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, $mRedirectTitles, $mNormalizedTitles;
  36
  37         public function __construct($query) {
  38                 parent :: __construct($query, __CLASS__);
  39
  40                 $this->mAllPages = array ();
  41                 $this->mGoodTitles = array ();
  42                 $this->mMissingTitles = array ();
  43                 $this->mRedirectTitles = array ();
  44                 $this->mNormalizedTitles = array ();
  45         }
  46
  47         /**
  48          * Title objects that were found in the database.
  49          * @return array page_id (int) => Title (obj)
  50          */
  51         public function getGoodTitles() {
  52                 return $this->mGoodTitles;
  53         }
  54
  55         /**
  56          * Title objects that were NOT found in the database.
  57          * @return array of Title objects
  58          */
  59         public function getMissingTitles() {
  60                 return $this->mMissingTitles;
  61         }
  62
  63         /**
  64          * Get a list of redirects when doing redirect resolution
  65          * @return array prefixed_title (string) => prefixed_title (string)
  66          */
  67         public function getRedirectTitles() {
  68                 return $this->mRedirectTitles;
  69         }
  70
  71         /**
  72          * Get a list of title normalizations - maps the title given
  73          * with its normalized version.
  74          * @return array raw_prefixed_title (string) => prefixed_title (string)
  75          */
  76         public function getNormalizedTitles() {
  77                 return $this->mNormalizedTitles;
  78         }
  79
  80         /**
  81          * Returns the number of unique pages (not revisions) in the set.
  82          */
  83         public function getGoodTitleCount() {
  84                 return count($this->getGoodTitles());
  85         }
  86
  87         /**
  88          * Get the list of revision IDs (requested with revids= parameter)
  89          */
  90         public function getRevisionIDs() {
  91                 $this->dieUsage(__FUNCTION__ . ' is not implemented', 'notimplemented');
  92         }
  93
  94         /**
  95          * Returns the number of revisions (requested with revids= parameter)
  96          */
  97         public function getRevisionCount() {
  98                 return 0; // TODO: implement
  99         }
 100
 101         /**
 102          * This method populates internal variables with page information
 103          * based on the given array of title strings.
 104          *
 105          * Steps:
 106          * #1 For each title, get data from `page` table
 107          * #2 If page was not found in the DB, store it as missing
 108          *
 109          * Additionally, when resolving redirects:
 110          * #3 If no more redirects left, stop.
 111          * #4 For each redirect, get its links from `pagelinks` table.
 112          * #5 Substitute the original LinkBatch object with the new list
 113          * #6 Repeat from step #1
 114          */
 115         private function populateTitles($titles, $redirects) {
 116
 117                 $pageFlds = array (
 118                         'page_id',
 119                         'page_namespace',
 120                         'page_title'
 121                 );
 122                 if ($redirects) {
 123                         $pageFlds[] = 'page_is_redirect';
 124                 }
 125
 126                 // Get validated and normalized title objects
 127                 $linkBatch = $this->processTitlesStrArray($titles);
 128
 129                 $db = $this->getDB();
 130
 131                 //
 132                 // Repeat until all redirects have been resolved
 133                 //
 134                 while (false !== ($set = $linkBatch->constructSet('page', $db))) {
 135
 136                         // Hack: get the ns:titles stored in array(ns => array(titles)) format
 137                         $remaining = $linkBatch->data;
 138
 139                         $redirectIds = array ();
 140
 141                         //
 142                         // Get data about $linkBatch from `page` table
 143                         //
 144                         $this->profileDBIn();
 145                         $res = $db->select('page', $pageFlds, $set, __CLASS__ . '::' . __FUNCTION__);
 146                         $this->profileDBOut();
 147                         while ($row = $db->fetchObject($res)) {
 148
 149                                 unset ($remaining[$row->page_namespace][$row->page_title]);
 150                                 $title = Title :: makeTitle($row->page_namespace, $row->page_title);
 151                                 $this->mAllPages[$row->page_namespace][$row->page_title] = $row->page_id;
 152
 153                                 if ($redirects && $row->page_is_redirect == '1') {
 154                                         $redirectIds[$row->page_id] = $title;
 155                                 } else {
 156                                         $this->mGoodTitles[$row->page_id] = $title;
 157                                 }
 158                         }
 159                         $db->freeResult($res);
 160
 161                         //
 162                         // The remaining titles in $remaining are non-existant pages
 163                         //
 164                         foreach ($remaining as $ns => $dbkeys) {
 165                                 foreach ($dbkeys as $dbkey => $nothing) {
 166                                         $this->mMissingTitles[] = Title :: makeTitle($ns, $dbkey);
 167                                         $this->mAllPages[$ns][$dbkey] = 0;
 168                                 }
 169                         }
 170
 171                         if (!$redirects || empty ($redirectIds))
 172                                 break;
 173
 174                         //
 175                         // Resolve redirects by querying the pagelinks table, and repeat the process
 176                         //
 177
 178                         // Create a new linkBatch object for the next pass
 179                         $linkBatch = new LinkBatch();
 180
 181                         // find redirect targets for all redirect pages
 182                         $this->profileDBIn();
 183                         $res = $db->select('pagelinks', array (
 184                                 'pl_from',
 185                                 'pl_namespace',
 186                                 'pl_title'
 187                         ), array (
 188                                 'pl_from' => array_keys($redirectIds
 189                         )), __CLASS__ . '::' . __FUNCTION__);
 190                         $this->profileDBOut();
 191
 192                         while ($row = $db->fetchObject($res)) {
 193
 194                                 // Bug 7304 workaround
 195                                 // ( http://bugzilla.wikipedia.org/show_bug.cgi?id=7304 )
 196                                 // A redirect page may have more than one link.
 197                                 // This code will only use the first link returned.
 198                                 if (isset ($redirectIds[$row->pl_from])) { // remove line when 7304 is fixed
 199
 200                                         $titleStrFrom = $redirectIds[$row->pl_from]->getPrefixedText();
 201                                         $titleStrTo = Title :: makeTitle($row->pl_namespace, $row->pl_title)->getPrefixedText();
 202                                         $this->mRedirectTitles[$titleStrFrom] = $titleStrTo;
 203
 204                                         unset ($redirectIds[$row->pl_from]); // remove line when 7304 is fixed
 205
 206                                         // Avoid an infinite loop by checking if we have already processed this target
 207                                         if (!isset ($this->mAllPages[$row->pl_namespace][$row->pl_title])) {
 208                                                 $linkBatch->add($row->pl_namespace, $row->pl_title);
 209                                         }
 210                                 }
 211                         }
 212                         $db->freeResult($res);
 213                 }
 214         }
 215
 216         /**
 217          * Given an array of title strings, convert them into Title objects.
 218          * This method validates access rights for the title,
 219          * and appends normalization values to the output.
 220          *
 221          * @return LinkBatch of title objects.
 222          */
 223         private function processTitlesStrArray($titles) {
 224
 225                 $linkBatch = new LinkBatch();
 226
 227                 foreach ($titles as $titleString) {
 228                         $titleObj = Title :: newFromText($titleString);
 229
 230                         // Validation
 231                         if (!$titleObj)
 232                                 $this->dieUsage("bad title $titleString", 'invalidtitle');
 233                         if ($titleObj->getNamespace() < 0)
 234                                 $this->dieUsage("No support for special page $titleString has been implemented", 'unsupportednamespace');
 235                         if (!$titleObj->userCanRead())
 236                                 $this->dieUsage("No read permission for $titleString", 'titleaccessdenied');
 237
 238                         $linkBatch->addObj($titleObj);
 239
 240                         // Make sure we remember the original title that was given to us
 241                         // This way the caller can correlate new titles with the originally requested,
 242                         // i.e. namespace is localized or capitalization is different
 243                         if ($titleString !== $titleObj->getPrefixedText()) {
 244                                 $this->mNormalizedTitles[$titleString] = $titleObj->getPrefixedText();
 245                         }
 246                 }
 247
 248                 return $linkBatch;
 249         }
 250
 251         private function populatePageIDs($pageids) {
 252                 $this->dieUsage(__FUNCTION__ . ' is not implemented', 'notimplemented');
 253         }
 254
 255         public function execute() {
 256                 $titles = $pageids = $revids = $redirects = null;
 257                 extract($this->extractRequestParams());
 258
 259                 // Only one of the titles/pageids/revids is allowed at the same time
 260                 $dataSource = null;
 261                 if (isset ($titles))
 262                         $dataSource = 'titles';
 263                 if (isset ($pageids)) {
 264                         if (isset ($dataSource))
 265                                 $this->dieUsage("Cannot use 'pageids' at the same time as '$dataSource'", 'multisource');
 266                         $dataSource = 'pageids';
 267                 }
 268                 if (isset ($revids)) {
 269                         if (isset ($dataSource))
 270                                 $this->dieUsage("Cannot use 'revids' at the same time as '$dataSource'", 'multisource');
 271                         $dataSource = 'revids';
 272                 }
 273
 274                 switch ($dataSource) {
 275                         case 'titles' :
 276                                 $this->populateTitles($titles, $redirects);
 277                                 break;
 278                         case 'pageids' :
 279                                 $this->populatePageIDs($pageids, $redirects);
 280                                 break;
 281                         case 'revids' :
 282                                 $this->populateRevIDs($revids);
 283                                 break;
 284                         default :
 285                                 // Do nothing - some queries do not need any of the data sources.
 286                                 break;
 287                 }
 288         }
 289
 290         protected function getAllowedParams() {
 291                 return array (
 292                         'titles' => array (
 293                                 GN_ENUM_ISMULTI => true
 294                         ),
 295                         'pageids' => array (
 296                                 GN_ENUM_TYPE => 'integer',
 297                                 GN_ENUM_ISMULTI => true
 298                         ),
 299                         'revids' => array (
 300                                 GN_ENUM_TYPE => 'integer',
 301                                 GN_ENUM_ISMULTI => true
 302                         ),
 303                         'redirects' => false
 304                 );
 305         }
 306
 307         protected function getParamDescription() {
 308                 return array (
 309                         'titles' => 'A list of titles to work on',
 310                         'pageids' => 'A list of page IDs to work on',
 311                         'revids' => 'A list of revision IDs to work on',
 312                         'redirects' => 'Automatically resolve redirects'
 313                 );
 314         }
 315 }
 316 ?>