includes/api/ApiPageSet.php

   1 <?php
   2
   3
   4 /*
   5  * Created on Sep 24, 2006
   6  *
   7  * API for MediaWiki 1.8+
   8  *
   9  * Copyright (C) 2006 Yuri Astrakhan <FirstnameLastname@gmail.com>
  10  *
  11  * This program is free software; you can redistribute it and/or modify
  12  * it under the terms of the GNU General Public License as published by
  13  * the Free Software Foundation; either version 2 of the License, or
  14  * (at your option) any later version.
  15  *
  16  * This program is distributed in the hope that it will be useful,
  17  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  18  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
  19  * GNU General Public License for more details.
  20  *
  21  * You should have received a copy of the GNU General Public License along
  22  * with this program; if not, write to the Free Software Foundation, Inc.,
  23  * 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.
  24  * http://www.gnu.org/copyleft/gpl.html
  25  */
  26
  27 if (!defined('MEDIAWIKI')) {
  28         // Eclipse helper - will be ignored in production
  29         require_once ('ApiQueryBase.php');
  30 }
  31
  32 class ApiPageSet extends ApiQueryBase {
  33
  34         private $mAllPages; // [ns][dbkey] => page_id or 0 when missing
  35         private $mGoodTitles, $mMissingTitles, $mMissingPageIDs, $mRedirectTitles, $mNormalizedTitles;
  36
  37         private $mRequestedFields;
  38
  39         public function __construct($query) {
  40                 parent :: __construct($query, __CLASS__);
  41
  42                 $this->mAllPages = array ();
  43                 $this->mGoodTitles = array ();
  44                 $this->mMissingTitles = array ();
  45                 $this->mMissingPageIDs = array ();
  46                 $this->mRedirectTitles = array ();
  47                 $this->mNormalizedTitles = array ();
  48
  49                 $this->mRequestedFields = array ();
  50         }
  51
  52         public function requestField($fieldName) {
  53                 $this->mRequestedFields[$fieldName] = null;
  54         }
  55
  56         public function getCustomField($fieldName) {
  57                 return $this->mRequestedFields[$fieldName];
  58         }
  59
  60         /**
  61          * Title objects that were found in the database.
  62          * @return array page_id (int) => Title (obj)
  63          */
  64         public function getGoodTitles() {
  65                 return $this->mGoodTitles;
  66         }
  67
  68         /**
  69          * Title objects that were NOT found in the database.
  70          * @return array of Title objects
  71          */
  72         public function getMissingTitles() {
  73                 return $this->mMissingTitles;
  74         }
  75
  76         /**
  77          * Page IDs that were not found in the database
  78          * @return array of page IDs
  79          */
  80         public function getMissingPageIDs() {
  81                 return $this->mMissingPageIDs;
  82         }
  83
  84         /**
  85          * Get a list of redirects when doing redirect resolution
  86          * @return array prefixed_title (string) => prefixed_title (string)
  87          */
  88         public function getRedirectTitles() {
  89                 return $this->mRedirectTitles;
  90         }
  91
  92         /**
  93          * Get a list of title normalizations - maps the title given
  94          * with its normalized version.
  95          * @return array raw_prefixed_title (string) => prefixed_title (string)
  96          */
  97         public function getNormalizedTitles() {
  98                 return $this->mNormalizedTitles;
  99         }
 100
 101         /**
 102          * Returns the number of unique pages (not revisions) in the set.
 103          */
 104         public function getGoodTitleCount() {
 105                 return count($this->getGoodTitles());
 106         }
 107
 108         /**
 109          * Get the list of revision IDs (requested with revids= parameter)
 110          */
 111         public function getRevisionIDs() {
 112                 $this->dieUsage(__METHOD__ . ' is not implemented', 'notimplemented');
 113         }
 114
 115         /**
 116          * Returns the number of revisions (requested with revids= parameter)
 117          */
 118         public function getRevisionCount() {
 119                 return 0; // TODO: implement
 120         }
 121
 122         /**
 123          * This method populates internal variables with page information
 124          * based on the given array of title strings.
 125          *
 126          * Steps:
 127          * #1 For each title, get data from `page` table
 128          * #2 If page was not found in the DB, store it as missing
 129          *
 130          * Additionally, when resolving redirects:
 131          * #3 If no more redirects left, stop.
 132          * #4 For each redirect, get its links from `pagelinks` table.
 133          * #5 Substitute the original LinkBatch object with the new list
 134          * #6 Repeat from step #1
 135          */
 136         private function populatePages($titles, $pageids, $redirects) {
 137                 if (!is_null($titles) && !is_null($pageids))
 138                         ApiBase :: dieDebug(__METHOD__, 'bad parameters');
 139                 $processTitles = !is_null($titles);
 140
 141                 // Ensure we get minimum required fields
 142                 $pageFlds = array (
 143                         'page_id' => null,
 144                         'page_namespace' => null,
 145                         'page_title' => null
 146                 );
 147
 148                 // only store non-default fields
 149                 $this->mRequestedFields = array_diff_key($this->mRequestedFields, $pageFlds);
 150
 151                 if ($redirects)
 152                         $pageFlds['page_is_redirect'] = null;
 153
 154                 $pageFlds = array_keys(array_merge($pageFlds, $this->mRequestedFields));
 155
 156                 $db = $this->getDB();
 157
 158                 if ($processTitles) {
 159
 160                         // Get validated and normalized title objects
 161                         $linkBatch = $this->processTitlesStrArray($titles);
 162
 163                         $set = $linkBatch->constructSet('page', $db);
 164                 } else {
 165                         $set = array (
 166                                 'page_id' => $pageids
 167                         );
 168                 }
 169
 170                 //
 171                 // Repeat until all redirects have been resolved
 172                 // The infinite loop is prevented by keeping all known pages in $this->mAllPages
 173                 //
 174                 do {
 175                         if ($processTitles) {
 176                                 // Hack: get the ns:titles stored in array(ns => array(titles)) format
 177                                 $remaining = $linkBatch->data;
 178                         } else {
 179                                 $remaining = array_flip($pageids); // turn pageids into keys
 180                         }
 181
 182                         $redirectIds = array ();
 183
 184                         //
 185                         // Get data about $linkBatch from `page` table
 186                         //
 187                         $this->profileDBIn();
 188                         $res = $db->select('page', $pageFlds, $set, __METHOD__);
 189                         $this->profileDBOut();
 190                         while ($row = $db->fetchObject($res)) {
 191
 192                                 $pageId = intval($row->page_id);
 193
 194                                 if ($processTitles)
 195                                         unset ($remaining[$row->page_namespace][$row->page_title]);
 196                                 else
 197                                         unset ($remaining[$pageId]);
 198
 199                                 $title = Title :: makeTitle($row->page_namespace, $row->page_title);
 200                                 $this->mAllPages[$row->page_namespace][$row->page_title] = $pageId;
 201
 202                                 if ($redirects && $row->page_is_redirect == '1') {
 203                                         $redirectIds[$pageId] = $title;
 204                                 } else {
 205                                         $this->mGoodTitles[$pageId] = $title;
 206                                 }
 207
 208                                 foreach ($this->mRequestedFields as $fieldName => & $fieldValues) {
 209                                         $fieldValues[$pageId] = $row-> $fieldName;
 210                                 }
 211                         }
 212                         $db->freeResult($res);
 213
 214                         if ($processTitles) {
 215                                 // The remaining titles in $remaining are non-existant pages
 216                                 foreach ($remaining as $ns => $dbkeys) {
 217                                         foreach ($dbkeys as $dbkey => $nothing) {
 218                                                 $this->mMissingTitles[] = Title :: makeTitle($ns, $dbkey);
 219                                                 $this->mAllPages[$ns][$dbkey] = 0;
 220                                         }
 221                                 }
 222                         } else {
 223                                 // The remaining pageids in $remaining do not exist
 224                                 foreach ($remaining as $pageid => $ignore) {
 225                                         $this->mMissingPageIDs[] = $pageid;
 226                                 }
 227                         }
 228
 229                         if (!$redirects || empty ($redirectIds))
 230                                 break;
 231
 232                         //
 233                         // Resolve redirects by querying the pagelinks table, and repeat the process
 234                         // Create a new linkBatch object for the next pass
 235                         //
 236                         $linkBatch = $this->resolveRedirectList($redirectIds);
 237
 238                         // Redirects are always titles
 239                         $processTitles = true;
 240                 }
 241                 while (false !== ($set = $linkBatch->constructSet('page', $db)));
 242         }
 243
 244         private function resolveRedirectList($redirectIds) {
 245
 246                 $linkBatch = new LinkBatch();
 247                 $db = $this->getDB();
 248
 249                 // find redirect targets for all redirect pages
 250                 $this->profileDBIn();
 251                 $res = $db->select('pagelinks', array (
 252                         'pl_from',
 253                         'pl_namespace',
 254                         'pl_title'
 255                 ), array (
 256                         'pl_from' => array_keys($redirectIds
 257                 )), __METHOD__);
 258                 $this->profileDBOut();
 259
 260                 while ($row = $db->fetchObject($res)) {
 261
 262                         $plfrom = intval($row->pl_from);
 263
 264                         // Bug 7304 workaround
 265                         // ( http://bugzilla.wikipedia.org/show_bug.cgi?id=7304 )
 266                         // A redirect page may have more than one link.
 267                         // This code will only use the first link returned.
 268                         if (isset ($redirectIds[$plfrom])) { // remove line when bug 7304 is fixed
 269
 270                                 $titleStrFrom = $redirectIds[$plfrom]->getPrefixedText();
 271                                 $titleStrTo = Title :: makeTitle($row->pl_namespace, $row->pl_title)->getPrefixedText();
 272                                 unset ($redirectIds[$plfrom]); // remove line when bug 7304 is fixed
 273
 274                                 // Avoid an infinite loop by checking if we have already processed this target
 275                                 if (!isset ($this->mAllPages[$row->pl_namespace][$row->pl_title])) {
 276                                         $linkBatch->add($row->pl_namespace, $row->pl_title);
 277                                 }
 278                         } else {
 279                                 // This redirect page has more than one link.
 280                                 // This is very slow, but safer until bug 7304 is resolved
 281                                 $title = Title :: newFromID($plfrom);
 282                                 $titleStrFrom = $title->getPrefixedText();
 283
 284                                 $article = new Article($title);
 285                                 $text = $article->getContent();
 286                                 $titleTo = Title :: newFromRedirect($text);
 287                                 $titleStrTo = $titleTo->getPrefixedText();
 288
 289                                 if (is_null($titleStrTo))
 290                                         ApiBase :: dieDebug(__METHOD__, 'Bug7304 workaround: redir target from {$title->getPrefixedText()} not found');
 291
 292                                 // Avoid an infinite loop by checking if we have already processed this target
 293                                 if (!isset ($this->mAllPages[$titleTo->getNamespace()][$titleTo->getDBkey()])) {
 294                                         $linkBatch->addObj($titleTo);
 295                                 }
 296                         }
 297
 298                         $this->mRedirectTitles[$titleStrFrom] = $titleStrTo;
 299                 }
 300                 $db->freeResult($res);
 301
 302                 return $linkBatch;
 303         }
 304
 305         /**
 306          * Given an array of title strings, convert them into Title objects.
 307          * This method validates access rights for the title,
 308          * and appends normalization values to the output.
 309          *
 310          * @return LinkBatch of title objects.
 311          */
 312         private function processTitlesStrArray($titles) {
 313
 314                 $linkBatch = new LinkBatch();
 315
 316                 foreach ($titles as $titleString) {
 317                         $titleObj = Title :: newFromText($titleString);
 318
 319                         // Validation
 320                         if (!$titleObj)
 321                                 $this->dieUsage("bad title $titleString", 'invalidtitle');
 322                         if ($titleObj->getNamespace() < 0)
 323                                 $this->dieUsage("No support for special page $titleString has been implemented", 'unsupportednamespace');
 324                         if (!$titleObj->userCanRead())
 325                                 $this->dieUsage("No read permission for $titleString", 'titleaccessdenied');
 326
 327                         $linkBatch->addObj($titleObj);
 328
 329                         // Make sure we remember the original title that was given to us
 330                         // This way the caller can correlate new titles with the originally requested,
 331                         // i.e. namespace is localized or capitalization is different
 332                         if ($titleString !== $titleObj->getPrefixedText()) {
 333                                 $this->mNormalizedTitles[$titleString] = $titleObj->getPrefixedText();
 334                         }
 335                 }
 336
 337                 return $linkBatch;
 338         }
 339
 340         private function populateRevIDs($revids) {
 341                 $this->dieUsage(__METHOD__ . ' is not implemented', 'notimplemented');
 342         }
 343
 344         public function execute() {
 345                 $this->profileIn();
 346                 $titles = $pageids = $revids = $redirects = null;
 347                 extract($this->extractRequestParams());
 348
 349                 // Only one of the titles/pageids/revids is allowed at the same time
 350                 $dataSource = null;
 351                 if (isset ($titles))
 352                         $dataSource = 'titles';
 353                 if (isset ($pageids)) {
 354                         if (isset ($dataSource))
 355                                 $this->dieUsage("Cannot use 'pageids' at the same time as '$dataSource'", 'multisource');
 356                         $dataSource = 'pageids';
 357                 }
 358                 if (isset ($revids)) {
 359                         if (isset ($dataSource))
 360                                 $this->dieUsage("Cannot use 'revids' at the same time as '$dataSource'", 'multisource');
 361                         $dataSource = 'revids';
 362                 }
 363
 364                 switch ($dataSource) {
 365                         case 'titles' :
 366                         case 'pageids' :
 367                                 $this->populatePages($titles, $pageids, $redirects);
 368                                 break;
 369                         case 'revids' :
 370                                 $this->populateRevIDs($revids);
 371                                 break;
 372                         default :
 373                                 // Do nothing - some queries do not need any of the data sources.
 374                                 break;
 375                 }
 376                 $this->profileOut();
 377         }
 378
 379         /**
 380          * This method is used by generators to pass the list of pageIDs internaly
 381          */
 382         public function executeForPageIDs($pageIDs) {
 383                 $this->profileIn();
 384                 $pageIDs = array_map( 'intval', $pageIDs );     // paranoia
 385                 $this->populatePages(null, $pageIDs, $this->getParameter('redirects'));
 386                 $this->profileOut();
 387         }
 388
 389         protected function getAllowedParams() {
 390                 return array (
 391                         'titles' => array (
 392                                 ApiBase :: PARAM_ISMULTI => true
 393                         ),
 394                         'pageids' => array (
 395                                 ApiBase :: PARAM_TYPE => 'integer',
 396                                 ApiBase :: PARAM_ISMULTI => true
 397                         ),
 398                         'revids' => array (
 399                                 ApiBase :: PARAM_TYPE => 'integer',
 400                                 ApiBase :: PARAM_ISMULTI => true
 401                         ),
 402                         'redirects' => false
 403                 );
 404         }
 405
 406         protected function getParamDescription() {
 407                 return array (
 408                         'titles' => 'A list of titles to work on',
 409                         'pageids' => 'A list of page IDs to work on',
 410                         'revids' => 'A list of revision IDs to work on',
 411                         'redirects' => 'Automatically resolve redirects'
 412                 );
 413         }
 414
 415         public function getVersion() {
 416                 return __CLASS__ . ': $Id$';
 417         }
 418 }
 419 ?>