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  * @addtogroup 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;
  46
  47         public function execute() {
  48                 $limit = $startid = $endid = $start = $end = $dir = $prop = $user = $excludeuser = $diffto = $difftoprev = $diffformat = $token = null;
  49                 extract($this->extractRequestParams());
  50
  51                 // If any of those parameters are used, work in 'enumeration' mode.
  52                 // Enum mode can only be used when exactly one page is provided.
  53                 // Enumerating revisions on multiple pages make it extremely
  54                 // difficult to manage continuations and require additional SQL indexes
  55                 $enumRevMode = (!is_null($user) || !is_null($excludeuser) || !is_null($limit) || !is_null($startid) || !is_null($endid) || $dir === 'newer' || !is_null($start) || !is_null($end) | $difftoprev);
  56
  57
  58                 $pageSet = $this->getPageSet();
  59                 $pageCount = $pageSet->getGoodTitleCount();
  60                 $revCount = $pageSet->getRevisionCount();
  61
  62                 // Optimization -- nothing to do
  63                 if ($revCount === 0 && $pageCount === 0)
  64                         return;
  65
  66                 if ($revCount > 0 && $enumRevMode)
  67                         $this->dieUsage('The revids= parameter may not be used with the list options (limit, startid, endid, dirNewer, start, end).', 'revids');
  68
  69                 if ($pageCount > 1 && $enumRevMode)
  70                         $this->dieUsage('titles, pageids or a generator was used to supply multiple pages, but the limit, startid, endid, dirNewer, user, excludeuser, start, end and difftoprev parameters may only be used on a single page.', 'multpages');
  71
  72                 $this->addTables('revision');
  73                 $this->addWhere('rev_deleted=0');
  74
  75                 $prop = array_flip($prop);
  76
  77                 // These field are needed regardless of the client requesting them
  78                 $this->addFields('rev_id');
  79                 $this->addFields('rev_page');
  80
  81                 // Optional fields
  82                 $this->fld_ids = isset ($prop['ids']);
  83                 // $this->addFieldsIf('rev_text_id', $this->fld_ids); // should this be exposed?
  84                 $this->fld_flags = $this->addFieldsIf('rev_minor_edit', isset ($prop['flags']));
  85                 $this->fld_timestamp = $this->addFieldsIf('rev_timestamp', isset ($prop['timestamp']));
  86                 $this->fld_comment = $this->addFieldsIf('rev_comment', isset ($prop['comment']));
  87                 $this->fld_size = $this->addFieldsIf('rev_len', isset ($prop['size']));
  88                 if(!is_null($token))
  89                 {
  90                         $this->tok_rollback = $this->getTokenFlag($token, 'rollback');
  91                 }
  92
  93                 if($diffto || $difftoprev)
  94                         switch($diffformat)
  95                         {
  96                                 // To add your own formatter, create a subclass of DiffFormatter
  97                                 // in DifferenceEngine.php, then add it here
  98                                 case 'traditional':
  99                                         $this->formatter = new DiffFormatter;
 100                                         break;
 101                                 case 'unified':
 102                                         $this->formatter = new UnifiedDiffFormatter;
 103                                         break;
 104                                 case 'array':
 105                                         $this->formatter = new ArrayDiffFormatter;
 106                                         break;
 107                         }
 108                 if($diffto)
 109                 {
 110                         global $wgContLang;
 111                         $difftoRev = Revision::newFromID($diffto);
 112                         if(!($difftoRev instanceof Revision))
 113                                 $this->dieUsage("There is no revision with ID $diffto", 'nosuchrev');
 114                         $this->diffOldText = $difftoRev->revText();
 115                         if($this->diffOldText == '') // deleted revision
 116                                 $this->dieUsage("There is no revision with ID $diffto", 'nosuchrev'); // fake non-existence
 117                         $this->diffOldText = explode("\n", $wgContLang->segmentForDiff($this->diffOldText));
 118                         $this->diffto = $diffto;
 119                 }
 120                 else
 121                         $this->diffto = false;
 122                 if($difftoprev)
 123                 {
 124                         $this->revCache = array();
 125                         $this->difftoprev = true;
 126                 }
 127                 else
 128                         $this->difftoprev = false;
 129
 130                 if (isset ($prop['user'])) {
 131                         $this->addFields('rev_user');
 132                         $this->addFields('rev_user_text');
 133                         $this->fld_user = true;
 134                 }
 135                 else if($this->tok_rollback)
 136                         $this->addFields('rev_user_text');
 137
 138                 if (isset ($prop['content']) || !is_null($diffto) || $difftoprev) {
 139
 140                         // For each page we will request, the user must have read rights for that page
 141                         foreach ($pageSet->getGoodTitles() as $title) {
 142                                 if( !$title->userCanRead() )
 143                                         $this->dieUsage(
 144                                                 'The current user is not allowed to read ' . $title->getPrefixedText(),
 145                                                 'accessdenied');
 146                         }
 147
 148                         $this->addTables('text');
 149                         $this->addWhere('rev_text_id=old_id');
 150                         $this->addFields('old_id');
 151                         $this->addFields('old_text');
 152                         $this->addFields('old_flags');
 153
 154                         $this->fld_content = isset($prop['content']);
 155
 156                         $this->expandTemplates = $expandtemplates;
 157                 }
 158
 159                 $userMax = ($this->fld_content || $diffto || $difftoprev ? 50 : 500);
 160                 $botMax = ($this->fld_content || $diffto || $difftoprev ? 200 : 10000);
 161
 162                 if ($enumRevMode) {
 163
 164                         // This is mostly to prevent parameter errors (and optimize SQL?)
 165                         if (!is_null($startid) && !is_null($start))
 166                                 $this->dieUsage('start and startid cannot be used together', 'badparams');
 167
 168                         if (!is_null($endid) && !is_null($end))
 169                                 $this->dieUsage('end and endid cannot be used together', 'badparams');
 170
 171                         if(!is_null($user) && !is_null( $excludeuser))
 172                                 $this->dieUsage('user and excludeuser cannot be used together', 'badparams');
 173
 174                         // This code makes an assumption that sorting by rev_id and rev_timestamp produces
 175                         // the same result. This way users may request revisions starting at a given time,
 176                         // but to page through results use the rev_id returned after each page.
 177                         // Switching to rev_id removes the potential problem of having more than
 178                         // one row with the same timestamp for the same page.
 179                         // The order needs to be the same as start parameter to avoid SQL filesort.
 180
 181                         if (is_null($startid) && is_null($endid))
 182                                 $this->addWhereRange('rev_timestamp', $dir, $start, $end);
 183                         else
 184                                 $this->addWhereRange('rev_id', $dir, $startid, $endid);
 185
 186                         // must manually initialize unset limit
 187                         if (is_null($limit))
 188                                 $limit = 10;
 189                         $this->validateLimit('limit', $limit, 1, $userMax, $botMax);
 190
 191                         // There is only one ID, use it
 192                         $this->addWhereFld('rev_page', current(array_keys($pageSet->getGoodTitles())));
 193
 194                         if(!is_null($user)) {
 195                                 $this->addWhereFld('rev_user_text', $user);
 196                         } elseif (!is_null( $excludeuser)) {
 197                                 $this->addWhere('rev_user_text != ' . $this->getDB()->addQuotes($excludeuser));
 198                         }
 199                 }
 200                 elseif ($revCount > 0) {
 201                         $this->validateLimit('rev_count', $revCount, 1, $userMax, $botMax);
 202
 203                         // Get all revision IDs
 204                         $this->addWhereFld('rev_id', array_keys($pageSet->getRevisionIDs()));
 205
 206                         // assumption testing -- we should never get more then $revCount rows.
 207                         $limit = $revCount;
 208                 }
 209                 elseif ($pageCount > 0) {
 210                         // When working in multi-page non-enumeration mode,
 211                         // limit to the latest revision only
 212                         $this->addTables('page');
 213                         $this->addWhere('page_id=rev_page');
 214                         $this->addWhere('page_latest=rev_id');
 215                         $this->validateLimit('page_count', $pageCount, 1, $userMax, $botMax);
 216
 217                         // Get all page IDs
 218                         $this->addWhereFld('page_id', array_keys($pageSet->getGoodTitles()));
 219
 220                         // assumption testing -- we should never get more then $pageCount rows.
 221                         $limit = $pageCount;
 222                 } else
 223                         ApiBase :: dieDebug(__METHOD__, 'param validation?');
 224
 225                 $this->addOption('LIMIT', $limit +1);
 226
 227                 $data = array ();
 228                 $count = 0;
 229                 $res = $this->select(__METHOD__);
 230
 231                 $db = $this->getDB();
 232                 while ($row = $db->fetchObject($res)) {
 233
 234                         if (++ $count > $limit) {
 235                                 // We've reached the one extra which shows that there are additional pages to be had. Stop here...
 236                                 if (!$enumRevMode)
 237                                         ApiBase :: dieDebug(__METHOD__, 'Got more rows then expected'); // bug report
 238                                 $this->setContinueEnumParameter('startid', intval($row->rev_id));
 239                                 break;
 240                         }
 241
 242                         $this->getResult()->addValue(
 243                                 array (
 244                                         'query',
 245                                         'pages',
 246                                         intval($row->rev_page),
 247                                         'revisions'),
 248                                 null,
 249                                 $this->extractRowInfo($row));
 250                 }
 251                 $db->freeResult($res);
 252
 253                 if($this->difftoprev)
 254                 {
 255                         global $wgContLang;
 256                         ksort($this->revCache, SORT_NUMERIC);
 257                         $previousRevID = null;
 258                         $oldText = null;
 259                         $data =& $this->getResult()->getData();
 260                         $pageID = current(array_keys($pageSet->getGoodTitles()));
 261                         $this->diffArr = array();
 262                         foreach($this->revCache as $revid => $revtext)
 263                         {
 264                                 $r = array();
 265                                 if(is_null($previousRevID))
 266                                 {
 267                                         // First run
 268                                         $previousRevID = $revid;
 269                                         $oldText = explode("\n", $wgContLang->segmentForDiff($revtext));
 270                                         continue;
 271                                 }
 272                                 $newText = explode("\n", $wgContLang->segmentForDiff($revtext));
 273                                 $diff = new Diff($oldText, $newText);
 274                                 $r['from'] = $previousRevID;
 275                                 $formatted = $this->formatter->format($diff);
 276                                 if(!is_array($formatted))
 277                                         ApiResult::setContent($r, $wgContLang->unsegmentForDiff($this->formatter->format($diff)));
 278                                 else
 279                                 {
 280                                         $r['differences'] = $formatted;
 281                                         $this->getResult()->setIndexedTagName($r['differences'], 'diff');
 282                                 }
 283                                 $this->diffArr[$revid] = $r;
 284
 285                                 $previousRevID = $revid;
 286                                 $oldText = $newText;
 287                         }
 288
 289                         if ( $this->diffArr ) {
 290                                 # Populate the query result with the contents of $this->diffArr.
 291                                 $knownrevs = array_keys($this->diffArr);
 292                                 $i = count($knownrevs) - 1;
 293                                 foreach($data['query']['pages'][$pageID]['revisions'] as &$rev) {
 294                                         if ( $i >= 0 && isset ( $this->diffArr[$knownrevs[$i]] ) )
 295                                                 $rev['difftoprev'] = $this->diffArr[$knownrevs[$i]];
 296                                         $i --;
 297                                 }
 298                         }
 299                 }
 300
 301                 // Ensure that all revisions are shown as '<rev>' elements
 302                 $result = $this->getResult();
 303                 if ($result->getIsRawMode()) {
 304                         $data =& $result->getData();
 305                         foreach ($data['query']['pages'] as & $page) {
 306                                 if (is_array($page) && array_key_exists('revisions', $page)) {
 307                                         $result->setIndexedTagName($page['revisions'], 'rev');
 308                                 }
 309                         }
 310                 }
 311         }
 312
 313         private function extractRowInfo($row) {
 314
 315                 $vals = array ();
 316
 317                 if ($this->fld_ids) {
 318                         $vals['revid'] = intval($row->rev_id);
 319                         // $vals['oldid'] = intval($row->rev_text_id);  // todo: should this be exposed?
 320                 }
 321
 322                 if ($this->fld_flags && $row->rev_minor_edit)
 323                         $vals['minor'] = '';
 324
 325                 if ($this->fld_user) {
 326                         $vals['user'] = $row->rev_user_text;
 327                         if (!$row->rev_user)
 328                                 $vals['anon'] = '';
 329                 }
 330
 331                 if ($this->fld_timestamp) {
 332                         $vals['timestamp'] = wfTimestamp(TS_ISO_8601, $row->rev_timestamp);
 333                 }
 334
 335                 if ($this->fld_size && !is_null($row->rev_len)) {
 336                         $vals['size'] = intval($row->rev_len);
 337                 }
 338
 339                 if ($this->fld_comment && !empty ($row->rev_comment)) {
 340                         $vals['comment'] = $row->rev_comment;
 341                 }
 342
 343                 if($this->tok_rollback || ($this->fld_content && $this->expandTemplates))
 344                         $title = Title::newFromID($row->rev_page);
 345
 346                 if($this->tok_rollback) {
 347                         global $wgUser;
 348                         $vals['rollbacktoken'] = $wgUser->editToken(array($title->getPrefixedText(), $row->rev_user_text));
 349                 }
 350
 351                 if ($this->fld_content || $this->diffto || $this->difftoprev)
 352                         $text = Revision :: getRevisionText($row);
 353                 if ($this->fld_content) {
 354                         if ($this->expandTemplates) {
 355                                 global $wgParser;
 356                                 $text = $wgParser->preprocess( $text, $title, new ParserOptions() );
 357                         }
 358                         ApiResult :: setContent($vals, $text);
 359                 }
 360
 361                 if($this->diffto)
 362                 {
 363                         global $wgContLang;
 364                         $newText = explode("\n", $wgContLang->segmentForDiff($text));
 365                         $diff = new Diff($this->diffOldText, $newText);
 366                         $vals['diffto']['from'] = $this->diffto;
 367                         $arraydiff = $this->formatter instanceof ArrayDiffFormatter;
 368                         if( $arraydiff ) {
 369                                 $changes = $wgContLang->unsegmentForDiff($this->formatter->format($diff));
 370                                 $this->getResult()->setIndexedTagName( $changes, 'change' );
 371                                 $vals['diffto'] = $changes;
 372                         } else {
 373                                 ApiResult::setContent($vals['diffto'], $wgContLang->unsegmentForDiff($this->formatter->format($diff)));
 374                         }
 375                 }
 376                 if($this->difftoprev)
 377                         // Cache the revision's content for later use
 378                         $this->revCache[$row->rev_id] = $text;
 379
 380                 return $vals;
 381         }
 382
 383         protected function getAllowedParams() {
 384                 return array (
 385                         'prop' => array (
 386                                 ApiBase :: PARAM_ISMULTI => true,
 387                                 ApiBase :: PARAM_DFLT => 'ids|timestamp|flags|comment|user',
 388                                 ApiBase :: PARAM_TYPE => array (
 389                                         'ids',
 390                                         'flags',
 391                                         'timestamp',
 392                                         'user',
 393                                         'size',
 394                                         'comment',
 395                                         'content',
 396                                 )
 397                         ),
 398                         'limit' => array (
 399                                 ApiBase :: PARAM_TYPE => 'limit',
 400                                 ApiBase :: PARAM_MIN => 1,
 401                                 ApiBase :: PARAM_MAX => ApiBase :: LIMIT_BIG1,
 402                                 ApiBase :: PARAM_MAX2 => ApiBase :: LIMIT_BIG2
 403                         ),
 404                         'startid' => array (
 405                                 ApiBase :: PARAM_TYPE => 'integer'
 406                         ),
 407                         'endid' => array (
 408                                 ApiBase :: PARAM_TYPE => 'integer'
 409                         ),
 410                         'start' => array (
 411                                 ApiBase :: PARAM_TYPE => 'timestamp'
 412                         ),
 413                         'end' => array (
 414                                 ApiBase :: PARAM_TYPE => 'timestamp'
 415                         ),
 416                         'dir' => array (
 417                                 ApiBase :: PARAM_DFLT => 'older',
 418                                 ApiBase :: PARAM_TYPE => array (
 419                                         'newer',
 420                                         'older'
 421                                 )
 422                         ),
 423                         'user' => array(
 424                                 ApiBase :: PARAM_TYPE => 'user'
 425                         ),
 426                         'excludeuser' => array(
 427                                 ApiBase :: PARAM_TYPE => 'user'
 428                         ),
 429
 430                         'expandtemplates' => false,
 431                         'diffto' => array(
 432                                 ApiBase :: PARAM_TYPE => 'integer'
 433                         ),
 434                         'difftoprev' => false,
 435                         'diffformat' => array(
 436                                 ApiBase :: PARAM_TYPE => array(
 437                                         'traditional',
 438                                         'unified',
 439                                         'array',
 440                                 ),
 441                                 ApiBase :: PARAM_DFLT => 'unified'
 442                         ),
 443                         'token' => array(
 444                                 ApiBase :: PARAM_TYPE => array(
 445                                         'rollback'
 446                                 ),
 447                                 ApiBase :: PARAM_ISMULTI => true
 448                         ),
 449                 );
 450         }
 451
 452         protected function getParamDescription() {
 453                 return array (
 454                         'prop' => 'Which properties to get for each revision.',
 455                         'limit' => 'limit how many revisions will be returned (enum)',
 456                         'startid' => 'from which revision id to start enumeration (enum)',
 457                         'endid' => 'stop revision enumeration on this revid (enum)',
 458                         'start' => 'from which revision timestamp to start enumeration (enum)',
 459                         'end' => 'enumerate up to this timestamp (enum)',
 460                         'dir' => 'direction of enumeration - towards "newer" or "older" revisions (enum)',
 461                         'user' => 'only include revisions made by user',
 462                         'excludeuser' => 'exclude revisions made by user',
 463                         'expandtemplates' => 'expand templates in revision content',
 464                         'diffto' => 'Revision number to compare all revisions with',
 465                         'difftoprev' => 'Diff each revision to the previous one (enum)',
 466                         'diffformat' => 'Format to use for diffs',
 467                         'token' => 'Which tokens to obtain for each revision',
 468                 );
 469         }
 470
 471         protected function getDescription() {
 472                 return array (
 473                         'Get revision information.',
 474                         'This module may be used in several ways:',
 475                         ' 1) Get data about a set of pages (last revision), by setting titles or pageids parameter.',
 476                         ' 2) Get revisions for one given page, by using titles/pageids with start/end/limit params.',
 477                         ' 3) Get data about a set of revisions by setting their IDs with revids parameter.',
 478                         'All parameters marked as (enum) may only be used with a single page (#2).'
 479                 );
 480         }
 481
 482         protected function getExamples() {
 483                 return array (
 484                         'Get data with content for the last revision of the "Main Page":',
 485                         '  api.php?action=query&prop=revisions&titles=Main%20Page&rvprop=timestamp|user|comment|content',
 486                         'Get last 5 revisions of the "Main Page":',
 487                         '  api.php?action=query&prop=revisions&titles=Main%20Page&rvlimit=5&rvprop=timestamp|user|comment',
 488                         'Get first 5 revisions of the "Main Page":',
 489                         '  api.php?action=query&prop=revisions&titles=Main%20Page&rvlimit=5&rvprop=timestamp|user|comment&rvdir=newer',
 490                         'Get first 5 revisions of the "Main Page" made after 2006-05-01:',
 491                         '  api.php?action=query&prop=revisions&titles=Main%20Page&rvlimit=5&rvprop=timestamp|user|comment&rvdir=newer&rvstart=20060501000000',
 492                         'Get first 5 revisions of the "Main Page" that were not made made by anonymous user "127.0.0.1"',
 493                         '  api.php?action=query&prop=revisions&titles=Main%20Page&rvlimit=5&rvprop=timestamp|user|comment&rvexcludeuser=127.0.0.1',
 494                         'Get first 5 revisions of the "Main Page" that were made by the user "MediaWiki default"',
 495                         '  api.php?action=query&prop=revisions&titles=Main%20Page&rvlimit=5&rvprop=timestamp|user|comment&rvuser=MediaWiki%20default',
 496                 );
 497         }
 498
 499         public function getVersion() {
 500                 return __CLASS__ . ': $Id$';
 501         }
 502 }
 503