includes/Categoryfinder.php

   1 <?php
   2 /**
   3  * Recent changes filtering by category.
   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  */
  22
  23 /**
  24  * The "Categoryfinder" class takes a list of articles, creates an internal
  25  * representation of all their parent categories (as well as parents of
  26  * parents etc.). From this representation, it determines which of these
  27  * articles are in one or all of a given subset of categories.
  28  *
  29  * Example use :
  30  * <code>
  31  *     # Determines whether the article with the page_id 12345 is in both
  32  *     # "Category 1" and "Category 2" or their subcategories, respectively
  33  *
  34  *     $cf = new Categoryfinder;
  35  *     $cf->seed(
  36  *         array( 12345 ),
  37  *         array( 'Category 1', 'Category 2' ),
  38  *         'AND'
  39  *     );
  40  *     $a = $cf->run();
  41  *     print implode( ',' , $a );
  42  * </code>
  43  *
  44  */
  45 class Categoryfinder {
  46         /** @var int[] The original article IDs passed to the seed function */
  47         protected $articles = array();
  48
  49         /** @var array Array of DBKEY category names for categories that don't have a page */
  50         protected $deadend = array();
  51
  52         /** @var array Array of [ID => array()] */
  53         protected $parents = array();
  54
  55         /** @var array Array of article/category IDs */
  56         protected $next = array();
  57
  58         /** @var array Array of DBKEY category names */
  59         protected $targets = array();
  60
  61         /** @var array */
  62         protected $name2id = array();
  63
  64         /** @var  "AND" or "OR" */
  65         protected $mode;
  66
  67         /** @var DatabaseBase Read-DB slave */
  68         protected $dbr;
  69
  70         function __construct() {
  71         }
  72
  73         /**
  74          * Initializes the instance. Do this prior to calling run().
  75          * @param array $article_ids Array of article IDs
  76          * @param array $categories FIXME
  77          * @param string $mode FIXME, default 'AND'.
  78          * @todo FIXME: $categories/$mode
  79          */
  80         function seed( $article_ids, $categories, $mode = 'AND' ) {
  81                 $this->articles = $article_ids;
  82                 $this->next = $article_ids;
  83                 $this->mode = $mode;
  84
  85                 # Set the list of target categories; convert them to DBKEY form first
  86                 $this->targets = array();
  87                 foreach ( $categories as $c ) {
  88                         $ct = Title::makeTitleSafe( NS_CATEGORY, $c );
  89                         if ( $ct ) {
  90                                 $c = $ct->getDBkey();
  91                                 $this->targets[$c] = $c;
  92                         }
  93                 }
  94         }
  95
  96         /**
  97          * Iterates through the parent tree starting with the seed values,
  98          * then checks the articles if they match the conditions
  99          * @return array of page_ids (those given to seed() that match the conditions)
 100          */
 101         function run() {
 102                 $this->dbr = wfGetDB( DB_SLAVE );
 103                 while ( count( $this->next ) > 0 ) {
 104                         $this->scan_next_layer();
 105                 }
 106
 107                 # Now check if this applies to the individual articles
 108                 $ret = array();
 109
 110                 foreach ( $this->articles as $article ) {
 111                         $conds = $this->targets;
 112                         if ( $this->check( $article, $conds ) ) {
 113                                 # Matches the conditions
 114                                 $ret[] = $article;
 115                         }
 116                 }
 117                 return $ret;
 118         }
 119
 120         /**
 121          * This functions recurses through the parent representation, trying to match the conditions
 122          * @param int $id The article/category to check
 123          * @param array $conds The array of categories to match
 124          * @param array $path Used to check for recursion loops
 125          * @return bool Does this match the conditions?
 126          */
 127         function check( $id, &$conds, $path = array() ) {
 128                 // Check for loops and stop!
 129                 if ( in_array( $id, $path ) ) {
 130                         return false;
 131                 }
 132
 133                 $path[] = $id;
 134
 135                 # Shortcut (runtime paranoia): No conditions=all matched
 136                 if ( count( $conds ) == 0 ) {
 137                         return true;
 138                 }
 139
 140                 if ( !isset( $this->parents[$id] ) ) {
 141                         return false;
 142                 }
 143
 144                 # iterate through the parents
 145                 foreach ( $this->parents[$id] as $p ) {
 146                         $pname = $p->cl_to;
 147
 148                         # Is this a condition?
 149                         if ( isset( $conds[$pname] ) ) {
 150                                 # This key is in the category list!
 151                                 if ( $this->mode == 'OR' ) {
 152                                         # One found, that's enough!
 153                                         $conds = array();
 154                                         return true;
 155                                 } else {
 156                                         # Assuming "AND" as default
 157                                         unset( $conds[$pname] );
 158                                         if ( count( $conds ) == 0 ) {
 159                                                 # All conditions met, done
 160                                                 return true;
 161                                         }
 162                                 }
 163                         }
 164
 165                         # Not done yet, try sub-parents
 166                         if ( !isset( $this->name2id[$pname] ) ) {
 167                                 # No sub-parent
 168                                 continue;
 169                         }
 170                         $done = $this->check( $this->name2id[$pname], $conds, $path );
 171                         if ( $done || count( $conds ) == 0 ) {
 172                                 # Subparents have done it!
 173                                 return true;
 174                         }
 175                 }
 176                 return false;
 177         }
 178
 179         /**
 180          * Scans a "parent layer" of the articles/categories in $this->next
 181          */
 182         function scan_next_layer() {
 183                 wfProfileIn( __METHOD__ );
 184
 185                 # Find all parents of the article currently in $this->next
 186                 $layer = array();
 187                 $res = $this->dbr->select(
 188                         /* FROM   */ 'categorylinks',
 189                         /* SELECT */ '*',
 190                         /* WHERE  */ array( 'cl_from' => $this->next ),
 191                         __METHOD__ . '-1'
 192                 );
 193                 foreach ( $res as $o ) {
 194                         $k = $o->cl_to;
 195
 196                         # Update parent tree
 197                         if ( !isset( $this->parents[$o->cl_from] ) ) {
 198                                 $this->parents[$o->cl_from] = array();
 199                         }
 200                         $this->parents[$o->cl_from][$k] = $o;
 201
 202                         # Ignore those we already have
 203                         if ( in_array( $k, $this->deadend ) ) {
 204                                 continue;
 205                         }
 206
 207                         if ( isset( $this->name2id[$k] ) ) {
 208                                 continue;
 209                         }
 210
 211                         # Hey, new category!
 212                         $layer[$k] = $k;
 213                 }
 214
 215                 $this->next = array();
 216
 217                 # Find the IDs of all category pages in $layer, if they exist
 218                 if ( count( $layer ) > 0 ) {
 219                         $res = $this->dbr->select(
 220                                 /* FROM   */ 'page',
 221                                 /* SELECT */ array( 'page_id', 'page_title' ),
 222                                 /* WHERE  */ array( 'page_namespace' => NS_CATEGORY, 'page_title' => $layer ),
 223                                 __METHOD__ . '-2'
 224                         );
 225                         foreach ( $res as $o ) {
 226                                 $id = $o->page_id;
 227                                 $name = $o->page_title;
 228                                 $this->name2id[$name] = $id;
 229                                 $this->next[] = $id;
 230                                 unset( $layer[$name] );
 231                         }
 232                 }
 233
 234                 # Mark dead ends
 235                 foreach ( $layer as $v ) {
 236                         $this->deadend[$v] = $v;
 237                 }
 238
 239                 wfProfileOut( __METHOD__ );
 240         }
 241 }