includes/api/ApiQueryBacklinks.php

   1 <?php
   2
   3 /*
   4  * Created on Oct 16, 2006
   5  *
   6  * API for MediaWiki 1.8+
   7  *
   8  * Copyright (C) 2006 Yuri Astrakhan <Firstname><Lastname>@gmail.com
   9  *
  10  * This program is free software; you can redistribute it and/or modify
  11  * it under the terms of the GNU General Public License as published by
  12  * the Free Software Foundation; either version 2 of the License, or
  13  * (at your option) any later version.
  14  *
  15  * This program is distributed in the hope that it will be useful,
  16  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  17  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
  18  * GNU General Public License for more details.
  19  *
  20  * You should have received a copy of the GNU General Public License along
  21  * with this program; if not, write to the Free Software Foundation, Inc.,
  22  * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
  23  * http://www.gnu.org/copyleft/gpl.html
  24  */
  25
  26 if ( !defined( 'MEDIAWIKI' ) ) {
  27         // Eclipse helper - will be ignored in production
  28         require_once( "ApiQueryBase.php" );
  29 }
  30
  31 /**
  32  * This is a three-in-one module to query:
  33  *   * backlinks  - links pointing to the given page,
  34  *   * embeddedin - what pages transclude the given page within themselves,
  35  *   * imageusage - what pages use the given image
  36  *
  37  * @ingroup API
  38  */
  39 class ApiQueryBacklinks extends ApiQueryGeneratorBase {
  40
  41         private $params, $rootTitle, $contID, $redirID, $redirect;
  42         private $bl_ns, $bl_from, $bl_table, $bl_code, $bl_title, $bl_sort, $bl_fields, $hasNS;
  43         private $pageMap, $resultArr;
  44
  45         // output element name, database column field prefix, database table
  46         private $backlinksSettings = array(
  47                 'backlinks' => array(
  48                         'code' => 'bl',
  49                         'prefix' => 'pl',
  50                         'linktbl' => 'pagelinks'
  51                 ),
  52                 'embeddedin' => array(
  53                         'code' => 'ei',
  54                         'prefix' => 'tl',
  55                         'linktbl' => 'templatelinks'
  56                 ),
  57                 'imageusage' => array(
  58                         'code' => 'iu',
  59                         'prefix' => 'il',
  60                         'linktbl' => 'imagelinks'
  61                 )
  62         );
  63
  64         public function __construct( $query, $moduleName ) {
  65                 extract( $this->backlinksSettings[$moduleName] );
  66                 $this->resultArr = array();
  67
  68                 parent::__construct( $query, $moduleName, $code );
  69                 $this->bl_ns = $prefix . '_namespace';
  70                 $this->bl_from = $prefix . '_from';
  71                 $this->bl_table = $linktbl;
  72                 $this->bl_code = $code;
  73
  74                 $this->hasNS = $moduleName !== 'imageusage';
  75                 if ( $this->hasNS ) {
  76                         $this->bl_title = $prefix . '_title';
  77                         $this->bl_sort = "{$this->bl_ns}, {$this->bl_title}, {$this->bl_from}";
  78                         $this->bl_fields = array(
  79                                 $this->bl_ns,
  80                                 $this->bl_title
  81                         );
  82                 } else {
  83                         $this->bl_title = $prefix . '_to';
  84                         $this->bl_sort = "{$this->bl_title}, {$this->bl_from}";
  85                         $this->bl_fields = array(
  86                                 $this->bl_title
  87                         );
  88                 }
  89         }
  90
  91         public function execute() {
  92                 $this->run();
  93         }
  94
  95         public function executeGenerator( $resultPageSet ) {
  96                 $this->run( $resultPageSet );
  97         }
  98
  99         private function prepareFirstQuery( $resultPageSet = null ) {
 100                 /* SELECT page_id, page_title, page_namespace, page_is_redirect
 101                  * FROM pagelinks, page WHERE pl_from=page_id
 102                  * AND pl_title='Foo' AND pl_namespace=0
 103                  * LIMIT 11 ORDER BY pl_from
 104                  */
 105                 $db = $this->getDB();
 106                 $this->addTables( array( $this->bl_table, 'page' ) );
 107                 $this->addWhere( "{$this->bl_from}=page_id" );
 108                 if ( is_null( $resultPageSet ) ) {
 109                         $this->addFields( array( 'page_id', 'page_title', 'page_namespace' ) );
 110                 } else {
 111                         $this->addFields( $resultPageSet->getPageTableFields() );
 112                 }
 113
 114                 $this->addFields( 'page_is_redirect' );
 115                 $this->addWhereFld( $this->bl_title, $this->rootTitle->getDBkey() );
 116
 117                 if ( $this->hasNS ) {
 118                         $this->addWhereFld( $this->bl_ns, $this->rootTitle->getNamespace() );
 119                 }
 120                 $this->addWhereFld( 'page_namespace', $this->params['namespace'] );
 121
 122                 if ( !is_null( $this->contID ) ) {
 123                         $this->addWhere( "{$this->bl_from}>={$this->contID}" );
 124                 }
 125
 126                 if ( $this->params['filterredir'] == 'redirects' ) {
 127                         $this->addWhereFld( 'page_is_redirect', 1 );
 128                 } elseif ( $this->params['filterredir'] == 'nonredirects' && !$this->redirect ) {
 129                         // bug 22245 - Check for !redirect, as filtering nonredirects, when getting what links to them is contradictory
 130                         $this->addWhereFld( 'page_is_redirect', 0 );
 131                 }
 132
 133                 $this->addOption( 'LIMIT', $this->params['limit'] + 1 );
 134                 $this->addOption( 'ORDER BY', $this->bl_from );
 135                 $this->addOption( 'STRAIGHT_JOIN' );
 136         }
 137
 138         private function prepareSecondQuery( $resultPageSet = null ) {
 139                 /* SELECT page_id, page_title, page_namespace, page_is_redirect, pl_title, pl_namespace
 140                    FROM pagelinks, page WHERE pl_from=page_id
 141                    AND (pl_title='Foo' AND pl_namespace=0) OR (pl_title='Bar' AND pl_namespace=1)
 142                    ORDER BY pl_namespace, pl_title, pl_from LIMIT 11
 143                  */
 144                 $db = $this->getDB();
 145                 $this->addTables( array( 'page', $this->bl_table ) );
 146                 $this->addWhere( "{$this->bl_from}=page_id" );
 147
 148                 if ( is_null( $resultPageSet ) ) {
 149                         $this->addFields( array( 'page_id', 'page_title', 'page_namespace', 'page_is_redirect' ) );
 150                 } else {
 151                         $this->addFields( $resultPageSet->getPageTableFields() );
 152                 }
 153
 154                 $this->addFields( $this->bl_title );
 155                 if ( $this->hasNS ) {
 156                         $this->addFields( $this->bl_ns );
 157                 }
 158
 159                 // We can't use LinkBatch here because $this->hasNS may be false
 160                 $titleWhere = array();
 161                 foreach ( $this->redirTitles as $t ) {
 162                         $titleWhere[] = "{$this->bl_title} = " . $db->addQuotes( $t->getDBkey() ) .
 163                                         ( $this->hasNS ? " AND {$this->bl_ns} = '{$t->getNamespace()}'" : '' );
 164                 }
 165                 $this->addWhere( $db->makeList( $titleWhere, LIST_OR ) );
 166                 $this->addWhereFld( 'page_namespace', $this->params['namespace'] );
 167
 168                 if ( !is_null( $this->redirID ) ) {
 169                         $first = $this->redirTitles[0];
 170                         $title = $db->strencode( $first->getDBkey() );
 171                         $ns = $first->getNamespace();
 172                         $from = $this->redirID;
 173                         if ( $this->hasNS ) {
 174                                 $this->addWhere( "{$this->bl_ns} > $ns OR " .
 175                                                 "({$this->bl_ns} = $ns AND " .
 176                                                 "({$this->bl_title} > '$title' OR " .
 177                                                 "({$this->bl_title} = '$title' AND " .
 178                                                 "{$this->bl_from} >= $from)))" );
 179                         } else {
 180                                 $this->addWhere( "{$this->bl_title} > '$title' OR " .
 181                                                 "({$this->bl_title} = '$title' AND " .
 182                                                 "{$this->bl_from} >= $from)" );
 183                         }
 184                 }
 185                 if ( $this->params['filterredir'] == 'redirects' ) {
 186                         $this->addWhereFld( 'page_is_redirect', 1 );
 187                 } elseif ( $this->params['filterredir'] == 'nonredirects' ) {
 188                         $this->addWhereFld( 'page_is_redirect', 0 );
 189                 }
 190
 191                 $this->addOption( 'LIMIT', $this->params['limit'] + 1 );
 192                 $this->addOption( 'ORDER BY', $this->bl_sort );
 193                 $this->addOption( 'USE INDEX', array( 'page' => 'PRIMARY' ) );
 194         }
 195
 196         private function run( $resultPageSet = null ) {
 197                 $this->params = $this->extractRequestParams( false );
 198                 $this->redirect = isset( $this->params['redirect'] ) && $this->params['redirect'];
 199                 $userMax = ( $this->redirect ? ApiBase::LIMIT_BIG1 / 2 : ApiBase::LIMIT_BIG1 );
 200                 $botMax  = ( $this->redirect ? ApiBase::LIMIT_BIG2 / 2 : ApiBase::LIMIT_BIG2 );
 201                 if ( $this->params['limit'] == 'max' ) {
 202                         $this->params['limit'] = $this->getMain()->canApiHighLimits() ? $botMax : $userMax;
 203                         $this->getResult()->addValue( 'limits', $this->getModuleName(), $this->params['limit'] );
 204                 }
 205
 206                 $this->processContinue();
 207                 $this->prepareFirstQuery( $resultPageSet );
 208
 209                 $db = $this->getDB();
 210                 $res = $this->select( __METHOD__ . '::firstQuery' );
 211
 212                 $count = 0;
 213                 $this->pageMap = array(); // Maps ns and title to pageid
 214                 $this->continueStr = null;
 215                 $this->redirTitles = array();
 216                 foreach ( $res as $row ) {
 217                         if ( ++ $count > $this->params['limit'] ) {
 218                                 // We've reached the one extra which shows that there are additional pages to be had. Stop here...
 219                                 // Continue string preserved in case the redirect query doesn't pass the limit
 220                                 $this->continueStr = $this->getContinueStr( $row->page_id );
 221                                 break;
 222                         }
 223
 224                         if ( is_null( $resultPageSet ) ) {
 225                                 $this->extractRowInfo( $row );
 226                         } else {
 227                                 $this->pageMap[$row->page_namespace][$row->page_title] = $row->page_id;
 228                                 if ( $row->page_is_redirect ) {
 229                                         $this->redirTitles[] = Title::makeTitle( $row->page_namespace, $row->page_title );
 230                                 }
 231
 232                                 $resultPageSet->processDbRow( $row );
 233                         }
 234                 }
 235
 236                 if ( $this->redirect && count( $this->redirTitles ) ) {
 237                         $this->resetQueryParams();
 238                         $this->prepareSecondQuery( $resultPageSet );
 239                         $res = $this->select( __METHOD__ . '::secondQuery' );
 240                         $count = 0;
 241                         foreach ( $res as $row ) {
 242                                 if ( ++$count > $this->params['limit'] ) {
 243                                         // We've reached the one extra which shows that there are additional pages to be had. Stop here...
 244                                         // We need to keep the parent page of this redir in
 245                                         if ( $this->hasNS ) {
 246                                                 $parentID = $this->pageMap[$row-> { $this->bl_ns } ][$row-> { $this->bl_title } ];
 247                                         } else {
 248                                                 $parentID = $this->pageMap[NS_IMAGE][$row-> { $this->bl_title } ];
 249                                         }
 250                                         $this->continueStr = $this->getContinueRedirStr( $parentID, $row->page_id );
 251                                         break;
 252                                 }
 253
 254                                 if ( is_null( $resultPageSet ) ) {
 255                                         $this->extractRedirRowInfo( $row );
 256                                 } else {
 257                                         $resultPageSet->processDbRow( $row );
 258                                 }
 259                         }
 260                 }
 261                 if ( is_null( $resultPageSet ) ) {
 262                         // Try to add the result data in one go and pray that it fits
 263                         $fit = $this->getResult()->addValue( 'query', $this->getModuleName(), array_values( $this->resultArr ) );
 264                         if ( !$fit ) {
 265                                 // It didn't fit. Add elements one by one until the
 266                                 // result is full.
 267                                 foreach ( $this->resultArr as $pageID => $arr ) {
 268                                         // Add the basic entry without redirlinks first
 269                                         $fit = $this->getResult()->addValue(
 270                                                 array( 'query', $this->getModuleName() ),
 271                                                 null, array_diff_key( $arr, array( 'redirlinks' => '' ) ) );
 272                                         if ( !$fit ) {
 273                                                 $this->continueStr = $this->getContinueStr( $pageID );
 274                                                 break;
 275                                         }
 276
 277                                         $hasRedirs = false;
 278                                         foreach ( (array)@$arr['redirlinks'] as $key => $redir ) {
 279                                                 $fit = $this->getResult()->addValue(
 280                                                         array( 'query', $this->getModuleName(), $pageID, 'redirlinks' ),
 281                                                         $key, $redir );
 282                                                 if ( !$fit ) {
 283                                                         $this->continueStr = $this->getContinueRedirStr( $pageID, $redir['pageid'] );
 284                                                         break;
 285                                                 }
 286                                                 $hasRedirs = true;
 287                                         }
 288                                         if ( $hasRedirs ) {
 289                                                 $this->getResult()->setIndexedTagName_internal(
 290                                                         array( 'query', $this->getModuleName(), $pageID, 'redirlinks' ),
 291                                                         $this->bl_code );
 292                                         }
 293                                         if ( !$fit ) {
 294                                                 break;
 295                                         }
 296                                 }
 297                         }
 298
 299                         $this->getResult()->setIndexedTagName_internal(
 300                                 array( 'query', $this->getModuleName() ),
 301                                 $this->bl_code
 302                         );
 303                 }
 304                 if ( !is_null( $this->continueStr ) ) {
 305                         $this->setContinueEnumParameter( 'continue', $this->continueStr );
 306                 }
 307         }
 308
 309         private function extractRowInfo( $row ) {
 310                 $this->pageMap[$row->page_namespace][$row->page_title] = $row->page_id;
 311                 $t = Title::makeTitle( $row->page_namespace, $row->page_title );
 312                 $a = array( 'pageid' => intval( $row->page_id ) );
 313                 ApiQueryBase::addTitleInfo( $a, $t );
 314                 if ( $row->page_is_redirect ) {
 315                         $a['redirect'] = '';
 316                         $this->redirTitles[] = $t;
 317                 }
 318                 // Put all the results in an array first
 319                 $this->resultArr[$a['pageid']] = $a;
 320         }
 321
 322         private function extractRedirRowInfo( $row ) {
 323                 $a['pageid'] = intval( $row->page_id );
 324                 ApiQueryBase::addTitleInfo( $a, Title::makeTitle( $row->page_namespace, $row->page_title ) );
 325                 if ( $row->page_is_redirect ) {
 326                         $a['redirect'] = '';
 327                 }
 328                 $ns = $this->hasNS ? $row-> { $this->bl_ns } : NS_FILE;
 329                 $parentID = $this->pageMap[$ns][$row-> { $this->bl_title } ];
 330                 // Put all the results in an array first
 331                 $this->resultArr[$parentID]['redirlinks'][] = $a;
 332                 $this->getResult()->setIndexedTagName( $this->resultArr[$parentID]['redirlinks'], $this->bl_code );
 333         }
 334
 335         protected function processContinue() {
 336                 if ( !is_null( $this->params['continue'] ) ) {
 337                         $this->parseContinueParam();
 338                 } else {
 339                         if ( $this->params['title'] !== '' ) {
 340                                 $title = Title::newFromText( $this->params['title'] );
 341                                 if ( !$title ) {
 342                                         $this->dieUsageMsg( array( 'invalidtitle', $this->params['title'] ) );
 343                                 } else {
 344                                         $this->rootTitle = $title;
 345                                 }
 346                         } else {
 347                                 $this->dieUsageMsg( array( 'missingparam', 'title' ) );
 348                         }
 349                 }
 350
 351                 // only image titles are allowed for the root in imageinfo mode
 352                 if ( !$this->hasNS && $this->rootTitle->getNamespace() !== NS_FILE ) {
 353                         $this->dieUsage( "The title for {$this->getModuleName()} query must be an image", 'bad_image_title' );
 354                 }
 355         }
 356
 357         protected function parseContinueParam() {
 358                 $continueList = explode( '|', $this->params['continue'] );
 359                 // expected format:
 360                 // ns | key | id1 [| id2]
 361                 // ns+key: root title
 362                 // id1: first-level page ID to continue from
 363                 // id2: second-level page ID to continue from
 364
 365                 // null stuff out now so we know what's set and what isn't
 366                 $this->rootTitle = $this->contID = $this->redirID = null;
 367                 $rootNs = intval( $continueList[0] );
 368                 if ( $rootNs === 0 && $continueList[0] !== '0' ) {
 369                         // Illegal continue parameter
 370                         $this->dieUsage( 'Invalid continue param. You should pass the original value returned by the previous query', '_badcontinue' );
 371                 }
 372                 $this->rootTitle = Title::makeTitleSafe( $rootNs, $continueList[1] );
 373
 374                 if ( !$this->rootTitle ) {
 375                         $this->dieUsage( 'Invalid continue param. You should pass the original value returned by the previous query', '_badcontinue' );
 376                 }
 377                 $contID = intval( $continueList[2] );
 378
 379                 if ( $contID === 0 && $continueList[2] !== '0' ) {
 380                         $this->dieUsage( 'Invalid continue param. You should pass the original value returned by the previous query', '_badcontinue' );
 381                 }
 382                 $this->contID = $contID;
 383                 $redirID = intval( @$continueList[3] );
 384
 385                 if ( $redirID === 0 && @$continueList[3] !== '0' ) {
 386                         // This one isn't required
 387                         return;
 388                 }
 389                 $this->redirID = $redirID;
 390
 391         }
 392
 393         protected function getContinueStr( $lastPageID ) {
 394                 return $this->rootTitle->getNamespace() .
 395                 '|' . $this->rootTitle->getDBkey() .
 396                 '|' . $lastPageID;
 397         }
 398
 399         protected function getContinueRedirStr( $lastPageID, $lastRedirID ) {
 400                 return $this->getContinueStr( $lastPageID ) . '|' . $lastRedirID;
 401         }
 402
 403         public function getAllowedParams() {
 404                 $retval = array(
 405                         'title' => null,
 406                         'continue' => null,
 407                         'namespace' => array(
 408                                 ApiBase::PARAM_ISMULTI => true,
 409                                 ApiBase::PARAM_TYPE => 'namespace'
 410                         ),
 411                         'filterredir' => array(
 412                                 ApiBase::PARAM_DFLT => 'all',
 413                                 ApiBase::PARAM_TYPE => array(
 414                                         'all',
 415                                         'redirects',
 416                                         'nonredirects'
 417                                 )
 418                         ),
 419                         'limit' => array(
 420                                 ApiBase::PARAM_DFLT => 10,
 421                                 ApiBase::PARAM_TYPE => 'limit',
 422                                 ApiBase::PARAM_MIN => 1,
 423                                 ApiBase::PARAM_MAX => ApiBase::LIMIT_BIG1,
 424                                 ApiBase::PARAM_MAX2 => ApiBase::LIMIT_BIG2
 425                         )
 426                 );
 427                 if ( $this->getModuleName() == 'embeddedin' ) {
 428                         return $retval;
 429                 }
 430                 $retval['redirect'] = false;
 431                 return $retval;
 432         }
 433
 434         public function getParamDescription() {
 435                 $retval = array(
 436                         'title' => 'Title to search',
 437                         'continue' => 'When more results are available, use this to continue',
 438                         'namespace' => 'The namespace to enumerate',
 439                 );
 440                 if ( $this->getModuleName() != 'embeddedin' ) {
 441                         return array_merge( $retval, array(
 442                                 'redirect' => 'If linking page is a redirect, find all pages that link to that redirect as well. Maximum limit is halved.',
 443                                 'filterredir' => "How to filter for redirects. If set to nonredirects when {$this->bl_code}redirect is enabled, this is only applied to the second level",
 444                                 '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)."
 445                         ) );
 446                 }
 447                 return array_merge( $retval, array(
 448                         'filterredir' => 'How to filter for redirects',
 449                         'limit' => 'How many total pages to return'
 450                 ) );
 451         }
 452
 453         public function getDescription() {
 454                 switch ( $this->getModuleName() ) {
 455                         case 'backlinks':
 456                                 return 'Find all pages that link to the given page';
 457                         case 'embeddedin':
 458                                 return 'Find all pages that embed (transclude) the given title';
 459                         case 'imageusage':
 460                                 return 'Find all pages that use the given image title.';
 461                         default:
 462                                 ApiBase::dieDebug( __METHOD__, 'Unknown module name' );
 463                 }
 464         }
 465
 466         public function getPossibleErrors() {
 467                 return array_merge( parent::getPossibleErrors(), array(
 468                         array( 'invalidtitle', 'title' ),
 469                         array( 'missingparam', 'title' ),
 470                         array( 'code' => 'bad_image_title', 'info' => "The title for {$this->getModuleName()} query must be an image" ),
 471                         array( 'code' => '_badcontinue', 'info' => 'Invalid continue param. You should pass the original value returned by the previous query' ),
 472                 ) );
 473         }
 474
 475         protected function getExamples() {
 476                 static $examples = array(
 477                         'backlinks' => array(
 478                                 'api.php?action=query&list=backlinks&bltitle=Main%20Page',
 479                                 'api.php?action=query&generator=backlinks&gbltitle=Main%20Page&prop=info'
 480                         ),
 481                         'embeddedin' => array(
 482                                 'api.php?action=query&list=embeddedin&eititle=Template:Stub',
 483                                 'api.php?action=query&generator=embeddedin&geititle=Template:Stub&prop=info'
 484                         ),
 485                         'imageusage' => array(
 486                                 'api.php?action=query&list=imageusage&iutitle=File:Albert%20Einstein%20Head.jpg',
 487                                 'api.php?action=query&generator=imageusage&giutitle=File:Albert%20Einstein%20Head.jpg&prop=info'
 488                         )
 489                 );
 490
 491                 return $examples[$this->getModuleName()];
 492         }
 493
 494         public function getVersion() {
 495                 return __CLASS__ . ': $Id$';
 496         }
 497 }