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  * 59 Temple Place - Suite 330, Boston, MA 02111-1307, 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, $contRedirs, $contLevel, $contTitle, $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('page', $this->bl_table));
 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                 $this->addFields('page_is_redirect');
 113                 $this->addWhereFld($this->bl_title, $this->rootTitle->getDBKey());
 114                 if($this->hasNS)
 115                         $this->addWhereFld($this->bl_ns, $this->rootTitle->getNamespace());
 116                 $this->addWhereFld('page_namespace', $this->params['namespace']);
 117                 if(!is_null($this->contID))
 118                         $this->addWhere("{$this->bl_from}>={$this->contID}");
 119                 if($this->params['filterredir'] == 'redirects')
 120                         $this->addWhereFld('page_is_redirect', 1);
 121                 if($this->params['filterredir'] == 'nonredirects')
 122                         $this->addWhereFld('page_is_redirect', 0);
 123                 $this->addOption('LIMIT', $this->params['limit'] + 1);
 124                 $this->addOption('ORDER BY', $this->bl_from);
 125         }
 126
 127         private function prepareSecondQuery($resultPageSet = null) {
 128                 /* SELECT page_id, page_title, page_namespace, page_is_redirect, pl_title, pl_namespace
 129                    FROM pagelinks, page WHERE pl_from=page_id
 130                    AND (pl_title='Foo' AND pl_namespace=0) OR (pl_title='Bar' AND pl_namespace=1)
 131                    ORDER BY pl_namespace, pl_title, pl_from LIMIT 11
 132                  */
 133                 $db = $this->getDB();
 134                 $this->addTables(array('page', $this->bl_table));
 135                 $this->addWhere("{$this->bl_from}=page_id");
 136                 if(is_null($resultPageSet))
 137                         $this->addFields(array('page_id', 'page_title', 'page_namespace', 'page_is_redirect'));
 138                 else
 139                         $this->addFields($resultPageSet->getPageTableFields());
 140                 $this->addFields($this->bl_title);
 141                 if($this->hasNS)
 142                         $this->addFields($this->bl_ns);
 143                 // We can't use LinkBatch here because $this->hasNS may be false
 144                 $titleWhere = array();
 145                 foreach($this->redirTitles as $t)
 146                         $titleWhere[] = "{$this->bl_title} = ".$db->addQuotes($t->getDBKey()).
 147                                         ($this->hasNS ? " AND {$this->bl_ns} = '{$t->getNamespace()}'" : "");
 148                 $this->addWhere($db->makeList($titleWhere, LIST_OR));
 149                 $this->addWhereFld('page_namespace', $this->params['namespace']);
 150                 if(!is_null($this->redirID))
 151                 {
 152                         $first = $this->redirTitles[0];
 153                         $title = $db->strencode($first->getDBKey());
 154                         $ns = $first->getNamespace();
 155                         $from = $this->redirID;
 156                         if($this->hasNS)
 157                                 $this->addWhere("{$this->bl_ns} > $ns OR ".
 158                                                 "({$this->bl_ns} = $ns AND ".
 159                                                 "({$this->bl_title} > '$title' OR ".
 160                                                 "({$this->bl_title} = '$title' AND ".
 161                                                 "{$this->bl_from} >= $from)))");
 162                         else
 163                                 $this->addWhere("{$this->bl_title} > '$title' OR ".
 164                                                 "({$this->bl_title} = '$title' AND ".
 165                                                 "{$this->bl_from} >= $from)");
 166
 167                 }
 168                 if($this->params['filterredir'] == 'redirects')
 169                         $this->addWhereFld('page_is_redirect', 1);
 170                 if($this->params['filterredir'] == 'nonredirects')
 171                         $this->addWhereFld('page_is_redirect', 0);
 172                 $this->addOption('LIMIT', $this->params['limit'] + 1);
 173                 $this->addOption('ORDER BY', $this->bl_sort);
 174                 $this->addOption('USE INDEX', array('page' => 'PRIMARY'));
 175         }
 176
 177         private function run($resultPageSet = null) {
 178                 $this->params = $this->extractRequestParams(false);
 179                 $this->redirect = isset($this->params['redirect']) && $this->params['redirect'];
 180                 $userMax = ( $this->redirect ? ApiBase::LIMIT_BIG1/2 : ApiBase::LIMIT_BIG1 );
 181                 $botMax  = ( $this->redirect ? ApiBase::LIMIT_BIG2/2 : ApiBase::LIMIT_BIG2 );
 182                 if( $this->params['limit'] == 'max' ) {
 183                         $this->params['limit'] = $this->getMain()->canApiHighLimits() ? $botMax : $userMax;
 184                         $this->getResult()->addValue( 'limits', $this->getModuleName(), $this->params['limit'] );
 185                 }
 186
 187                 $this->processContinue();
 188                 $this->prepareFirstQuery($resultPageSet);
 189
 190                 $db = $this->getDB();
 191                 $res = $this->select(__METHOD__.'::firstQuery');
 192
 193                 $count = 0;
 194                 $this->pageMap = array(); // Maps ns and title to pageid
 195                 $this->continueStr = null;
 196                 $this->redirTitles = array();
 197                 while ($row = $db->fetchObject($res)) {
 198                         if (++ $count > $this->params['limit']) {
 199                                 // We've reached the one extra which shows that there are additional pages to be had. Stop here...
 200                                 // Continue string preserved in case the redirect query doesn't pass the limit
 201                                 $this->continueStr = $this->getContinueStr($row->page_id);
 202                                 break;
 203                         }
 204
 205                         if (is_null($resultPageSet))
 206                                 $this->extractRowInfo($row);
 207                         else
 208                         {
 209                                 $this->pageMap[$row->page_namespace][$row->page_title] = $row->page_id;
 210                                 if($row->page_is_redirect)
 211                                         $this->redirTitles[] = Title::makeTitle($row->page_namespace, $row->page_title);
 212                                 $resultPageSet->processDbRow($row);
 213                         }
 214                 }
 215                 $db->freeResult($res);
 216
 217                 if($this->redirect && count($this->redirTitles))
 218                 {
 219                         $this->resetQueryParams();
 220                         $this->prepareSecondQuery($resultPageSet);
 221                         $res = $this->select(__METHOD__.'::secondQuery');
 222                         $count = 0;
 223                         while($row = $db->fetchObject($res))
 224                         {
 225                                 if(++$count > $this->params['limit'])
 226                                 {
 227                                         // We've reached the one extra which shows that there are additional pages to be had. Stop here...
 228                                         // We need to keep the parent page of this redir in
 229                                         if($this->hasNS)
 230                                                 $parentID = $this->pageMap[$row->{$this->bl_ns}][$row->{$this->bl_title}];
 231                                         else
 232                                                 $parentID = $this->pageMap[NS_IMAGE][$row->{$this->bl_title}];
 233                                         $this->continueStr = $this->getContinueRedirStr($parentID, $row->page_id);
 234                                         break;
 235                                 }
 236
 237                                 if(is_null($resultPageSet))
 238                                         $this->extractRedirRowInfo($row);
 239                                 else
 240                                         $resultPageSet->processDbRow($row);
 241                         }
 242                         $db->freeResult($res);
 243                 }
 244                 if (is_null($resultPageSet)) {
 245                         // Try to add the result data in one go and pray that it fits
 246                         $fit = $this->getResult()->addValue('query', $this->getModuleName(), array_values($this->resultArr));
 247                         if(!$fit)
 248                         {
 249                                 // It didn't fit. Add elements one by one until the
 250                                 // result is full.
 251                                 foreach($this->resultArr as $pageID => $arr)
 252                                 {
 253                                         // Add the basic entry without redirlinks first
 254                                         $fit = $this->getResult()->addValue(
 255                                                 array('query', $this->getModuleName()),
 256                                                 null, array_diff_key($arr, array('redirlinks' => '')));
 257                                         if(!$fit)
 258                                         {
 259                                                 $this->continueStr = $this->getContinueStr($pageID);
 260                                                 break;
 261                                         }
 262
 263                                         $hasRedirs = false;
 264                                         foreach((array)@$arr['redirlinks'] as $key => $redir)
 265                                         {
 266                                                 $fit = $this->getResult()->addValue(
 267                                                         array('query', $this->getModuleName(), $pageID, 'redirlinks'),
 268                                                         $key, $redir);
 269                                                 if(!$fit)
 270                                                 {
 271                                                         $this->continueStr = $this->getContinueRedirStr($pageID, $redir['pageid']);
 272                                                         break;
 273                                                 }
 274                                                 $hasRedirs = true;
 275                                         }
 276                                         if($hasRedirs)
 277                                                 $this->getResult()->setIndexedTagName_internal(
 278                                                         array('query', $this->getModuleName(), $pageID, 'redirlinks'),
 279                                                         $this->bl_code);
 280                                         if(!$fit)
 281                                                 break;
 282                                 }
 283                         }
 284
 285                         $this->getResult()->setIndexedTagName_internal(
 286                                          array('query', $this->getModuleName()),
 287                                          $this->bl_code);
 288                 }
 289                 if(!is_null($this->continueStr))
 290                         $this->setContinueEnumParameter('continue', $this->continueStr);
 291         }
 292
 293         private function extractRowInfo($row) {
 294                 $this->pageMap[$row->page_namespace][$row->page_title] = $row->page_id;
 295                 $t = Title::makeTitle($row->page_namespace, $row->page_title);
 296                 $a = array('pageid' => intval($row->page_id));
 297                 ApiQueryBase::addTitleInfo($a, $t);
 298                 if($row->page_is_redirect)
 299                 {
 300                         $a['redirect'] = '';
 301                         $this->redirTitles[] = $t;
 302                 }
 303                 // Put all the results in an array first
 304                 $this->resultArr[$a['pageid']] = $a;
 305         }
 306
 307         private function extractRedirRowInfo($row)
 308         {
 309                 $a['pageid'] = intval($row->page_id);
 310                 ApiQueryBase::addTitleInfo($a, Title::makeTitle($row->page_namespace, $row->page_title));
 311                 if($row->page_is_redirect)
 312                         $a['redirect'] = '';
 313                 $ns = $this->hasNS ? $row->{$this->bl_ns} : NS_FILE;
 314                 $parentID = $this->pageMap[$ns][$row->{$this->bl_title}];
 315                 // Put all the results in an array first
 316                 $this->resultArr[$parentID]['redirlinks'][] = $a;
 317                 $this->getResult()->setIndexedTagName($this->resultArr[$parentID]['redirlinks'], $this->bl_code);
 318         }
 319
 320         protected function processContinue() {
 321                 if (!is_null($this->params['continue']))
 322                         $this->parseContinueParam();
 323                 else {
 324                         if ( $this->params['title'] !== "" ) {
 325                                 $title = Title::newFromText( $this->params['title'] );
 326                                 if ( !$title ) {
 327                                         $this->dieUsageMsg(array('invalidtitle', $this->params['title']));
 328                                 } else {
 329                                         $this->rootTitle = $title;
 330                                 }
 331                         } else {
 332                                 $this->dieUsageMsg(array('missingparam', 'title'));
 333                         }
 334                 }
 335
 336                 // only image titles are allowed for the root in imageinfo mode
 337                 if (!$this->hasNS && $this->rootTitle->getNamespace() !== NS_FILE)
 338                         $this->dieUsage("The title for {$this->getModuleName()} query must be an image", 'bad_image_title');
 339         }
 340
 341         protected function parseContinueParam() {
 342                 $continueList = explode('|', $this->params['continue']);
 343                 // expected format:
 344                 // ns | key | id1 [| id2]
 345                 // ns+key: root title
 346                 // id1: first-level page ID to continue from
 347                 // id2: second-level page ID to continue from
 348
 349                 // null stuff out now so we know what's set and what isn't
 350                 $this->rootTitle = $this->contID = $this->redirID = null;
 351                 $rootNs = intval($continueList[0]);
 352                 if($rootNs === 0 && $continueList[0] !== '0')
 353                         // Illegal continue parameter
 354                         $this->dieUsage("Invalid continue param. You should pass the original value returned by the previous query", "_badcontinue");
 355                 $this->rootTitle = Title::makeTitleSafe($rootNs, $continueList[1]);
 356                 if(!$this->rootTitle)
 357                         $this->dieUsage("Invalid continue param. You should pass the original value returned by the previous query", "_badcontinue");
 358                 $contID = intval($continueList[2]);
 359                 if($contID === 0 && $continueList[2] !== '0')
 360                         $this->dieUsage("Invalid continue param. You should pass the original value returned by the previous query", "_badcontinue");
 361                 $this->contID = $contID;
 362                 $redirID = intval(@$continueList[3]);
 363                 if($redirID === 0 && @$continueList[3] !== '0')
 364                         // This one isn't required
 365                         return;
 366                 $this->redirID = $redirID;
 367
 368         }
 369
 370         protected function getContinueStr($lastPageID) {
 371                 return $this->rootTitle->getNamespace() .
 372                 '|' . $this->rootTitle->getDBkey() .
 373                 '|' . $lastPageID;
 374         }
 375
 376         protected function getContinueRedirStr($lastPageID, $lastRedirID) {
 377                 return $this->getContinueStr($lastPageID) . '|' . $lastRedirID;
 378         }
 379
 380         public function getAllowedParams() {
 381                 $retval =  array (
 382                         'title' => null,
 383                         'continue' => null,
 384                         'namespace' => array (
 385                                 ApiBase :: PARAM_ISMULTI => true,
 386                                 ApiBase :: PARAM_TYPE => 'namespace'
 387                         ),
 388                         'filterredir' => array(
 389                                 ApiBase :: PARAM_DFLT => 'all',
 390                                 ApiBase :: PARAM_TYPE => array(
 391                                         'all',
 392                                         'redirects',
 393                                         'nonredirects'
 394                                 )
 395                         ),
 396                         'limit' => array (
 397                                 ApiBase :: PARAM_DFLT => 10,
 398                                 ApiBase :: PARAM_TYPE => 'limit',
 399                                 ApiBase :: PARAM_MIN => 1,
 400                                 ApiBase :: PARAM_MAX => ApiBase :: LIMIT_BIG1,
 401                                 ApiBase :: PARAM_MAX2 => ApiBase :: LIMIT_BIG2
 402                         )
 403                 );
 404                 if($this->getModuleName() == 'embeddedin')
 405                         return $retval;
 406                 $retval['redirect'] = false;
 407                 return $retval;
 408         }
 409
 410         public function getParamDescription() {
 411                 $retval = array (
 412                         'title' => 'Title to search.',
 413                         'continue' => 'When more results are available, use this to continue.',
 414                         'namespace' => 'The namespace to enumerate.',
 415                         'filterredir' => 'How to filter for redirects'
 416                 );
 417                 if($this->getModuleName() != 'embeddedin')
 418                         return array_merge($retval, array(
 419                                 'redirect' => 'If linking page is a redirect, find all pages that link to that redirect as well. Maximum limit is halved.',
 420                                 'limit' => "How many total pages to return. If {$this->bl_code}redirect is enabled, limit applies to each level separately (which means you may get up to 2 * limit results)."
 421                         ));
 422                 return array_merge($retval, array(
 423                         'limit' => "How many total pages to return."
 424                 ));
 425         }
 426
 427         public function getDescription() {
 428                 switch ($this->getModuleName()) {
 429                         case 'backlinks' :
 430                                 return 'Find all pages that link to the given page';
 431                         case 'embeddedin' :
 432                                 return 'Find all pages that embed (transclude) the given title';
 433                         case 'imageusage' :
 434                                 return 'Find all pages that use the given image title.';
 435                         default :
 436                                 ApiBase :: dieDebug(__METHOD__, 'Unknown module name');
 437                 }
 438         }
 439
 440         protected function getExamples() {
 441                 static $examples = array (
 442                         'backlinks' => array (
 443                                 "api.php?action=query&list=backlinks&bltitle=Main%20Page",
 444                                 "api.php?action=query&generator=backlinks&gbltitle=Main%20Page&prop=info"
 445                         ),
 446                         'embeddedin' => array (
 447                                 "api.php?action=query&list=embeddedin&eititle=Template:Stub",
 448                                 "api.php?action=query&generator=embeddedin&geititle=Template:Stub&prop=info"
 449                         ),
 450                         'imageusage' => array (
 451                                 "api.php?action=query&list=imageusage&iutitle=File:Albert%20Einstein%20Head.jpg",
 452                                 "api.php?action=query&generator=imageusage&giutitle=File:Albert%20Einstein%20Head.jpg&prop=info"
 453                         )
 454                 );
 455
 456                 return $examples[$this->getModuleName()];
 457         }
 458
 459         public function getVersion() {
 460                 return __CLASS__ . ': $Id$';
 461         }
 462 }