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