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  * @deprecated since 1.31
  47  */
  48 class CategoryFinder {
  49         /** @var int[] The original article IDs passed to the seed function */
  50         protected $articles = [];
  51
  52         /** @var array Array of DBKEY category names for categories that don't have a page */
  53         protected $deadend = [];
  54
  55         /** @var array Array of [ ID => [] ] */
  56         protected $parents = [];
  57
  58         /** @var array Array of article/category IDs */
  59         protected $next = [];
  60
  61         /** @var int Max layer depth **/
  62         protected $maxdepth = -1;
  63
  64         /** @var array Array of DBKEY category names */
  65         protected $targets = [];
  66
  67         /** @var array */
  68         protected $name2id = [];
  69
  70         /** @var string "AND" or "OR" */
  71         protected $mode;
  72
  73         /** @var IDatabase Read-DB replica DB */
  74         protected $dbr;
  75
  76         /**
  77          * Initializes the instance. Do this prior to calling run().
  78          * @param array $articleIds Array of article IDs
  79          * @param array $categories FIXME
  80          * @param string $mode FIXME, default 'AND'.
  81          * @param int $maxdepth Maximum layer depth. Where:
  82          *      -1 means deep recursion (default);
  83          *       0 means no-parents;
  84          *       1 means one parent layer, etc.
  85          * @todo FIXME: $categories/$mode
  86          */
  87         public function seed( $articleIds, $categories, $mode = 'AND', $maxdepth = -1 ) {
  88                 $this->articles = $articleIds;
  89                 $this->next = $articleIds;
  90                 $this->mode = $mode;
  91                 $this->maxdepth = $maxdepth;
  92
  93                 # Set the list of target categories; convert them to DBKEY form first
  94                 $this->targets = [];
  95                 foreach ( $categories as $c ) {
  96                         $ct = Title::makeTitleSafe( NS_CATEGORY, $c );
  97                         if ( $ct ) {
  98                                 $c = $ct->getDBkey();
  99                                 $this->targets[$c] = $c;
 100                         }
 101                 }
 102         }
 103
 104         /**
 105          * Iterates through the parent tree starting with the seed values,
 106          * then checks the articles if they match the conditions
 107          * @return array Array of page_ids (those given to seed() that match the conditions)
 108          */
 109         public function run() {
 110                 $this->dbr = wfGetDB( DB_REPLICA );
 111
 112                 $i = 0;
 113                 $dig = true;
 114                 while ( count( $this->next ) && $dig ) {
 115                         $this->scanNextLayer();
 116
 117                         // Is there any depth limit?
 118                         if ( $this->maxdepth !== -1 ) {
 119                                 $dig = $i < $this->maxdepth;
 120                                 $i++;
 121                         }
 122                 }
 123
 124                 # Now check if this applies to the individual articles
 125                 $ret = [];
 126
 127                 foreach ( $this->articles as $article ) {
 128                         $conds = $this->targets;
 129                         if ( $this->check( $article, $conds ) ) {
 130                                 # Matches the conditions
 131                                 $ret[] = $article;
 132                         }
 133                 }
 134                 return $ret;
 135         }
 136
 137         /**
 138          * Get the parents. Only really useful if run() has been called already
 139          * @return array
 140          */
 141         public function getParents() {
 142                 return $this->parents;
 143         }
 144
 145         /**
 146          * This functions recurses through the parent representation, trying to match the conditions
 147          * @param int $id The article/category to check
 148          * @param array $conds The array of categories to match
 149          * @param array $path Used to check for recursion loops
 150          * @return bool Does this match the conditions?
 151          */
 152         private function check( $id, &$conds, $path = [] ) {
 153                 // Check for loops and stop!
 154                 if ( in_array( $id, $path ) ) {
 155                         return false;
 156                 }
 157
 158                 $path[] = $id;
 159
 160                 # Shortcut (runtime paranoia): No conditions=all matched
 161                 if ( count( $conds ) == 0 ) {
 162                         return true;
 163                 }
 164
 165                 if ( !isset( $this->parents[$id] ) ) {
 166                         return false;
 167                 }
 168
 169                 # iterate through the parents
 170                 foreach ( $this->parents[$id] as $p ) {
 171                         $pname = $p->cl_to;
 172
 173                         # Is this a condition?
 174                         if ( isset( $conds[$pname] ) ) {
 175                                 # This key is in the category list!
 176                                 if ( $this->mode == 'OR' ) {
 177                                         # One found, that's enough!
 178                                         $conds = [];
 179                                         return true;
 180                                 } else {
 181                                         # Assuming "AND" as default
 182                                         unset( $conds[$pname] );
 183                                         if ( count( $conds ) == 0 ) {
 184                                                 # All conditions met, done
 185                                                 return true;
 186                                         }
 187                                 }
 188                         }
 189
 190                         # Not done yet, try sub-parents
 191                         if ( !isset( $this->name2id[$pname] ) ) {
 192                                 # No sub-parent
 193                                 continue;
 194                         }
 195                         $done = $this->check( $this->name2id[$pname], $conds, $path );
 196                         if ( $done || count( $conds ) == 0 ) {
 197                                 # Subparents have done it!
 198                                 return true;
 199                         }
 200                 }
 201                 return false;
 202         }
 203
 204         /**
 205          * Scans a "parent layer" of the articles/categories in $this->next
 206          */
 207         private function scanNextLayer() {
 208                 # Find all parents of the article currently in $this->next
 209                 $layer = [];
 210                 $res = $this->dbr->select(
 211                         /* FROM   */ 'categorylinks',
 212                         /* SELECT */ [ 'cl_to', 'cl_from' ],
 213                         /* WHERE  */ [ 'cl_from' => $this->next ],
 214                         __METHOD__ . '-1'
 215                 );
 216                 foreach ( $res as $row ) {
 217                         $k = $row->cl_to;
 218
 219                         # Update parent tree
 220                         if ( !isset( $this->parents[$row->cl_from] ) ) {
 221                                 $this->parents[$row->cl_from] = [];
 222                         }
 223                         $this->parents[$row->cl_from][$k] = $row;
 224
 225                         # Ignore those we already have
 226                         if ( in_array( $k, $this->deadend ) ) {
 227                                 continue;
 228                         }
 229
 230                         if ( isset( $this->name2id[$k] ) ) {
 231                                 continue;
 232                         }
 233
 234                         # Hey, new category!
 235                         $layer[$k] = $k;
 236                 }
 237
 238                 $this->next = [];
 239
 240                 # Find the IDs of all category pages in $layer, if they exist
 241                 if ( count( $layer ) > 0 ) {
 242                         $res = $this->dbr->select(
 243                                 /* FROM   */ 'page',
 244                                 /* SELECT */ [ 'page_id', 'page_title' ],
 245                                 /* WHERE  */ [ 'page_namespace' => NS_CATEGORY, 'page_title' => $layer ],
 246                                 __METHOD__ . '-2'
 247                         );
 248                         foreach ( $res as $row ) {
 249                                 $id = $row->page_id;
 250                                 $name = $row->page_title;
 251                                 $this->name2id[$name] = $id;
 252                                 $this->next[] = $id;
 253                                 unset( $layer[$name] );
 254                         }
 255                 }
 256
 257                 # Mark dead ends
 258                 foreach ( $layer as $v ) {
 259                         $this->deadend[$v] = $v;
 260                 }
 261         }
 262 }