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 use Wikimedia\Rdbms\IDatabase;
  24
  25 /**
  26  * The "CategoryFinder" class takes a list of articles, creates an internal
  27  * representation of all their parent categories (as well as parents of
  28  * parents etc.). From this representation, it determines which of these
  29  * articles are in one or all of a given subset of categories.
  30  *
  31  * Example use :
  32  * @code
  33  *     # Determines whether the article with the page_id 12345 is in both
  34  *     # "Category 1" and "Category 2" or their subcategories, respectively
  35  *
  36  *     $cf = new CategoryFinder;
  37  *     $cf->seed(
  38  *         [ 12345 ],
  39  *         [ 'Category 1', 'Category 2' ],
  40  *         'AND'
  41  *     );
  42  *     $a = $cf->run();
  43  *     print implode( ',' , $a );
  44  * @endcode
  45  */
  46 class CategoryFinder {
  47         /** @var int[] The original article IDs passed to the seed function */
  48         protected $articles = [];
  49
  50         /** @var array Array of DBKEY category names for categories that don't have a page */
  51         protected $deadend = [];
  52
  53         /** @var array Array of [ ID => [] ] */
  54         protected $parents = [];
  55
  56         /** @var array Array of article/category IDs */
  57         protected $next = [];
  58
  59         /** @var int Max layer depth **/
  60         protected $maxdepth = -1;
  61
  62         /** @var array Array of DBKEY category names */
  63         protected $targets = [];
  64
  65         /** @var array */
  66         protected $name2id = [];
  67
  68         /** @var string "AND" or "OR" */
  69         protected $mode;
  70
  71         /** @var IDatabase Read-DB replica DB */
  72         protected $dbr;
  73
  74         /**
  75          * Initializes the instance. Do this prior to calling run().
  76          * @param array $articleIds Array of article IDs
  77          * @param array $categories FIXME
  78          * @param string $mode FIXME, default 'AND'.
  79          * @param int $maxdepth Maximum layer depth. Where:
  80          *      -1 means deep recursion (default);
  81          *       0 means no-parents;
  82          *       1 means one parent layer, etc.
  83          * @todo FIXME: $categories/$mode
  84          */
  85         public function seed( $articleIds, $categories, $mode = 'AND', $maxdepth = -1 ) {
  86                 $this->articles = $articleIds;
  87                 $this->next = $articleIds;
  88                 $this->mode = $mode;
  89                 $this->maxdepth = $maxdepth;
  90
  91                 # Set the list of target categories; convert them to DBKEY form first
  92                 $this->targets = [];
  93                 foreach ( $categories as $c ) {
  94                         $ct = Title::makeTitleSafe( NS_CATEGORY, $c );
  95                         if ( $ct ) {
  96                                 $c = $ct->getDBkey();
  97                                 $this->targets[$c] = $c;
  98                         }
  99                 }
 100         }
 101
 102         /**
 103          * Iterates through the parent tree starting with the seed values,
 104          * then checks the articles if they match the conditions
 105          * @return array Array of page_ids (those given to seed() that match the conditions)
 106          */
 107         public function run() {
 108                 $this->dbr = wfGetDB( DB_REPLICA );
 109
 110                 $i = 0;
 111                 $dig = true;
 112                 while ( count( $this->next ) && $dig ) {
 113                         $this->scanNextLayer();
 114
 115                         // Is there any depth limit?
 116                         if ( $this->maxdepth !== -1 ) {
 117                                 $dig = $i < $this->maxdepth;
 118                                 $i++;
 119                         }
 120                 }
 121
 122                 # Now check if this applies to the individual articles
 123                 $ret = [];
 124
 125                 foreach ( $this->articles as $article ) {
 126                         $conds = $this->targets;
 127                         if ( $this->check( $article, $conds ) ) {
 128                                 # Matches the conditions
 129                                 $ret[] = $article;
 130                         }
 131                 }
 132                 return $ret;
 133         }
 134
 135         /**
 136          * Get the parents. Only really useful if run() has been called already
 137          * @return array
 138          */
 139         public function getParents() {
 140                 return $this->parents;
 141         }
 142
 143         /**
 144          * This functions recurses through the parent representation, trying to match the conditions
 145          * @param int $id The article/category to check
 146          * @param array $conds The array of categories to match
 147          * @param array $path Used to check for recursion loops
 148          * @return bool Does this match the conditions?
 149          */
 150         private function check( $id, &$conds, $path = [] ) {
 151                 // Check for loops and stop!
 152                 if ( in_array( $id, $path ) ) {
 153                         return false;
 154                 }
 155
 156                 $path[] = $id;
 157
 158                 # Shortcut (runtime paranoia): No conditions=all matched
 159                 if ( count( $conds ) == 0 ) {
 160                         return true;
 161                 }
 162
 163                 if ( !isset( $this->parents[$id] ) ) {
 164                         return false;
 165                 }
 166
 167                 # iterate through the parents
 168                 foreach ( $this->parents[$id] as $p ) {
 169                         $pname = $p->cl_to;
 170
 171                         # Is this a condition?
 172                         if ( isset( $conds[$pname] ) ) {
 173                                 # This key is in the category list!
 174                                 if ( $this->mode == 'OR' ) {
 175                                         # One found, that's enough!
 176                                         $conds = [];
 177                                         return true;
 178                                 } else {
 179                                         # Assuming "AND" as default
 180                                         unset( $conds[$pname] );
 181                                         if ( count( $conds ) == 0 ) {
 182                                                 # All conditions met, done
 183                                                 return true;
 184                                         }
 185                                 }
 186                         }
 187
 188                         # Not done yet, try sub-parents
 189                         if ( !isset( $this->name2id[$pname] ) ) {
 190                                 # No sub-parent
 191                                 continue;
 192                         }
 193                         $done = $this->check( $this->name2id[$pname], $conds, $path );
 194                         if ( $done || count( $conds ) == 0 ) {
 195                                 # Subparents have done it!
 196                                 return true;
 197                         }
 198                 }
 199                 return false;
 200         }
 201
 202         /**
 203          * Scans a "parent layer" of the articles/categories in $this->next
 204          */
 205         private function scanNextLayer() {
 206                 # Find all parents of the article currently in $this->next
 207                 $layer = [];
 208                 $res = $this->dbr->select(
 209                         /* FROM   */ 'categorylinks',
 210                         /* SELECT */ [ 'cl_to', 'cl_from' ],
 211                         /* WHERE  */ [ 'cl_from' => $this->next ],
 212                         __METHOD__ . '-1'
 213                 );
 214                 foreach ( $res as $o ) {
 215                         $k = $o->cl_to;
 216
 217                         # Update parent tree
 218                         if ( !isset( $this->parents[$o->cl_from] ) ) {
 219                                 $this->parents[$o->cl_from] = [];
 220                         }
 221                         $this->parents[$o->cl_from][$k] = $o;
 222
 223                         # Ignore those we already have
 224                         if ( in_array( $k, $this->deadend ) ) {
 225                                 continue;
 226                         }
 227
 228                         if ( isset( $this->name2id[$k] ) ) {
 229                                 continue;
 230                         }
 231
 232                         # Hey, new category!
 233                         $layer[$k] = $k;
 234                 }
 235
 236                 $this->next = [];
 237
 238                 # Find the IDs of all category pages in $layer, if they exist
 239                 if ( count( $layer ) > 0 ) {
 240                         $res = $this->dbr->select(
 241                                 /* FROM   */ 'page',
 242                                 /* SELECT */ [ 'page_id', 'page_title' ],
 243                                 /* WHERE  */ [ 'page_namespace' => NS_CATEGORY, 'page_title' => $layer ],
 244                                 __METHOD__ . '-2'
 245                         );
 246                         foreach ( $res as $o ) {
 247                                 $id = $o->page_id;
 248                                 $name = $o->page_title;
 249                                 $this->name2id[$name] = $id;
 250                                 $this->next[] = $id;
 251                                 unset( $layer[$name] );
 252                         }
 253                 }
 254
 255                 # Mark dead ends
 256                 foreach ( $layer as $v ) {
 257                         $this->deadend[$v] = $v;
 258                 }
 259         }
 260 }