includes/Category.php

   1 <?php
   2 /**
   3  * Representation for a 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  * @author Simetrical
  22  */
  23
  24 /**
  25  * Category objects are immutable, strictly speaking. If you call methods that change the database,
  26  * like to refresh link counts, the objects will be appropriately reinitialized.
  27  * Member variables are lazy-initialized.
  28  */
  29 class Category {
  30         /** Name of the category, normalized to DB-key form */
  31         private $mName = null;
  32         private $mID = null;
  33         /**
  34          * Category page title
  35          * @var Title
  36          */
  37         private $mTitle = null;
  38         /** Counts of membership (cat_pages, cat_subcats, cat_files) */
  39         private $mPages = null, $mSubcats = null, $mFiles = null;
  40
  41         const LOAD_ONLY = 0;
  42         const LAZY_INIT_ROW = 1;
  43
  44         const ROW_COUNT_SMALL = 100;
  45
  46         private function __construct() {
  47         }
  48
  49         /**
  50          * Set up all member variables using a database query.
  51          * @param int $mode One of (Category::LOAD_ONLY, Category::LAZY_INIT_ROW)
  52          * @throws MWException
  53          * @return bool True on success, false on failure.
  54          */
  55         protected function initialize( $mode = self::LOAD_ONLY ) {
  56                 if ( $this->mName === null && $this->mID === null ) {
  57                         throw new MWException( __METHOD__ . ' has both names and IDs null' );
  58                 } elseif ( $this->mID === null ) {
  59                         $where = [ 'cat_title' => $this->mName ];
  60                 } elseif ( $this->mName === null ) {
  61                         $where = [ 'cat_id' => $this->mID ];
  62                 } else {
  63                         # Already initialized
  64                         return true;
  65                 }
  66
  67                 $dbr = wfGetDB( DB_REPLICA );
  68                 $row = $dbr->selectRow(
  69                         'category',
  70                         [ 'cat_id', 'cat_title', 'cat_pages', 'cat_subcats', 'cat_files' ],
  71                         $where,
  72                         __METHOD__
  73                 );
  74
  75                 if ( !$row ) {
  76                         # Okay, there were no contents.  Nothing to initialize.
  77                         if ( $this->mTitle ) {
  78                                 # If there is a title object but no record in the category table,
  79                                 # treat this as an empty category.
  80                                 $this->mID = false;
  81                                 $this->mName = $this->mTitle->getDBkey();
  82                                 $this->mPages = 0;
  83                                 $this->mSubcats = 0;
  84                                 $this->mFiles = 0;
  85
  86                                 # If the title exists, call refreshCounts to add a row for it.
  87                                 if ( $mode === self::LAZY_INIT_ROW && $this->mTitle->exists() ) {
  88                                         DeferredUpdates::addCallableUpdate( [ $this, 'refreshCounts' ] );
  89                                 }
  90
  91                                 return true;
  92                         } else {
  93                                 return false; # Fail
  94                         }
  95                 }
  96
  97                 $this->mID = $row->cat_id;
  98                 $this->mName = $row->cat_title;
  99                 $this->mPages = $row->cat_pages;
 100                 $this->mSubcats = $row->cat_subcats;
 101                 $this->mFiles = $row->cat_files;
 102
 103                 # (T15683) If the count is negative, then 1) it's obviously wrong
 104                 # and should not be kept, and 2) we *probably* don't have to scan many
 105                 # rows to obtain the correct figure, so let's risk a one-time recount.
 106                 if ( $this->mPages < 0 || $this->mSubcats < 0 || $this->mFiles < 0 ) {
 107                         $this->mPages = max( $this->mPages, 0 );
 108                         $this->mSubcats = max( $this->mSubcats, 0 );
 109                         $this->mFiles = max( $this->mFiles, 0 );
 110
 111                         if ( $mode === self::LAZY_INIT_ROW ) {
 112                                 DeferredUpdates::addCallableUpdate( [ $this, 'refreshCounts' ] );
 113                         }
 114                 }
 115
 116                 return true;
 117         }
 118
 119         /**
 120          * Factory function.
 121          *
 122          * @param string $name A category name (no "Category:" prefix).  It need
 123          *   not be normalized, with spaces replaced by underscores.
 124          * @return Category|bool Category, or false on a totally invalid name
 125          */
 126         public static function newFromName( $name ) {
 127                 $cat = new self();
 128                 $title = Title::makeTitleSafe( NS_CATEGORY, $name );
 129
 130                 if ( !is_object( $title ) ) {
 131                         return false;
 132                 }
 133
 134                 $cat->mTitle = $title;
 135                 $cat->mName = $title->getDBkey();
 136
 137                 return $cat;
 138         }
 139
 140         /**
 141          * Factory function.
 142          *
 143          * @param Title $title Title for the category page
 144          * @return Category|bool On a totally invalid name
 145          */
 146         public static function newFromTitle( $title ) {
 147                 $cat = new self();
 148
 149                 $cat->mTitle = $title;
 150                 $cat->mName = $title->getDBkey();
 151
 152                 return $cat;
 153         }
 154
 155         /**
 156          * Factory function.
 157          *
 158          * @param int $id A category id
 159          * @return Category
 160          */
 161         public static function newFromID( $id ) {
 162                 $cat = new self();
 163                 $cat->mID = intval( $id );
 164                 return $cat;
 165         }
 166
 167         /**
 168          * Factory function, for constructing a Category object from a result set
 169          *
 170          * @param object $row Result set row, must contain the cat_xxx fields. If the
 171          *   fields are null, the resulting Category object will represent an empty
 172          *   category if a title object was given. If the fields are null and no
 173          *   title was given, this method fails and returns false.
 174          * @param Title|null $title Optional title object for the category represented by
 175          *   the given row. May be provided if it is already known, to avoid having
 176          *   to re-create a title object later.
 177          * @return Category|false
 178          */
 179         public static function newFromRow( $row, $title = null ) {
 180                 $cat = new self();
 181                 $cat->mTitle = $title;
 182
 183                 # NOTE: the row often results from a LEFT JOIN on categorylinks. This may result in
 184                 #       all the cat_xxx fields being null, if the category page exists, but nothing
 185                 #       was ever added to the category. This case should be treated link an empty
 186                 #       category, if possible.
 187
 188                 if ( $row->cat_title === null ) {
 189                         if ( $title === null ) {
 190                                 # the name is probably somewhere in the row, for example as page_title,
 191                                 # but we can't know that here...
 192                                 return false;
 193                         } else {
 194                                 # if we have a title object, fetch the category name from there
 195                                 $cat->mName = $title->getDBkey();
 196                         }
 197
 198                         $cat->mID = false;
 199                         $cat->mSubcats = 0;
 200                         $cat->mPages = 0;
 201                         $cat->mFiles = 0;
 202                 } else {
 203                         $cat->mName = $row->cat_title;
 204                         $cat->mID = $row->cat_id;
 205                         $cat->mSubcats = $row->cat_subcats;
 206                         $cat->mPages = $row->cat_pages;
 207                         $cat->mFiles = $row->cat_files;
 208                 }
 209
 210                 return $cat;
 211         }
 212
 213         /**
 214          * @return mixed DB key name, or false on failure
 215          */
 216         public function getName() {
 217                 return $this->getX( 'mName' );
 218         }
 219
 220         /**
 221          * @return mixed Category ID, or false on failure
 222          */
 223         public function getID() {
 224                 return $this->getX( 'mID' );
 225         }
 226
 227         /**
 228          * @return mixed Total number of member pages, or false on failure
 229          */
 230         public function getPageCount() {
 231                 return $this->getX( 'mPages' );
 232         }
 233
 234         /**
 235          * @return mixed Number of subcategories, or false on failure
 236          */
 237         public function getSubcatCount() {
 238                 return $this->getX( 'mSubcats' );
 239         }
 240
 241         /**
 242          * @return mixed Number of member files, or false on failure
 243          */
 244         public function getFileCount() {
 245                 return $this->getX( 'mFiles' );
 246         }
 247
 248         /**
 249          * @return Title|bool Title for this category, or false on failure.
 250          */
 251         public function getTitle() {
 252                 if ( $this->mTitle ) {
 253                         return $this->mTitle;
 254                 }
 255
 256                 if ( !$this->initialize( self::LAZY_INIT_ROW ) ) {
 257                         return false;
 258                 }
 259
 260                 $this->mTitle = Title::makeTitleSafe( NS_CATEGORY, $this->mName );
 261                 return $this->mTitle;
 262         }
 263
 264         /**
 265          * Fetch a TitleArray of up to $limit category members, beginning after the
 266          * category sort key $offset.
 267          * @param int|bool $limit
 268          * @param string $offset
 269          * @return TitleArray TitleArray object for category members.
 270          */
 271         public function getMembers( $limit = false, $offset = '' ) {
 272                 $dbr = wfGetDB( DB_REPLICA );
 273
 274                 $conds = [ 'cl_to' => $this->getName(), 'cl_from = page_id' ];
 275                 $options = [ 'ORDER BY' => 'cl_sortkey' ];
 276
 277                 if ( $limit ) {
 278                         $options['LIMIT'] = $limit;
 279                 }
 280
 281                 if ( $offset !== '' ) {
 282                         $conds[] = 'cl_sortkey > ' . $dbr->addQuotes( $offset );
 283                 }
 284
 285                 $result = TitleArray::newFromResult(
 286                         $dbr->select(
 287                                 [ 'page', 'categorylinks' ],
 288                                 [ 'page_id', 'page_namespace', 'page_title', 'page_len',
 289                                         'page_is_redirect', 'page_latest' ],
 290                                 $conds,
 291                                 __METHOD__,
 292                                 $options
 293                         )
 294                 );
 295
 296                 return $result;
 297         }
 298
 299         /**
 300          * Generic accessor
 301          * @param string $key
 302          * @return mixed
 303          */
 304         private function getX( $key ) {
 305                 if ( $this->{$key} === null && !$this->initialize( self::LAZY_INIT_ROW ) ) {
 306                         return false;
 307                 }
 308                 return $this->{$key};
 309         }
 310
 311         /**
 312          * Refresh the counts for this category.
 313          *
 314          * @return bool True on success, false on failure
 315          */
 316         public function refreshCounts() {
 317                 if ( wfReadOnly() ) {
 318                         return false;
 319                 }
 320
 321                 # If we have just a category name, find out whether there is an
 322                 # existing row. Or if we have just an ID, get the name, because
 323                 # that's what categorylinks uses.
 324                 if ( !$this->initialize( self::LOAD_ONLY ) ) {
 325                         return false;
 326                 }
 327
 328                 $dbw = wfGetDB( DB_MASTER );
 329                 # Avoid excess contention on the same category (T162121)
 330                 $name = __METHOD__ . ':' . md5( $this->mName );
 331                 $scopedLock = $dbw->getScopedLockAndFlush( $name, __METHOD__, 0 );
 332                 if ( !$scopedLock ) {
 333                         return false;
 334                 }
 335
 336                 $dbw->startAtomic( __METHOD__ );
 337
 338                 // Lock the `category` row before locking `categorylinks` rows to try
 339                 // to avoid deadlocks with LinksDeletionUpdate (T195397)
 340                 $dbw->lockForUpdate( 'category', [ 'cat_title' => $this->mName ], __METHOD__ );
 341
 342                 // Lock all the `categorylinks` records and gaps for this category;
 343                 // this is a separate query due to postgres limitations
 344                 $dbw->selectRowCount(
 345                         [ 'categorylinks', 'page' ],
 346                         '*',
 347                         [ 'cl_to' => $this->mName, 'page_id = cl_from' ],
 348                         __METHOD__,
 349                         [ 'LOCK IN SHARE MODE' ]
 350                 );
 351                 // Get the aggregate `categorylinks` row counts for this category
 352                 $catCond = $dbw->conditional( [ 'page_namespace' => NS_CATEGORY ], 1, 'NULL' );
 353                 $fileCond = $dbw->conditional( [ 'page_namespace' => NS_FILE ], 1, 'NULL' );
 354                 $result = $dbw->selectRow(
 355                         [ 'categorylinks', 'page' ],
 356                         [
 357                                 'pages' => 'COUNT(*)',
 358                                 'subcats' => "COUNT($catCond)",
 359                                 'files' => "COUNT($fileCond)"
 360                         ],
 361                         [ 'cl_to' => $this->mName, 'page_id = cl_from' ],
 362                         __METHOD__
 363                 );
 364
 365                 $shouldExist = $result->pages > 0 || $this->getTitle()->exists();
 366
 367                 if ( $this->mID ) {
 368                         if ( $shouldExist ) {
 369                                 # The category row already exists, so do a plain UPDATE instead
 370                                 # of INSERT...ON DUPLICATE KEY UPDATE to avoid creating a gap
 371                                 # in the cat_id sequence. The row may or may not be "affected".
 372                                 $dbw->update(
 373                                         'category',
 374                                         [
 375                                                 'cat_pages' => $result->pages,
 376                                                 'cat_subcats' => $result->subcats,
 377                                                 'cat_files' => $result->files
 378                                         ],
 379                                         [ 'cat_title' => $this->mName ],
 380                                         __METHOD__
 381                                 );
 382                         } else {
 383                                 # The category is empty and has no description page, delete it
 384                                 $dbw->delete(
 385                                         'category',
 386                                         [ 'cat_title' => $this->mName ],
 387                                         __METHOD__
 388                                 );
 389                                 $this->mID = false;
 390                         }
 391                 } elseif ( $shouldExist ) {
 392                         # The category row doesn't exist but should, so create it. Use
 393                         # upsert in case of races.
 394                         $dbw->upsert(
 395                                 'category',
 396                                 [
 397                                         'cat_title' => $this->mName,
 398                                         'cat_pages' => $result->pages,
 399                                         'cat_subcats' => $result->subcats,
 400                                         'cat_files' => $result->files
 401                                 ],
 402                                 [ 'cat_title' ],
 403                                 [
 404                                         'cat_pages' => $result->pages,
 405                                         'cat_subcats' => $result->subcats,
 406                                         'cat_files' => $result->files
 407                                 ],
 408                                 __METHOD__
 409                         );
 410                         // @todo: Should we update $this->mID here? Or not since Category
 411                         // objects tend to be short lived enough to not matter?
 412                 }
 413
 414                 $dbw->endAtomic( __METHOD__ );
 415
 416                 # Now we should update our local counts.
 417                 $this->mPages = $result->pages;
 418                 $this->mSubcats = $result->subcats;
 419                 $this->mFiles = $result->files;
 420
 421                 return true;
 422         }
 423
 424         /**
 425          * Call refreshCounts() if there are no entries in the categorylinks table
 426          * or if the category table has a row that states that there are no entries
 427          *
 428          * Due to lock errors or other failures, the precomputed counts can get out of sync,
 429          * making it hard to know when to delete the category row without checking the
 430          * categorylinks table.
 431          *
 432          * @return bool Whether links were refreshed
 433          * @since 1.32
 434          */
 435         public function refreshCountsIfEmpty() {
 436                 return $this->refreshCountsIfSmall( 0 );
 437         }
 438
 439         /**
 440          * Call refreshCounts() if there are few entries in the categorylinks table
 441          *
 442          * Due to lock errors or other failures, the precomputed counts can get out of sync,
 443          * making it hard to know when to delete the category row without checking the
 444          * categorylinks table.
 445          *
 446          * This method will do a non-locking select first to reduce contention.
 447          *
 448          * @param int $maxSize Only refresh if there are this or less many backlinks
 449          * @return bool Whether links were refreshed
 450          * @since 1.34
 451          */
 452         public function refreshCountsIfSmall( $maxSize = self::ROW_COUNT_SMALL ) {
 453                 $dbw = wfGetDB( DB_MASTER );
 454                 $dbw->startAtomic( __METHOD__ );
 455
 456                 $typeOccurances = $dbw->selectFieldValues(
 457                         'categorylinks',
 458                         'cl_type',
 459                         [ 'cl_to' => $this->getName() ],
 460                         __METHOD__,
 461                         [ 'LIMIT' => $maxSize + 1 ]
 462                 );
 463
 464                 if ( !$typeOccurances ) {
 465                         $doRefresh = true; // delete any category table entry
 466                 } elseif ( count( $typeOccurances ) <= $maxSize ) {
 467                         $countByType = array_count_values( $typeOccurances );
 468                         $doRefresh = !$dbw->selectField(
 469                                 'category',
 470                                 '1',
 471                                 [
 472                                         'cat_title' => $this->getName(),
 473                                         'cat_pages' => $countByType['page'] ?? 0,
 474                                         'cat_subcats' => $countByType['subcat'] ?? 0,
 475                                         'cat_files' => $countByType['file'] ?? 0
 476                                 ],
 477                                 __METHOD__
 478                         );
 479                 } else {
 480                         $doRefresh = false; // category is too big
 481                 }
 482
 483                 $dbw->endAtomic( __METHOD__ );
 484
 485                 if ( $doRefresh ) {
 486                         $this->refreshCounts(); // update the row
 487
 488                         return true;
 489                 }
 490
 491                 return false;
 492         }
 493 }