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