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