 'redirects' ) {
 223                         $this->addWhereFld( 'page_is_redirect', 1 );
 224                 } elseif ( $this->params['filterredir'] == 'nonredirects' ) {
 225                         $this->addWhereFld( 'page_is_redirect', 0 );
 226                 }
 227
 228                 $this->addOption( 'LIMIT', $this->params['limit'] + 1 );
 229                 $orderBy = array();
 230                 $sort = ( $this->params['dir'] == 'descending' ? ' DESC' : '' );
 231                 // Don't order by namespace/title if it's constant in the WHERE clause
 232                 if ( $this->hasNS && count( array_unique( $allRedirNs ) ) != 1 ) {
 233                         $orderBy[] = $this->bl_ns . $sort;
 234                 }
 235                 if ( count( array_unique( $allRedirDBkey ) ) != 1 ) {
 236                         $orderBy[] = $this->bl_title . $sort;
 237                 }
 238                 $orderBy[] = $this->bl_from . $sort;
 239                 $this->addOption( 'ORDER BY', $orderBy );
 240                 $this->addOption( 'USE INDEX', array( 'page' => 'PRIMARY' ) );
 241         }
 242
 243         /**
 244          * @param $resultPageSet ApiPageSet
 245          * @return void
 246          */
 247         private function run( $resultPageSet = null ) {
 248                 $this->params = $this->extractRequestParams( false );
 249                 $this->redirect = isset( $this->params['redirect'] ) && $this->params['redirect'];
 250                 $userMax = ( $this->redirect ? ApiBase::LIMIT_BIG1 / 2 : ApiBase::LIMIT_BIG1 );
 251                 $botMax = ( $this->redirect ? ApiBase::LIMIT_BIG2 / 2 : ApiBase::LIMIT_BIG2 );
 252
 253                 $result = $this->getResult();
 254
 255                 if ( $this->params['limit'] == 'max' ) {
 256                         $this->params['limit'] = $this->getMain()->canApiHighLimits() ? $botMax : $userMax;
 257                         $result->setParsedLimit( $this->getModuleName(), $this->params['limit'] );
 258                 } else {
 259                         $this->params['limit'] = intval( $this->params['limit'] );
 260                         $this->validateLimit( 'limit', $this->params['limit'], 1, $userMax, $botMax );
 261                 }
 262
 263                 $this->processContinue();
 264                 $this->prepareFirstQuery( $resultPageSet );
 265
 266                 $res = $this->select( __METHOD__ . '::firstQuery' );
 267
 268                 $count = 0;
 269
 270                 foreach ( $res as $row ) {
 271                         if ( ++ $count > $this->params['limit'] ) {
 272                                 // We've reached the one extra which shows that there are additional pages to be had. Stop here...
 273                                 // Continue string preserved in case the redirect query doesn't pass the limit
 274                                 $this->continueStr = $this->getContinueStr( $row->page_id );
 275                                 break;
 276                         }
 277
 278                         if ( is_null( $resultPageSet ) ) {
 279                                 $this->extractRowInfo( $row );
 280                         } else {
 281                                 $this->pageMap[$row->page_namespace][$row->page_title] = $row->page_id;
 282                                 if ( $row->page_is_redirect ) {
 283                                         $this->redirTitles[] = Title::makeTitle( $row->page_namespace, $row->page_title );
 284                                 }
 285
 286                                 $resultPageSet->processDbRow( $row );
 287                         }
 288                 }
 289
 290                 if ( $this->redirect && count( $this->redirTitles ) ) {
 291                         $this->resetQueryParams();
 292                         $this->prepareSecondQuery( $resultPageSet );
 293                         $res = $this->select( __METHOD__ . '::secondQuery' );
 294                         $count = 0;
 295                         foreach ( $res as $row ) {
 296                                 if ( ++$count > $this->params['limit'] ) {
 297                                         // We've reached the one extra which shows that there are additional pages to be had. Stop here...
 298                                         // We need to keep the parent page of this redir in
 299                                         if ( $this->hasNS ) {
 300                                                 $parentID = $this->pageMap[$row->{$this->bl_ns}][$row->{$this->bl_title}];
 301                                         } else {
 302                                                 $parentID = $this->pageMap[NS_FILE][$row->{$this->bl_title}];
 303                                         }
 304                                         $this->continueStr = $this->getContinueRedirStr( $parentID, $row->page_id );
 305                                         break;
 306                                 }
 307
 308                                 if ( is_null( $resultPageSet ) ) {
 309                                         $this->extractRedirRowInfo( $row );
 310                                 } else {
 311                                         $resultPageSet->processDbRow( $row );
 312                                 }
 313                         }
 314                 }
 315                 if ( is_null( $resultPageSet ) ) {
 316                         // Try to add the result data in one go and pray that it fits
 317                         $fit = $result->addValue( 'query', $this->getModuleName(), array_values( $this->resultArr ) );
 318                         if ( !$fit ) {
 319                                 // It didn't fit. Add elements one by one until the
 320                                 // result is full.
 321                                 foreach ( $this->resultArr as $pageID => $arr ) {
 322                                         // Add the basic entry without redirlinks first
 323                                         $fit = $result->addValue(
 324                                                 array( 'query', $this->getModuleName() ),
 325                                                 null, array_diff_key( $arr, array( 'redirlinks' => '' ) ) );
 326                                         if ( !$fit ) {
 327                                                 $this->continueStr = $this->getContinueStr( $pageID );
 328                                                 break;
 329                                         }
 330
 331                                         $hasRedirs = false;
 332                                         $redirLinks = isset( $arr['redirlinks'] ) ? $arr['redirlinks'] : array();
 333                                         foreach ( (array)$redirLinks as $key => $redir ) {
 334                                                 $fit = $result->addValue(
 335                                                         array( 'query', $this->getModuleName(), $pageID, 'redirlinks' ),
 336                                                         $key, $redir );
 337                                                 if ( !$fit ) {
 338                                                         $this->continueStr = $this->getContinueRedirStr( $pageID, $redir['pageid'] );
 339                                                         break;
 340                                                 }
 341                                                 $hasRedirs = true;
 342                                         }
 343                                         if ( $hasRedirs ) {
 344                                                 $result->setIndexedTagName_internal(
 345                                                         array( 'query', $this->getModuleName(), $pageID, 'redirlinks' ),
 346                                                         $this->bl_code );
 347                                         }
 348                                         if ( !$fit ) {
 349                                                 break;
 350                                         }
 351                                 }
 352                         }
 353
 354                         $result->setIndexedTagName_internal(
 355                                 array( 'query', $this->getModuleName() ),
 356                                 $this->bl_code
 357                         );
 358                 }
 359                 if ( !is_null( $this->continueStr ) ) {
 360                         $this->setContinueEnumParameter( 'continue', $this->continueStr );
 361                 }
 362         }
 363
 364         private function extractRowInfo( $row ) {
 365                 $this->pageMap[$row->page_namespace][$row->page_title] = $row->page_id;
 366                 $t = Title::makeTitle( $row->page_namespace, $row->page_title );
 367                 $a = array( 'pageid' => intval( $row->page_id ) );
 368                 ApiQueryBase::addTitleInfo( $a, $t );
 369                 if ( $row->page_is_redirect ) {
 370                         $a['redirect'] = '';
 371                         $this->redirTitles[] = $t;
 372                 }
 373                 // Put all the results in an array first
 374                 $this->resultArr[$a['pageid']] = $a;
 375         }
 376
 377         private function extractRedirRowInfo( $row ) {
 378                 $a['pageid'] = intval( $row->page_id );
 379                 ApiQueryBase::addTitleInfo( $a, Title::makeTitle( $row->page_namespace, $row->page_title ) );
 380                 if ( $row->page_is_redirect ) {
 381                         $a['redirect'] = '';
 382                 }
 383                 $ns = $this->hasNS ? $row->{$this->bl_ns} : NS_FILE;
 384                 $parentID = $this->pageMap[$ns][$row->{$this->bl_title}];
 385                 // Put all the results in an array first
 386                 $this->resultArr[$parentID]['redirlinks'][] = $a;
 387                 $this->getResult()->setIndexedTagName( $this->resultArr[$parentID]['redirlinks'], $this->bl_code );
 388         }
 389
 390         protected function processContinue() {
 391                 if ( !is_null( $this->params['continue'] ) ) {
 392                         $this->parseContinueParam();
 393                 } else {
 394                         $this->rootTitle = $this->getTitleOrPageId( $this->params )->getTitle();
 395                 }
 396
 397                 // only image titles are allowed for the root in imageinfo mode
 398                 if ( !$this->hasNS && $this->rootTitle->getNamespace() !== NS_FILE ) {
 399                         $this->dieUsage( "The title for {$this->getModuleName()} query must be an image", 'bad_image_title' );
 400                 }
 401         }
 402
 403         protected function parseContinueParam() {
 404                 $continueList = explode( '|', $this->params['continue'] );
 405                 // expected format:
 406                 // ns | key | id1 [| id2]
 407                 // ns+key: root title
 408                 // id1: first-level page ID to continue from
 409                 // id2: second-level page ID to continue from
 410
 411                 // null stuff out now so we know what's set and what isn't
 412                 $this->rootTitle = $this->contID = $this->redirID = null;
 413                 $rootNs = intval( $continueList[0] );
 414                 $this->dieContinueUsageIf( $rootNs === 0 && $continueList[0] !== '0' );
 415
 416                 $this->rootTitle = Title::makeTitleSafe( $rootNs, $continueList[1] );
 417                 $this->dieContinueUsageIf( !$this->rootTitle );
 418
 419                 $contID = intval( $continueList[2] );
 420                 $this->dieContinueUsageIf( $contID === 0 && $continueList[2] !== '0' );
 421
 422                 $this->contID = $contID;
 423                 $id2 = isset( $continueList[3] ) ? $continueList[3] : null;
 424                 $redirID = intval( $id2 );
 425
 426                 if ( $redirID === 0 && $id2 !== '0' ) {
 427                         // This one isn't required
 428                         return;
 429                 }
 430                 $this->redirID = $redirID;
 431
 432         }
 433
 434         protected function getContinueStr( $lastPageID ) {
 435                 return $this->rootTitle->getNamespace() .
 436                 '|' . $this->rootTitle->getDBkey() .
 437                 '|' . $lastPageID;
 438         }
 439
 440         protected function getContinueRedirStr( $lastPageID, $lastRedirID ) {
 441                 return $this->getContinueStr( $lastPageID ) . '|' . $lastRedirID;
 442         }
 443
 444         public function getAllowedParams() {
 445                 $retval = array(
 446                         'title' => array(
 447                                 ApiBase::PARAM_TYPE => 'string',
 448                         ),
 449                         'pageid' => array(
 450                                 ApiBase::PARAM_TYPE => 'integer',
 451                         ),
 452                         'continue' => null,
 453                         'namespace' => array(
 454                                 ApiBase::PARAM_ISMULTI => true,
 455                                 ApiBase::PARAM_TYPE => 'namespace'
 456                         ),
 457                         'dir' => array(
 458                                 ApiBase::PARAM_DFLT => 'ascending',
 459                                 ApiBase::PARAM_TYPE => array(
 460                                         'ascending',
 461                                         'descending'
 462                                 )
 463                         ),
 464                         'filterredir' => array(
 465                                 ApiBase::PARAM_DFLT => 'all',
 466                                 ApiBase::PARAM_TYPE => array(
 467                                         'all',
 468                                         'redirects',
 469                                         'nonredirects'
 470                                 )
 471                         ),
 472                         'limit' => array(
 473                                 ApiBase::PARAM_DFLT => 10,
 474                                 ApiBase::PARAM_TYPE => 'limit',
 475                                 ApiBase::PARAM_MIN => 1,
 476                                 ApiBase::PARAM_MAX => ApiBase::LIMIT_BIG1,
 477                                 ApiBase::PARAM_MAX2 => ApiBase::LIMIT_BIG2
 478                         )
 479                 );
 480                 if ( $this->getModuleName() == 'embeddedin' ) {
 481                         return $retval;
 482                 }
 483                 $retval['redirect'] = false;
 484                 return $retval;
 485         }
 486
 487         public function getParamDescription() {
 488                 $retval = array(
 489                         'title' => "Title to search. Cannot be used together with {$this->bl_code}pageid",
 490                         'pageid' => "Pageid to search. Cannot be used together with {$this->bl_code}title",
 491                         'continue' => 'When more results are available, use this to continue',
 492                         'namespace' => 'The namespace to enumerate',
 493                         'dir' => 'The direction in which to list',
 494                 );
 495                 if ( $this->getModuleName() != 'embeddedin' ) {
 496                         return array_merge( $retval, array(
 497                                 'redirect' => 'If linking page is a redirect, find all pages that link to that redirect as well. Maximum limit is halved.',
 498                                 'filterredir' => "How to filter for redirects. If set to nonredirects when {$this->bl_code}redirect is enabled, this is only applied to the second level",
 499                                 '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)."
 500                         ) );
 501                 }
 502                 return array_merge( $retval, array(
 503                         'filterredir' => 'How to filter for redirects',
 504                         'limit' => 'How many total pages to return'
 505                 ) );
 506         }
 507
 508         public function getResultProperties() {
 509                 return array(
 510                         '' => array(
 511                                 'pageid' => 'integer',
 512                                 'ns' => 'namespace',
 513                                 'title' => 'string',
 514                                 'redirect' => 'boolean'
 515                         )
 516                 );
 517         }
 518
 519         public function getDescription() {
 520                 switch ( $this->getModuleName() ) {
 521                         case 'backlinks':
 522                                 return 'Find all pages that link to the given page';
 523                         case 'embeddedin':
 524                                 return 'Find all pages that embed (transclude) the given title';
 525                         case 'imageusage':
 526                                 return 'Find all pages that use the given image title.';
 527                         default:
 528                                 ApiBase::dieDebug( __METHOD__, 'Unknown module name' );
 529                 }
 530         }
 531
 532         public function getPossibleErrors() {
 533                 return array_merge( parent::getPossibleErrors(),
 534                         $this->getTitleOrPageIdErrorMessage(),
 535                         array(
 536                                 array( 'code' => 'bad_image_title', 'info' => "The title for {$this->getModuleName()} query must be an image" ),
 537                         )
 538                 );
 539         }
 540
 541         public function getExamples() {
 542                 static $examples = array(
 543                         'backlinks' => array(
 544                                 'api.php?action=query&list=backlinks&bltitle=Main%20Page',
 545                                 'api.php?action=query&generator=backlinks&gbltitle=Main%20Page&prop=info'
 546                         ),
 547                         'embeddedin' => array(
 548                                 'api.php?action=query&list=embeddedin&eititle=Template:Stub',
 549                                 'api.php?action=query&generator=embeddedin&geititle=Template:Stub&prop=info'
 550                         ),
 551                         'imageusage' => array(
 552                                 'api.php?action=query&list=imageusage&iutitle=File:Albert%20Einstein%20Head.jpg',
 553                                 'api.php?action=query&generator=imageusage&giutitle=File:Albert%20Einstein%20Head.jpg&prop=info'
 554                         )
 555                 );
 556
 557                 return $examples[$this->getModuleName()];
 558         }
 559
 560         public function getHelpUrls() {
 561                 return $this->helpUrl;
 562         }
 563 }