includes/api/ApiQueryRevisions.php

   1 <?php
   2
   3 /*
   4  * Created on Sep 7, 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  * A query action to enumerate revisions of a given page, or show top revisions of multiple pages.
  33  * Various pieces of information may be shown - flags, comments, and the actual wiki markup of the rev.
  34  * In the enumeration mode, ranges of revisions may be requested and filtered.
  35  *
  36  * @ingroup API
  37  */
  38 class ApiQueryRevisions extends ApiQueryBase {
  39
  40         public function __construct($query, $moduleName) {
  41                 parent :: __construct($query, $moduleName, 'rv');
  42         }
  43
  44         private $fld_ids = false, $fld_flags = false, $fld_timestamp = false, $fld_size = false,
  45                         $fld_comment = false, $fld_user = false, $fld_content = false, $fld_tags = false;
  46
  47         protected function getTokenFunctions() {
  48                 // tokenname => function
  49                 // function prototype is func($pageid, $title, $rev)
  50                 // should return token or false
  51
  52                 // Don't call the hooks twice
  53                 if(isset($this->tokenFunctions))
  54                         return $this->tokenFunctions;
  55
  56                 // If we're in JSON callback mode, no tokens can be obtained
  57                 if(!is_null($this->getMain()->getRequest()->getVal('callback')))
  58                         return array();
  59
  60                 $this->tokenFunctions = array(
  61                         'rollback' => array( 'ApiQueryRevisions', 'getRollbackToken' )
  62                 );
  63                 wfRunHooks('APIQueryRevisionsTokens', array(&$this->tokenFunctions));
  64                 return $this->tokenFunctions;
  65         }
  66
  67         public static function getRollbackToken($pageid, $title, $rev)
  68         {
  69                 global $wgUser;
  70                 if(!$wgUser->isAllowed('rollback'))
  71                         return false;
  72                 return $wgUser->editToken(array($title->getPrefixedText(),
  73                                                 $rev->getUserText()));
  74         }
  75
  76         public function execute() {
  77                 $params = $this->extractRequestParams(false);
  78
  79                 // If any of those parameters are used, work in 'enumeration' mode.
  80                 // Enum mode can only be used when exactly one page is provided.
  81                 // Enumerating revisions on multiple pages make it extremely
  82                 // difficult to manage continuations and require additional SQL indexes
  83                 $enumRevMode = (!is_null($params['user']) || !is_null($params['excludeuser']) ||
  84                                 !is_null($params['limit']) || !is_null($params['startid']) ||
  85                                 !is_null($params['endid']) || $params['dir'] === 'newer' ||
  86                                 !is_null($params['start']) || !is_null($params['end']));
  87
  88
  89                 $pageSet = $this->getPageSet();
  90                 $pageCount = $pageSet->getGoodTitleCount();
  91                 $revCount = $pageSet->getRevisionCount();
  92
  93                 // Optimization -- nothing to do
  94                 if ($revCount === 0 && $pageCount === 0)
  95                         return;
  96
  97                 if ($revCount > 0 && $enumRevMode)
  98                         $this->dieUsage('The revids= parameter may not be used with the list options (limit, startid, endid, dirNewer, start, end).', 'revids');
  99
 100                 if ($pageCount > 1 && $enumRevMode)
 101                         $this->dieUsage('titles, pageids or a generator was used to supply multiple pages, but the limit, startid, endid, dirNewer, user, excludeuser, start and end parameters may only be used on a single page.', 'multpages');
 102
 103                 if (!is_null($params['diffto'])) {
 104                         if ($params['diffto'] == 'cur')
 105                                 $params['diffto'] = 0;
 106                         if ((!ctype_digit($params['diffto']) || $params['diffto'] < 0)
 107                                         && $params['diffto'] != 'prev' && $params['diffto'] != 'next')
 108                                 $this->dieUsage('rvdiffto must be set to a non-negative number, "prev", "next" or "cur"', 'diffto');
 109                         // Check whether the revision exists and is readable,
 110                         // DifferenceEngine returns a rather ambiguous empty
 111                         // string if that's not the case
 112                         if ($params['diffto'] != 0) {
 113                                 $difftoRev = Revision::newFromID($params['diffto']);
 114                                 if (!$difftoRev)
 115                                         $this->dieUsageMsg(array('nosuchrevid', $params['diffto']));
 116                                 if (!$difftoRev->userCan(Revision::DELETED_TEXT)) {
 117                                         $this->setWarning("Couldn't diff to r{$difftoRev->getID()}: content is hidden");
 118                                         $params['diffto'] = null;
 119                                 }
 120                         }
 121                 }
 122
 123                 $db = $this->getDB();
 124                 $this->addTables(array('page', 'revision'));
 125                 $this->addFields(Revision::selectFields());
 126                 $this->addWhere('page_id = rev_page');
 127
 128                 $prop = array_flip($params['prop']);
 129
 130                 // Optional fields
 131                 $this->fld_ids = isset ($prop['ids']);
 132                 // $this->addFieldsIf('rev_text_id', $this->fld_ids); // should this be exposed?
 133                 $this->fld_flags = isset ($prop['flags']);
 134                 $this->fld_timestamp = isset ($prop['timestamp']);
 135                 $this->fld_comment = isset ($prop['comment']);
 136                 $this->fld_size = isset ($prop['size']);
 137                 $this->fld_tags = isset ($prop['tags']);
 138                 $this->fld_user = isset ($prop['user']);
 139                 $this->token = $params['token'];
 140                 $this->diffto = $params['diffto'];
 141
 142                 if ( !is_null($this->token) || $pageCount > 0) {
 143                         $this->addFields( Revision::selectPageFields() );
 144                 }
 145
 146                 if ($this->fld_tags || !is_null($params['tag'])) {
 147                         $this->addTables('tag_summary');
 148                         $this->addJoinConds(array('tag_summary' => array('LEFT JOIN', array('rev_id=ts_rev_id'))));
 149                         $this->addFields('ts_tags');
 150                 }
 151
 152                 if( !is_null($params['tag']) ) {
 153                         $this->addWhereFld('ts_tags', $params['tag']);
 154                 }
 155
 156                 if (isset ($prop['content'])) {
 157
 158                         // For each page we will request, the user must have read rights for that page
 159                         foreach ($pageSet->getGoodTitles() as $title) {
 160                                 if( !$title->userCanRead() )
 161                                         $this->dieUsage(
 162                                                 'The current user is not allowed to read ' . $title->getPrefixedText(),
 163                                                 'accessdenied');
 164                         }
 165
 166                         $this->addTables('text');
 167                         $this->addWhere('rev_text_id=old_id');
 168                         $this->addFields('old_id');
 169                         $this->addFields(Revision::selectTextFields());
 170
 171                         $this->fld_content = true;
 172
 173                         $this->expandTemplates = $params['expandtemplates'];
 174                         $this->generateXML = $params['generatexml'];
 175                         if(isset($params['section']))
 176                                 $this->section = $params['section'];
 177                         else
 178                                 $this->section = false;
 179                 }
 180
 181                 $userMax = ( $this->fld_content ? ApiBase::LIMIT_SML1 : ApiBase::LIMIT_BIG1 );
 182                 $botMax  = ( $this->fld_content ? ApiBase::LIMIT_SML2 : ApiBase::LIMIT_BIG2 );
 183                 $limit = $params['limit'];
 184                 if( $limit == 'max' ) {
 185                         $limit = $this->getMain()->canApiHighLimits() ? $botMax : $userMax;
 186                         $this->getResult()->addValue( 'limits', $this->getModuleName(), $limit );
 187                 }
 188
 189                 if ($enumRevMode) {
 190
 191                         // This is mostly to prevent parameter errors (and optimize SQL?)
 192                         if (!is_null($params['startid']) && !is_null($params['start']))
 193                                 $this->dieUsage('start and startid cannot be used together', 'badparams');
 194
 195                         if (!is_null($params['endid']) && !is_null($params['end']))
 196                                 $this->dieUsage('end and endid cannot be used together', 'badparams');
 197
 198                         if(!is_null($params['user']) && !is_null($params['excludeuser']))
 199                                 $this->dieUsage('user and excludeuser cannot be used together', 'badparams');
 200
 201                         // This code makes an assumption that sorting by rev_id and rev_timestamp produces
 202                         // the same result. This way users may request revisions starting at a given time,
 203                         // but to page through results use the rev_id returned after each page.
 204                         // Switching to rev_id removes the potential problem of having more than
 205                         // one row with the same timestamp for the same page.
 206                         // The order needs to be the same as start parameter to avoid SQL filesort.
 207
 208                         if (is_null($params['startid']) && is_null($params['endid']))
 209                                 $this->addWhereRange('rev_timestamp', $params['dir'],
 210                                         $params['start'], $params['end']);
 211                         else {
 212                                 $this->addWhereRange('rev_id', $params['dir'],
 213                                         $params['startid'], $params['endid']);
 214                                 // One of start and end can be set
 215                                 // If neither is set, this does nothing
 216                                 $this->addWhereRange('rev_timestamp', $params['dir'],
 217                                         $params['start'], $params['end'], false);
 218                         }
 219
 220                         // must manually initialize unset limit
 221                         if (is_null($limit))
 222                                 $limit = 10;
 223                         $this->validateLimit('limit', $limit, 1, $userMax, $botMax);
 224
 225                         // There is only one ID, use it
 226                         $ids = array_keys($pageSet->getGoodTitles());
 227                         $this->addWhereFld('rev_page', reset($ids));
 228
 229                         if(!is_null($params['user'])) {
 230                                 $this->addWhereFld('rev_user_text', $params['user']);
 231                         } elseif (!is_null($params['excludeuser'])) {
 232                                 $this->addWhere('rev_user_text != ' .
 233                                         $db->addQuotes($params['excludeuser']));
 234                         }
 235                         if(!is_null($params['user']) || !is_null($params['excludeuser'])) {
 236                                 // Paranoia: avoid brute force searches (bug 17342)
 237                                 $this->addWhere($db->bitAnd('rev_deleted',Revision::DELETED_USER) . ' = 0');
 238                         }
 239                 }
 240                 elseif ($revCount > 0) {
 241                         $max = $this->getMain()->canApiHighLimits() ? $botMax : $userMax;
 242                         $revs = $pageSet->getRevisionIDs();
 243                         if(self::truncateArray($revs, $max))
 244                                 $this->setWarning("Too many values supplied for parameter 'revids': the limit is $max");
 245
 246                         // Get all revision IDs
 247                         $this->addWhereFld('rev_id', array_keys($revs));
 248
 249                         if(!is_null($params['continue']))
 250                                 $this->addWhere("rev_id >= '" . intval($params['continue']) . "'");
 251                         $this->addOption('ORDER BY', 'rev_id');
 252
 253                         // assumption testing -- we should never get more then $revCount rows.
 254                         $limit = $revCount;
 255                 }
 256                 elseif ($pageCount > 0) {
 257                         $max = $this->getMain()->canApiHighLimits() ? $botMax : $userMax;
 258                         $titles = $pageSet->getGoodTitles();
 259                         if(self::truncateArray($titles, $max))
 260                                 $this->setWarning("Too many values supplied for parameter 'titles': the limit is $max");
 261
 262                         // When working in multi-page non-enumeration mode,
 263                         // limit to the latest revision only
 264                         $this->addWhere('page_id=rev_page');
 265                         $this->addWhere('page_latest=rev_id');
 266
 267                         // Get all page IDs
 268                         $this->addWhereFld('page_id', array_keys($titles));
 269                         // Every time someone relies on equality propagation, god kills a kitten :)
 270                         $this->addWhereFld('rev_page', array_keys($titles));
 271
 272                         if(!is_null($params['continue']))
 273                         {
 274                                 $cont = explode('|', $params['continue']);
 275                                 if(count($cont) != 2)
 276                                         $this->dieUsage("Invalid continue param. You should pass the original " .
 277                                                         "value returned by the previous query", "_badcontinue");
 278                                 $pageid = intval($cont[0]);
 279                                 $revid = intval($cont[1]);
 280                                 $this->addWhere("rev_page > '$pageid' OR " .
 281                                                 "(rev_page = '$pageid' AND " .
 282                                                 "rev_id >= '$revid')");
 283                         }
 284                         $this->addOption('ORDER BY', 'rev_page, rev_id');
 285
 286                         // assumption testing -- we should never get more then $pageCount rows.
 287                         $limit = $pageCount;
 288                 } else
 289                         ApiBase :: dieDebug(__METHOD__, 'param validation?');
 290
 291                 $this->addOption('LIMIT', $limit +1);
 292
 293                 $data = array ();
 294                 $count = 0;
 295                 $res = $this->select(__METHOD__);
 296
 297                 while ($row = $db->fetchObject($res)) {
 298
 299                         if (++ $count > $limit) {
 300                                 // We've reached the one extra which shows that there are additional pages to be had. Stop here...
 301                                 if (!$enumRevMode)
 302                                         ApiBase :: dieDebug(__METHOD__, 'Got more rows then expected'); // bug report
 303                                 $this->setContinueEnumParameter('startid', intval($row->rev_id));
 304                                 break;
 305                         }
 306
 307                         //
 308                         $fit = $this->addPageSubItem($row->rev_page, $this->extractRowInfo($row), 'rev');
 309                         if(!$fit)
 310                         {
 311                                 if($enumRevMode)
 312                                         $this->setContinueEnumParameter('startid', intval($row->rev_id));
 313                                 else if($revCount > 0)
 314                                         $this->setContinueEnumParameter('continue', intval($row->rev_id));
 315                                 else
 316                                         $this->setContinueEnumParameter('continue', intval($row->rev_page) .
 317                                                 '|' . intval($row->rev_id));
 318                                 break;
 319                         }
 320                 }
 321                 $db->freeResult($res);
 322         }
 323
 324         private function extractRowInfo( $row ) {
 325                 $revision = new Revision( $row );
 326                 $title = $revision->getTitle();
 327                 $vals = array ();
 328
 329                 if ($this->fld_ids) {
 330                         $vals['revid'] = intval($revision->getId());
 331                         // $vals['oldid'] = intval($row->rev_text_id);  // todo: should this be exposed?
 332                         if (!is_null($revision->getParentId()))
 333                                 $vals['parentid'] = intval($revision->getParentId());
 334                 }
 335
 336                 if ($this->fld_flags && $revision->isMinor())
 337                         $vals['minor'] = '';
 338
 339                 if ($this->fld_user) {
 340                         if ($revision->isDeleted(Revision::DELETED_USER)) {
 341                                 $vals['userhidden'] = '';
 342                         } else {
 343                                 $vals['user'] = $revision->getUserText();
 344                                 if (!$revision->getUser())
 345                                         $vals['anon'] = '';
 346                         }
 347                 }
 348
 349                 if ($this->fld_timestamp) {
 350                         $vals['timestamp'] = wfTimestamp(TS_ISO_8601, $revision->getTimestamp());
 351                 }
 352
 353                 if ($this->fld_size && !is_null($revision->getSize())) {
 354                         $vals['size'] = intval($revision->getSize());
 355                 }
 356
 357                 if ($this->fld_comment) {
 358                         if ($revision->isDeleted(Revision::DELETED_COMMENT)) {
 359                                 $vals['commenthidden'] = '';
 360                         } else {
 361                                 $comment = $revision->getComment();
 362                                 if (strval($comment) !== '')
 363                                         $vals['comment'] = $comment;
 364                         }
 365                 }
 366
 367                 if ($this->fld_tags && $row->ts_tags)
 368                         $vals['tags'] = $row->ts_tags;
 369
 370                 if(!is_null($this->token))
 371                 {
 372                         $tokenFunctions = $this->getTokenFunctions();
 373                         foreach($this->token as $t)
 374                         {
 375                                 $val = call_user_func($tokenFunctions[$t], $title->getArticleID(), $title, $revision);
 376                                 if($val === false)
 377                                         $this->setWarning("Action '$t' is not allowed for the current user");
 378                                 else
 379                                         $vals[$t . 'token'] = $val;
 380                         }
 381                 }
 382
 383                 if ($this->fld_content && !$revision->isDeleted(Revision::DELETED_TEXT)) {
 384                         global $wgParser;
 385                         $text = $revision->getText();
 386                         # Expand templates after getting section content because
 387                         # template-added sections don't count and Parser::preprocess()
 388                         # will have less input
 389                         if ($this->section !== false) {
 390                                 $text = $wgParser->getSection( $text, $this->section, false);
 391                                 if($text === false)
 392                                         $this->dieUsage("There is no section {$this->section} in r".$revision->getId(), 'nosuchsection');
 393                         }
 394                         if ($this->generateXML) {
 395                                 $wgParser->startExternalParse( $title, new ParserOptions(), OT_PREPROCESS );
 396                                 $dom = $wgParser->preprocessToDom( $text );
 397                                 if ( is_callable( array( $dom, 'saveXML' ) ) ) {
 398                                         $xml = $dom->saveXML();
 399                                 } else {
 400                                         $xml = $dom->__toString();
 401                                 }
 402                                 $vals['parsetree'] = $xml;
 403
 404                         }
 405                         if ($this->expandTemplates) {
 406                                 $text = $wgParser->preprocess( $text, $title, new ParserOptions() );
 407                         }
 408                         ApiResult :: setContent($vals, $text);
 409                 } else if ($this->fld_content) {
 410                         $vals['texthidden'] = '';
 411                 }
 412
 413                 if (!is_null($this->diffto)) {
 414                         global $wgAPIMaxUncachedDiffs;
 415                         static $n = 0; // Numer of uncached diffs we've had
 416                         if($n< $wgAPIMaxUncachedDiffs) {
 417                                 $engine = new DifferenceEngine($title, $revision->getID(), $this->diffto);
 418                                 $difftext = $engine->getDiffBody();
 419                                 $vals['diff']['from'] = $engine->getOldid();
 420                                 $vals['diff']['to'] = $engine->getNewid();
 421                                 ApiResult::setContent($vals['diff'], $difftext);
 422                                 if(!$engine->wasCacheHit())
 423                                         $n++;
 424                         } else {
 425                                 $vals['diff']['notcached'] = '';
 426                         }
 427                 }
 428                 return $vals;
 429         }
 430
 431         public function getAllowedParams() {
 432                 return array (
 433                         'prop' => array (
 434                                 ApiBase :: PARAM_ISMULTI => true,
 435                                 ApiBase :: PARAM_DFLT => 'ids|timestamp|flags|comment|user',
 436                                 ApiBase :: PARAM_TYPE => array (
 437                                         'ids',
 438                                         'flags',
 439                                         'timestamp',
 440                                         'user',
 441                                         'size',
 442                                         'comment',
 443                                         'content',
 444                                         'tags'
 445                                 )
 446                         ),
 447                         'limit' => array (
 448                                 ApiBase :: PARAM_TYPE => 'limit',
 449                                 ApiBase :: PARAM_MIN => 1,
 450                                 ApiBase :: PARAM_MAX => ApiBase :: LIMIT_BIG1,
 451                                 ApiBase :: PARAM_MAX2 => ApiBase :: LIMIT_BIG2
 452                         ),
 453                         'startid' => array (
 454                                 ApiBase :: PARAM_TYPE => 'integer'
 455                         ),
 456                         'endid' => array (
 457                                 ApiBase :: PARAM_TYPE => 'integer'
 458                         ),
 459                         'start' => array (
 460                                 ApiBase :: PARAM_TYPE => 'timestamp'
 461                         ),
 462                         'end' => array (
 463                                 ApiBase :: PARAM_TYPE => 'timestamp'
 464                         ),
 465                         'dir' => array (
 466                                 ApiBase :: PARAM_DFLT => 'older',
 467                                 ApiBase :: PARAM_TYPE => array (
 468                                         'newer',
 469                                         'older'
 470                                 )
 471                         ),
 472                         'user' => array(
 473                                 ApiBase :: PARAM_TYPE => 'user'
 474                         ),
 475                         'excludeuser' => array(
 476                                 ApiBase :: PARAM_TYPE => 'user'
 477                         ),
 478                         'tag' => null,
 479                         'expandtemplates' => false,
 480                         'generatexml' => false,
 481                         'section' => null,
 482                         'token' => array(
 483                                 ApiBase :: PARAM_TYPE => array_keys($this->getTokenFunctions()),
 484                                 ApiBase :: PARAM_ISMULTI => true
 485                         ),
 486                         'continue' => null,
 487                         'diffto' => null,
 488                 );
 489         }
 490
 491         public function getParamDescription() {
 492                 return array (
 493                         'prop' => 'Which properties to get for each revision.',
 494                         'limit' => 'limit how many revisions will be returned (enum)',
 495                         'startid' => 'from which revision id to start enumeration (enum)',
 496                         'endid' => 'stop revision enumeration on this revid (enum)',
 497                         'start' => 'from which revision timestamp to start enumeration (enum)',
 498                         'end' => 'enumerate up to this timestamp (enum)',
 499                         'dir' => 'direction of enumeration - towards "newer" or "older" revisions (enum)',
 500                         'user' => 'only include revisions made by user',
 501                         'excludeuser' => 'exclude revisions made by user',
 502                         'tag' => 'only list revisions with this tag',
 503                         'expandtemplates' => 'expand templates in revision content',
 504                         'generatexml' => 'generate XML parse tree for revision content',
 505                         'section' => 'only retrieve the content of this section',
 506                         'token' => 'Which tokens to obtain for each revision',
 507                         'continue' => 'When more results are available, use this to continue',
 508                         'diffto' => array('Revision ID to diff each revision to.',
 509                                 'Use "prev", "next" and "cur" for the previous, next and current revision respectively.'),
 510                 );
 511         }
 512
 513         public function getDescription() {
 514                 return array (
 515                         'Get revision information.',
 516                         'This module may be used in several ways:',
 517                         ' 1) Get data about a set of pages (last revision), by setting titles or pageids parameter.',
 518                         ' 2) Get revisions for one given page, by using titles/pageids with start/end/limit params.',
 519                         ' 3) Get data about a set of revisions by setting their IDs with revids parameter.',
 520                         'All parameters marked as (enum) may only be used with a single page (#2).'
 521                 );
 522         }
 523
 524         protected function getExamples() {
 525                 return array (
 526                         'Get data with content for the last revision of titles "API" and "Main Page":',
 527                         '  api.php?action=query&prop=revisions&titles=API|Main%20Page&rvprop=timestamp|user|comment|content',
 528                         'Get last 5 revisions of the "Main Page":',
 529                         '  api.php?action=query&prop=revisions&titles=Main%20Page&rvlimit=5&rvprop=timestamp|user|comment',
 530                         'Get first 5 revisions of the "Main Page":',
 531                         '  api.php?action=query&prop=revisions&titles=Main%20Page&rvlimit=5&rvprop=timestamp|user|comment&rvdir=newer',
 532                         'Get first 5 revisions of the "Main Page" made after 2006-05-01:',
 533                         '  api.php?action=query&prop=revisions&titles=Main%20Page&rvlimit=5&rvprop=timestamp|user|comment&rvdir=newer&rvstart=20060501000000',
 534                         'Get first 5 revisions of the "Main Page" that were not made made by anonymous user "127.0.0.1"',
 535                         '  api.php?action=query&prop=revisions&titles=Main%20Page&rvlimit=5&rvprop=timestamp|user|comment&rvexcludeuser=127.0.0.1',
 536                         'Get first 5 revisions of the "Main Page" that were made by the user "MediaWiki default"',
 537                         '  api.php?action=query&prop=revisions&titles=Main%20Page&rvlimit=5&rvprop=timestamp|user|comment&rvuser=MediaWiki%20default',
 538                 );
 539         }
 540
 541         public function getVersion() {
 542                 return __CLASS__ . ': $Id$';
 543         }
 544 }