includes/specials/SpecialRandomInCategory.php

   1 <?php
   2 /**
   3  * Implements Special:RandomInCategory
   4  *
   5  * This program is free software; you can redistribute it and/or modify
   6  * it under the terms of the GNU General Public License as published by
   7  * the Free Software Foundation; either version 2 of the License, or
   8  * (at your option) any later version.
   9  *
  10  * This program is distributed in the hope that it will be useful,
  11  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  12  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
  13  * GNU General Public License for more details.
  14  *
  15  * You should have received a copy of the GNU General Public License along
  16  * with this program; if not, write to the Free Software Foundation, Inc.,
  17  * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
  18  * http://www.gnu.org/copyleft/gpl.html
  19  *
  20  * @file
  21  * @ingroup SpecialPage
  22  * @author Brian Wolff
  23  */
  24
  25 /**
  26  * Special page to direct the user to a random page
  27  *
  28  * @note The method used here is rather biased. It is assumed that
  29  * the use of this page will be people wanting to get a random page
  30  * out of a maintenance category, to fix it up. The method used by
  31  * this page should return different pages in an unpredictable fashion
  32  * which is hoped to be sufficient, even if some pages are selected
  33  * more often than others.
  34  *
  35  * A more unbiased method could be achieved by adding a cl_random field
  36  * to the categorylinks table.
  37  *
  38  * The method used here is as follows:
  39  *  * Find the smallest and largest timestamp in the category
  40  *  * Pick a random timestamp in between
  41  *  * Pick an offset between 0 and 30
  42  *  * Get the offset'ed page that is newer than the timestamp selected
  43  * The offset is meant to counter the fact the timestamps aren't usually
  44  * uniformly distributed, so if things are very non-uniform at least we
  45  * won't have the same page selected 99% of the time.
  46  *
  47  * @ingroup SpecialPage
  48  */
  49 class SpecialRandomInCategory extends SpecialPage {
  50         protected $extra = array(); // Extra SQL statements
  51         protected $category = false; // Title object of category
  52         protected $maxOffset = 30; // Max amount to fudge randomness by.
  53         private $maxTimestamp = null;
  54         private $minTimestamp = null;
  55
  56         public function __construct( $name = 'RandomInCategory' ) {
  57                 parent::__construct( $name );
  58         }
  59
  60         /**
  61          * Set which category to use.
  62          * @param Title $cat
  63          */
  64         public function setCategory( Title $cat ) {
  65                 $this->category = $cat;
  66                 $this->maxTimestamp = null;
  67                 $this->minTimestamp = null;
  68         }
  69
  70         public function execute( $par ) {
  71                 $cat = false;
  72
  73                 $categoryStr = $this->getRequest()->getText( 'category', $par );
  74
  75                 if ( $categoryStr ) {
  76                         $cat = Title::newFromText( $categoryStr, NS_CATEGORY );
  77                 }
  78
  79                 if ( $cat && $cat->getNamespace() !== NS_CATEGORY ) {
  80                         // Someone searching for something like "Wikipedia:Foo"
  81                         $cat = Title::makeTitleSafe( NS_CATEGORY, $categoryStr );
  82                 }
  83
  84                 if ( $cat ) {
  85                         $this->setCategory( $cat );
  86                 }
  87
  88                 if ( !$this->category && $categoryStr ) {
  89                         $this->setHeaders();
  90                         $this->getOutput()->addWikiMsg( 'randomincategory-invalidcategory',
  91                                 wfEscapeWikiText( $categoryStr ) );
  92
  93                         return;
  94                 } elseif ( !$this->category ) {
  95                         $this->setHeaders();
  96                         $input = Html::input( 'category' );
  97                         $submitText = $this->msg( 'randomincategory-selectcategory-submit' )->text();
  98                         $submit = Html::input( '', $submitText, 'submit' );
  99
 100                         $msg = $this->msg( 'randomincategory-selectcategory' );
 101                         $form = Html::rawElement( 'form', array( 'action' => wfScript() ),
 102                                 Html::hidden( 'title', $this->getPageTitle()->getPrefixedText() ) .
 103                                 $msg->rawParams( $input, $submit )->parse()
 104                         );
 105                         $this->getOutput()->addHtml( $form );
 106
 107                         return;
 108                 }
 109
 110                 $title = $this->getRandomTitle();
 111
 112                 if ( is_null( $title ) ) {
 113                         $this->setHeaders();
 114                         $this->getOutput()->addWikiMsg( 'randomincategory-nopages',
 115                                 $this->category->getText() );
 116
 117                         return;
 118                 }
 119
 120                 $query = $this->getRequest()->getValues();
 121                 unset( $query['title'] );
 122                 unset( $query['category'] );
 123                 $this->getOutput()->redirect( $title->getFullURL( $query ) );
 124         }
 125
 126         /**
 127          * Choose a random title.
 128          * @return Title|null Title object (or null if nothing to choose from)
 129          */
 130         public function getRandomTitle() {
 131                 // Convert to float, since we do math with the random number.
 132                 $rand = (float)wfRandom();
 133                 $title = null;
 134
 135                 // Given that timestamps are rather unevenly distributed, we also
 136                 // use an offset between 0 and 30 to make any biases less noticeable.
 137                 $offset = mt_rand( 0, $this->maxOffset );
 138
 139                 if ( mt_rand( 0, 1 ) ) {
 140                         $up = true;
 141                 } else {
 142                         $up = false;
 143                 }
 144
 145                 $row = $this->selectRandomPageFromDB( $rand, $offset, $up );
 146
 147                 // Try again without the timestamp offset (wrap around the end)
 148                 if ( !$row ) {
 149                         $row = $this->selectRandomPageFromDB( false, $offset, $up );
 150                 }
 151
 152                 // Maybe the category is really small and offset too high
 153                 if ( !$row ) {
 154                         $row = $this->selectRandomPageFromDB( $rand, 0, $up );
 155                 }
 156
 157                 // Just get the first entry.
 158                 if ( !$row ) {
 159                         $row = $this->selectRandomPageFromDB( false, 0, true );
 160                 }
 161
 162                 if ( $row ) {
 163                         return Title::makeTitle( $row->page_namespace, $row->page_title );
 164                 }
 165
 166                 return null;
 167         }
 168
 169         /**
 170          * @param float $rand Random number between 0 and 1
 171          * @param int $offset Extra offset to fudge randomness
 172          * @param bool $up True to get the result above the random number, false for below
 173          *
 174          * @note The $up parameter is supposed to counteract what would happen if there
 175          *   was a large gap in the distribution of cl_timestamp values. This way instead
 176          *   of things to the right of the gap being favoured, both sides of the gap
 177          *   are favoured.
 178          * @return array Query information.
 179          */
 180         protected function getQueryInfo( $rand, $offset, $up ) {
 181                 $op = $up ? '>=' : '<=';
 182                 $dir = $up ? 'ASC' : 'DESC';
 183                 if ( !$this->category instanceof Title ) {
 184                         throw new MWException( 'No category set' );
 185                 }
 186                 $qi = array(
 187                         'tables' => array( 'categorylinks', 'page' ),
 188                         'fields' => array( 'page_title', 'page_namespace' ),
 189                         'conds' => array_merge( array(
 190                                 'cl_to' => $this->category->getDBKey(),
 191                         ), $this->extra ),
 192                         'options' => array(
 193                                 'ORDER BY' => 'cl_timestamp ' . $dir,
 194                                 'LIMIT' => 1,
 195                                 'OFFSET' => $offset
 196                         ),
 197                         'join_conds' => array(
 198                                 'page' => array( 'INNER JOIN', 'cl_from = page_id' )
 199                         )
 200                 );
 201
 202                 $dbr = wfGetDB( DB_SLAVE );
 203                 $minClTime = $this->getTimestampOffset( $rand );
 204                 if ( $minClTime ) {
 205                         $qi['conds'][] = 'cl_timestamp ' . $op . ' ' .
 206                                 $dbr->addQuotes( $dbr->timestamp( $minClTime ) );
 207                 }
 208
 209                 return $qi;
 210         }
 211
 212         /**
 213          * @param float $rand Random number between 0 and 1
 214          *
 215          * @return int|bool A random (unix) timestamp from the range of the category or false on failure
 216          */
 217         protected function getTimestampOffset( $rand ) {
 218                 if ( $rand === false ) {
 219                         return false;
 220                 }
 221                 if ( !$this->minTimestamp || !$this->maxTimestamp ) {
 222                         try {
 223                                 list( $this->minTimestamp, $this->maxTimestamp ) = $this->getMinAndMaxForCat( $this->category );
 224                         } catch ( MWException $e ) {
 225                                 // Possibly no entries in category.
 226                                 return false;
 227                         }
 228                 }
 229
 230                 $ts = ( $this->maxTimestamp - $this->minTimestamp ) * $rand + $this->minTimestamp;
 231
 232                 return intval( $ts );
 233         }
 234
 235         /**
 236          * Get the lowest and highest timestamp for a category.
 237          *
 238          * @param Title $category
 239          * @return array The lowest and highest timestamp
 240          * @throws MWException If category has no entries.
 241          */
 242         protected function getMinAndMaxForCat( Title $category ) {
 243                 $dbr = wfGetDB( DB_SLAVE );
 244                 $res = $dbr->selectRow(
 245                         'categorylinks',
 246                         array(
 247                                 'low' => 'MIN( cl_timestamp )',
 248                                 'high' => 'MAX( cl_timestamp )'
 249                         ),
 250                         array(
 251                                 'cl_to' => $this->category->getDBKey(),
 252                         ),
 253                         __METHOD__,
 254                         array(
 255                                 'LIMIT' => 1
 256                         )
 257                 );
 258                 if ( !$res ) {
 259                         throw new MWException( 'No entries in category' );
 260                 }
 261
 262                 return array( wfTimestamp( TS_UNIX, $res->low ), wfTimestamp( TS_UNIX, $res->high ) );
 263         }
 264
 265         /**
 266          * @param float $rand A random number that is converted to a random timestamp
 267          * @param int $offset A small offset to make the result seem more "random"
 268          * @param bool $up Get the result above the random value
 269          * @param string $fname The name of the calling method
 270          * @return array Info for the title selected.
 271          */
 272         private function selectRandomPageFromDB( $rand, $offset, $up, $fname = __METHOD__ ) {
 273                 $dbr = wfGetDB( DB_SLAVE );
 274
 275                 $query = $this->getQueryInfo( $rand, $offset, $up );
 276                 $res = $dbr->select(
 277                         $query['tables'],
 278                         $query['fields'],
 279                         $query['conds'],
 280                         $fname,
 281                         $query['options'],
 282                         $query['join_conds']
 283                 );
 284
 285                 return $res->fetchObject();
 286         }
 287
 288         protected function getGroupName() {
 289                 return 'redirects';
 290         }
 291 }