includes/Collation.php

   1 <?php
   2 /**
   3  * Database row sorting.
   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 abstract class Collation {
  24         static $instance;
  25
  26         /**
  27          * @return Collation
  28          */
  29         static function singleton() {
  30                 if ( !self::$instance ) {
  31                         global $wgCategoryCollation;
  32                         self::$instance = self::factory( $wgCategoryCollation );
  33                 }
  34                 return self::$instance;
  35         }
  36
  37         /**
  38          * @throws MWException
  39          * @param $collationName string
  40          * @return Collation
  41          */
  42         static function factory( $collationName ) {
  43                 switch( $collationName ) {
  44                         case 'uppercase':
  45                                 return new UppercaseCollation;
  46                         case 'identity':
  47                                 return new IdentityCollation;
  48                         case 'uca-default':
  49                                 return new IcuCollation( 'root' );
  50                         default:
  51                                 # Provide a mechanism for extensions to hook in.
  52
  53                                 $collationObject = null;
  54                                 wfRunHooks( 'Collation::factory', array( $collationName, &$collationObject ) );
  55
  56                                 if ( $collationObject instanceof Collation ) {
  57                                         return $collationObject;
  58                                 }
  59
  60                                 // If all else fails...
  61                                 throw new MWException( __METHOD__.": unknown collation type \"$collationName\"" );
  62                 }
  63         }
  64
  65         /**
  66          * Given a string, convert it to a (hopefully short) key that can be used
  67          * for efficient sorting.  A binary sort according to the sortkeys
  68          * corresponds to a logical sort of the corresponding strings.  Current
  69          * code expects that a line feed character should sort before all others, but
  70          * has no other particular expectations (and that one can be changed if
  71          * necessary).
  72          *
  73          * @param string $string UTF-8 string
  74          * @return string Binary sortkey
  75          */
  76         abstract function getSortKey( $string );
  77
  78         /**
  79          * Given a string, return the logical "first letter" to be used for
  80          * grouping on category pages and so on.  This has to be coordinated
  81          * carefully with convertToSortkey(), or else the sorted list might jump
  82          * back and forth between the same "initial letters" or other pathological
  83          * behavior.  For instance, if you just return the first character, but "a"
  84          * sorts the same as "A" based on getSortKey(), then you might get a
  85          * list like
  86          *
  87          * == A ==
  88          * * [[Aardvark]]
  89          *
  90          * == a ==
  91          * * [[antelope]]
  92          *
  93          * == A ==
  94          * * [[Ape]]
  95          *
  96          * etc., assuming for the sake of argument that $wgCapitalLinks is false.
  97          *
  98          * @param string $string UTF-8 string
  99          * @return string UTF-8 string corresponding to the first letter of input
 100          */
 101         abstract function getFirstLetter( $string );
 102 }
 103
 104 class UppercaseCollation extends Collation {
 105         var $lang;
 106         function __construct() {
 107                 // Get a language object so that we can use the generic UTF-8 uppercase
 108                 // function there
 109                 $this->lang = Language::factory( 'en' );
 110         }
 111
 112         function getSortKey( $string ) {
 113                 return $this->lang->uc( $string );
 114         }
 115
 116         function getFirstLetter( $string ) {
 117                 if ( $string[0] == "\0" ) {
 118                         $string = substr( $string, 1 );
 119                 }
 120                 return $this->lang->ucfirst( $this->lang->firstChar( $string ) );
 121         }
 122 }
 123
 124 /**
 125  * Collation class that's essentially a no-op.
 126  *
 127  * Does sorting based on binary value of the string.
 128  * Like how things were pre 1.17.
 129  */
 130 class IdentityCollation extends Collation {
 131
 132         function getSortKey( $string ) {
 133                 return $string;
 134         }
 135
 136         function getFirstLetter( $string ) {
 137                 global $wgContLang;
 138                 // Copied from UppercaseCollation.
 139                 // I'm kind of unclear on when this could happen...
 140                 if ( $string[0] == "\0" ) {
 141                         $string = substr( $string, 1 );
 142                 }
 143                 return $wgContLang->firstChar( $string );
 144         }
 145 }
 146
 147
 148 class IcuCollation extends Collation {
 149         var $primaryCollator, $mainCollator, $locale;
 150         var $firstLetterData;
 151
 152         /**
 153          * Unified CJK blocks.
 154          *
 155          * The same definition of a CJK block must be used for both Collation and
 156          * generateCollationData.php. These blocks are omitted from the first
 157          * letter data, as an optimisation measure and because the default UCA table
 158          * is pretty useless for sorting Chinese text anyway. Japanese and Korean
 159          * blocks are not included here, because they are smaller and more useful.
 160          */
 161         static $cjkBlocks = array(
 162                 array( 0x2E80, 0x2EFF ), // CJK Radicals Supplement
 163                 array( 0x2F00, 0x2FDF ), // Kangxi Radicals
 164                 array( 0x2FF0, 0x2FFF ), // Ideographic Description Characters
 165                 array( 0x3000, 0x303F ), // CJK Symbols and Punctuation
 166                 array( 0x31C0, 0x31EF ), // CJK Strokes
 167                 array( 0x3200, 0x32FF ), // Enclosed CJK Letters and Months
 168                 array( 0x3300, 0x33FF ), // CJK Compatibility
 169                 array( 0x3400, 0x4DBF ), // CJK Unified Ideographs Extension A
 170                 array( 0x4E00, 0x9FFF ), // CJK Unified Ideographs
 171                 array( 0xF900, 0xFAFF ), // CJK Compatibility Ideographs
 172                 array( 0xFE30, 0xFE4F ), // CJK Compatibility Forms
 173                 array( 0x20000, 0x2A6DF ), // CJK Unified Ideographs Extension B
 174                 array( 0x2A700, 0x2B73F ), // CJK Unified Ideographs Extension C
 175                 array( 0x2B740, 0x2B81F ), // CJK Unified Ideographs Extension D
 176                 array( 0x2F800, 0x2FA1F ), // CJK Compatibility Ideographs Supplement
 177         );
 178
 179         const RECORD_LENGTH = 14;
 180
 181         function __construct( $locale ) {
 182                 if ( !extension_loaded( 'intl' ) ) {
 183                         throw new MWException( 'An ICU collation was requested, ' .
 184                                 'but the intl extension is not available.' );
 185                 }
 186                 $this->locale = $locale;
 187                 $this->mainCollator = Collator::create( $locale );
 188                 if ( !$this->mainCollator ) {
 189                         throw new MWException( "Invalid ICU locale specified for collation: $locale" );
 190                 }
 191
 192                 $this->primaryCollator = Collator::create( $locale );
 193                 $this->primaryCollator->setStrength( Collator::PRIMARY );
 194         }
 195
 196         function getSortKey( $string ) {
 197                 // intl extension produces non null-terminated
 198                 // strings. Appending '' fixes it so that it doesn't generate
 199                 // a warning on each access in debug php.
 200                 wfSuppressWarnings();
 201                 $key = $this->mainCollator->getSortKey( $string ) . '';
 202                 wfRestoreWarnings();
 203                 return $key;
 204         }
 205
 206         function getPrimarySortKey( $string ) {
 207                 wfSuppressWarnings();
 208                 $key = $this->primaryCollator->getSortKey( $string ) . '';
 209                 wfRestoreWarnings();
 210                 return $key;
 211         }
 212
 213         function getFirstLetter( $string ) {
 214                 $string = strval( $string );
 215                 if ( $string === '' ) {
 216                         return '';
 217                 }
 218
 219                 // Check for CJK
 220                 $firstChar = mb_substr( $string, 0, 1, 'UTF-8' );
 221                 if ( ord( $firstChar ) > 0x7f
 222                         && self::isCjk( utf8ToCodepoint( $firstChar ) ) )
 223                 {
 224                         return $firstChar;
 225                 }
 226
 227                 $sortKey = $this->getPrimarySortKey( $string );
 228
 229                 // Do a binary search to find the correct letter to sort under
 230                 $min = $this->findLowerBound(
 231                         array( $this, 'getSortKeyByLetterIndex' ),
 232                         $this->getFirstLetterCount(),
 233                         'strcmp',
 234                         $sortKey );
 235
 236                 if ( $min === false ) {
 237                         // Before the first letter
 238                         return '';
 239                 }
 240                 return $this->getLetterByIndex( $min );
 241         }
 242
 243         function getFirstLetterData() {
 244                 if ( $this->firstLetterData !== null ) {
 245                         return $this->firstLetterData;
 246                 }
 247
 248                 $cache = wfGetCache( CACHE_ANYTHING );
 249                 $cacheKey = wfMemcKey( 'first-letters', $this->locale );
 250                 $cacheEntry = $cache->get( $cacheKey );
 251
 252                 if ( $cacheEntry ) {
 253                         $this->firstLetterData = $cacheEntry;
 254                         return $this->firstLetterData;
 255                 }
 256
 257                 // Generate data from serialized data file
 258
 259                 $letters = wfGetPrecompiledData( "first-letters-{$this->locale}.ser" );
 260                 if ( $letters === false ) {
 261                         throw new MWException( "MediaWiki does not support ICU locale " .
 262                                 "\"{$this->locale}\"" );
 263                 }
 264
 265                 // Sort the letters.
 266                 //
 267                 // It's impossible to have the precompiled data file properly sorted,
 268                 // because the sort order changes depending on ICU version. If the
 269                 // array is not properly sorted, the binary search will return random
 270                 // results.
 271                 //
 272                 // We also take this opportunity to remove primary collisions.
 273                 $letterMap = array();
 274                 foreach ( $letters as $letter ) {
 275                         $key = $this->getPrimarySortKey( $letter );
 276                         if ( isset( $letterMap[$key] ) ) {
 277                                 // Primary collision
 278                                 // Keep whichever one sorts first in the main collator
 279                                 if ( $this->mainCollator->compare( $letter, $letterMap[$key] ) < 0 ) {
 280                                         $letterMap[$key] = $letter;
 281                                 }
 282                         } else {
 283                                 $letterMap[$key] = $letter;
 284                         }
 285                 }
 286                 ksort( $letterMap, SORT_STRING );
 287                 $data = array(
 288                         'chars' => array_values( $letterMap ),
 289                         'keys' => array_keys( $letterMap )
 290                 );
 291
 292                 // Reduce memory usage before caching
 293                 unset( $letterMap );
 294
 295                 // Save to cache
 296                 $this->firstLetterData = $data;
 297                 $cache->set( $cacheKey, $data, 86400 * 7 /* 1 week */ );
 298                 return $data;
 299         }
 300
 301         function getLetterByIndex( $index ) {
 302                 if ( $this->firstLetterData === null ) {
 303                         $this->getFirstLetterData();
 304                 }
 305                 return $this->firstLetterData['chars'][$index];
 306         }
 307
 308         function getSortKeyByLetterIndex( $index ) {
 309                 if ( $this->firstLetterData === null ) {
 310                         $this->getFirstLetterData();
 311                 }
 312                 return $this->firstLetterData['keys'][$index];
 313         }
 314
 315         function getFirstLetterCount() {
 316                 if ( $this->firstLetterData === null ) {
 317                         $this->getFirstLetterData();
 318                 }
 319                 return count( $this->firstLetterData['chars'] );
 320         }
 321
 322         /**
 323          * Do a binary search, and return the index of the largest item that sorts
 324          * less than or equal to the target value.
 325          *
 326          * @param $valueCallback array A function to call to get the value with
 327          *     a given array index.
 328          * @param $valueCount int The number of items accessible via $valueCallback,
 329          *     indexed from 0 to $valueCount - 1
 330          * @param $comparisonCallback array A callback to compare two values, returning
 331          *     -1, 0 or 1 in the style of strcmp().
 332          * @param $target string The target value to find.
 333          *
 334          * @return int|bool The item index of the lower bound, or false if the target value
 335          *     sorts before all items.
 336          */
 337         function findLowerBound( $valueCallback, $valueCount, $comparisonCallback, $target ) {
 338                 $min = 0;
 339                 $max = $valueCount - 1;
 340                 do {
 341                         $mid = $min + ( ( $max - $min ) >> 1 );
 342                         $item = call_user_func( $valueCallback, $mid );
 343                         $comparison = call_user_func( $comparisonCallback, $target, $item );
 344                         if ( $comparison > 0 ) {
 345                                 $min = $mid;
 346                         } elseif ( $comparison == 0 ) {
 347                                 $min = $mid;
 348                                 break;
 349                         } else {
 350                                 $max = $mid;
 351                         }
 352                 } while ( $min < $max - 1 );
 353
 354                 if ( $min == 0 && $max == 0 && $comparison > 0 ) {
 355                         // Before the first item
 356                         return false;
 357                 } else {
 358                         return $min;
 359                 }
 360         }
 361
 362         static function isCjk( $codepoint ) {
 363                 foreach ( self::$cjkBlocks as $block ) {
 364                         if ( $codepoint >= $block[0] && $codepoint <= $block[1] ) {
 365                                 return true;
 366                         }
 367                 }
 368                 return false;
 369         }
 370 }
 371