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