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                                 $match = array();
  52                                 if ( preg_match( '/^uca-([a-z-]+)$/', $collationName, $match ) ) {
  53                                         return new IcuCollation( $match[1] );
  54                                 }
  55
  56                                 # Provide a mechanism for extensions to hook in.
  57                                 $collationObject = null;
  58                                 wfRunHooks( 'Collation::factory', array( $collationName, &$collationObject ) );
  59
  60                                 if ( $collationObject instanceof Collation ) {
  61                                         return $collationObject;
  62                                 }
  63
  64                                 // If all else fails...
  65                                 throw new MWException( __METHOD__.": unknown collation type \"$collationName\"" );
  66                 }
  67         }
  68
  69         /**
  70          * Given a string, convert it to a (hopefully short) key that can be used
  71          * for efficient sorting.  A binary sort according to the sortkeys
  72          * corresponds to a logical sort of the corresponding strings.  Current
  73          * code expects that a line feed character should sort before all others, but
  74          * has no other particular expectations (and that one can be changed if
  75          * necessary).
  76          *
  77          * @param string $string UTF-8 string
  78          * @return string Binary sortkey
  79          */
  80         abstract function getSortKey( $string );
  81
  82         /**
  83          * Given a string, return the logical "first letter" to be used for
  84          * grouping on category pages and so on.  This has to be coordinated
  85          * carefully with convertToSortkey(), or else the sorted list might jump
  86          * back and forth between the same "initial letters" or other pathological
  87          * behavior.  For instance, if you just return the first character, but "a"
  88          * sorts the same as "A" based on getSortKey(), then you might get a
  89          * list like
  90          *
  91          * == A ==
  92          * * [[Aardvark]]
  93          *
  94          * == a ==
  95          * * [[antelope]]
  96          *
  97          * == A ==
  98          * * [[Ape]]
  99          *
 100          * etc., assuming for the sake of argument that $wgCapitalLinks is false.
 101          *
 102          * @param string $string UTF-8 string
 103          * @return string UTF-8 string corresponding to the first letter of input
 104          */
 105         abstract function getFirstLetter( $string );
 106 }
 107
 108 class UppercaseCollation extends Collation {
 109         var $lang;
 110         function __construct() {
 111                 // Get a language object so that we can use the generic UTF-8 uppercase
 112                 // function there
 113                 $this->lang = Language::factory( 'en' );
 114         }
 115
 116         function getSortKey( $string ) {
 117                 return $this->lang->uc( $string );
 118         }
 119
 120         function getFirstLetter( $string ) {
 121                 if ( $string[0] == "\0" ) {
 122                         $string = substr( $string, 1 );
 123                 }
 124                 return $this->lang->ucfirst( $this->lang->firstChar( $string ) );
 125         }
 126 }
 127
 128 /**
 129  * Collation class that's essentially a no-op.
 130  *
 131  * Does sorting based on binary value of the string.
 132  * Like how things were pre 1.17.
 133  */
 134 class IdentityCollation extends Collation {
 135
 136         function getSortKey( $string ) {
 137                 return $string;
 138         }
 139
 140         function getFirstLetter( $string ) {
 141                 global $wgContLang;
 142                 // Copied from UppercaseCollation.
 143                 // I'm kind of unclear on when this could happen...
 144                 if ( $string[0] == "\0" ) {
 145                         $string = substr( $string, 1 );
 146                 }
 147                 return $wgContLang->firstChar( $string );
 148         }
 149 }
 150
 151
 152 class IcuCollation extends Collation {
 153         var $primaryCollator, $mainCollator, $locale;
 154         var $firstLetterData;
 155
 156         /**
 157          * Unified CJK blocks.
 158          *
 159          * The same definition of a CJK block must be used for both Collation and
 160          * generateCollationData.php. These blocks are omitted from the first
 161          * letter data, as an optimisation measure and because the default UCA table
 162          * is pretty useless for sorting Chinese text anyway. Japanese and Korean
 163          * blocks are not included here, because they are smaller and more useful.
 164          */
 165         static $cjkBlocks = array(
 166                 array( 0x2E80, 0x2EFF ), // CJK Radicals Supplement
 167                 array( 0x2F00, 0x2FDF ), // Kangxi Radicals
 168                 array( 0x2FF0, 0x2FFF ), // Ideographic Description Characters
 169                 array( 0x3000, 0x303F ), // CJK Symbols and Punctuation
 170                 array( 0x31C0, 0x31EF ), // CJK Strokes
 171                 array( 0x3200, 0x32FF ), // Enclosed CJK Letters and Months
 172                 array( 0x3300, 0x33FF ), // CJK Compatibility
 173                 array( 0x3400, 0x4DBF ), // CJK Unified Ideographs Extension A
 174                 array( 0x4E00, 0x9FFF ), // CJK Unified Ideographs
 175                 array( 0xF900, 0xFAFF ), // CJK Compatibility Ideographs
 176                 array( 0xFE30, 0xFE4F ), // CJK Compatibility Forms
 177                 array( 0x20000, 0x2A6DF ), // CJK Unified Ideographs Extension B
 178                 array( 0x2A700, 0x2B73F ), // CJK Unified Ideographs Extension C
 179                 array( 0x2B740, 0x2B81F ), // CJK Unified Ideographs Extension D
 180                 array( 0x2F800, 0x2FA1F ), // CJK Compatibility Ideographs Supplement
 181         );
 182
 183         /**
 184          * Additional characters (or character groups) to be considered first-letters
 185          *
 186          * Generated based on the primary level of Unicode collation tailorings
 187          * available at http://developer.mimer.com/charts/tailorings.htm .
 188          *
 189          * Empty arrays are intended; this signifies that the data for the language is
 190          * available and that there are, in fact, no additional letters to consider.
 191          */
 192         static $tailoringFirstLetters = array(
 193                 // Verified by native speakers
 194                 'pl' => array( "Ą", "Ć", "Ę", "Ł", "Ń", "Ó", "Ś", "Ź", "Ż" ),
 195                 'fi' => array( "Å", "Ä", "Ö" ),
 196                 // Not verified, but likely correct
 197                 'af' => array(),
 198                 'ast' => array( "CH", "LL", "Ñ" ),
 199                 'az' => array( "Ç", "Ə", "Ğ", "İ", "Ö", "Ş", "Ü" ),
 200                 'be' => array( "Ё" ),
 201                 'bg' => array(),
 202                 'br' => array( "CH", "C'H" ),
 203                 'bs' => array( "Č", "Ć", "DŽ", "Đ", "LJ", "NJ", "Š", "Ž" ),
 204                 'ca' => array(),
 205                 'co' => array(),
 206                 'cs' => array( "Č", "CH", "Ř", "Š", "Ž" ),
 207                 'cy' => array( "CH", "DD", "FF", "NG", "LL", "PH", "RH", "TH" ),
 208                 'da' => array( "Æ", "Ø", "Å" ),
 209                 'de' => array(),
 210                 'dsb' => array( "Č", "Ć", "DŹ", "Ě", "CH", "Ł", "Ń", "Ŕ", "Š", "Ś", "Ž", "Ź" ),
 211                 'el' => array(),
 212                 'en' => array(),
 213                 'eo' => array( "Ĉ", "Ĝ", "Ĥ", "Ĵ", "Ŝ", "Ŭ" ),
 214                 'es' => array( "Ñ" ),
 215                 'et' => array( "Š", "Ž", "Õ", "Ä", "Ö", "Ü" ),
 216                 'eu' => array( "Ñ" ),
 217                 'fo' => array( "Á", "Ð", "Í", "Ó", "Ú", "Ý", "Æ", "Ø", "Å" ),
 218                 'fr' => array(),
 219                 'fur' => array( "À", "Á", "Â", "È", "Ì", "Ò", "Ù" ),
 220                 'fy' => array(),
 221                 'ga' => array(),
 222                 'gd' => array(),
 223                 'gl' => array( "CH", "LL", "Ñ" ),
 224                 'hr' => array( "Č", "Ć", "DŽ", "Đ", "LJ", "NJ", "Š", "Ž" ),
 225                 'hsb' => array( "Č", "DŹ", "Ě", "CH", "Ł", "Ń", "Ř", "Š", "Ć", "Ž" ),
 226                 'hu' => array( "CS", "DZ", "DZS", "GY", "LY", "NY", "Ö", "SZ", "TY", "Ü", "ZS" ),
 227                 'is' => array( "Á", "Ð", "É", "Í", "Ó", "Ú", "Ý", "Þ", "Æ", "Ö", "Å" ),
 228                 'it' => array(),
 229                 'kk' => array( "Ү", "І" ),
 230                 'kl' => array( "Æ", "Ø", "Å" ),
 231                 'ku' => array( "Ç", "Ê", "Î", "Ş", "Û" ),
 232                 'ky' => array( "Ё" ),
 233                 'la' => array(),
 234                 'lb' => array(),
 235                 'lt' => array( "Č", "Š", "Ž" ),
 236                 'lv' => array( "Č", "Ģ", "Ķ", "Ļ", "Ņ", "Š", "Ž" ),
 237                 'mk' => array(),
 238                 'mo' => array( "Ă", "Â", "Î", "Ş", "Ţ" ),
 239                 'mt' => array( "Ċ", "Ġ", "GĦ", "Ħ", "Ż" ),
 240                 'nl' => array(),
 241                 'no' => array( "Æ", "Ø", "Å" ),
 242                 'oc' => array(),
 243                 'pt' => array(),
 244                 'rm' => array(),
 245                 'ro' => array( "Ă", "Â", "Î", "Ş", "Ţ" ),
 246                 'ru' => array(),
 247                 'rup' => array( "Ă", "Â", "Î", "Ľ", "Ń", "Ş", "Ţ" ),
 248                 'sco' => array(),
 249                 'sk' => array( "Ä", "Č", "CH", "Ô", "Š", "Ž" ),
 250                 'sl' => array( "Č", "Š", "Ž" ),
 251                 'smn' => array( "Á", "Č", "Đ", "Ŋ", "Š", "Ŧ", "Ž", "Æ", "Ø", "Å", "Ä", "Ö" ),
 252                 'sq' => array( "Ç", "DH", "Ë", "GJ", "LL", "NJ", "RR", "SH", "TH", "XH", "ZH" ),
 253                 'sr' => array(),
 254                 'sv' => array( "Å", "Ä", "Ö" ),
 255                 'tk' => array( "Ç", "Ä", "Ž", "Ň", "Ö", "Ş", "Ü", "Ý" ),
 256                 'tl' => array( "Ñ", "NG" ), /* 'fil' in the data source */
 257                 'tr' => array( "Ç", "Ğ", "İ", "Ö", "Ş", "Ü" ),
 258                 'tt' => array( "Ә", "Ө", "Ү", "Җ", "Ң", "Һ" ),
 259                 'uk' => array( "Ґ", "Ь" ),
 260                 'uz' => array( "CH", "G'", "NG", "O'", "SH" ),
 261                 'vi' => array( "Ă", "Â", "Đ", "Ê", "Ô", "Ơ", "Ư" ),
 262         );
 263
 264         const RECORD_LENGTH = 14;
 265
 266         function __construct( $locale ) {
 267                 if ( !extension_loaded( 'intl' ) ) {
 268                         throw new MWException( 'An ICU collation was requested, ' .
 269                                 'but the intl extension is not available.' );
 270                 }
 271                 $this->locale = $locale;
 272                 $this->mainCollator = Collator::create( $locale );
 273                 if ( !$this->mainCollator ) {
 274                         throw new MWException( "Invalid ICU locale specified for collation: $locale" );
 275                 }
 276
 277                 $this->primaryCollator = Collator::create( $locale );
 278                 $this->primaryCollator->setStrength( Collator::PRIMARY );
 279         }
 280
 281         function getSortKey( $string ) {
 282                 // intl extension produces non null-terminated
 283                 // strings. Appending '' fixes it so that it doesn't generate
 284                 // a warning on each access in debug php.
 285                 wfSuppressWarnings();
 286                 $key = $this->mainCollator->getSortKey( $string ) . '';
 287                 wfRestoreWarnings();
 288                 return $key;
 289         }
 290
 291         function getPrimarySortKey( $string ) {
 292                 wfSuppressWarnings();
 293                 $key = $this->primaryCollator->getSortKey( $string ) . '';
 294                 wfRestoreWarnings();
 295                 return $key;
 296         }
 297
 298         function getFirstLetter( $string ) {
 299                 $string = strval( $string );
 300                 if ( $string === '' ) {
 301                         return '';
 302                 }
 303
 304                 // Check for CJK
 305                 $firstChar = mb_substr( $string, 0, 1, 'UTF-8' );
 306                 if ( ord( $firstChar ) > 0x7f
 307                         && self::isCjk( utf8ToCodepoint( $firstChar ) ) )
 308                 {
 309                         return $firstChar;
 310                 }
 311
 312                 $sortKey = $this->getPrimarySortKey( $string );
 313
 314                 // Do a binary search to find the correct letter to sort under
 315                 $min = $this->findLowerBound(
 316                         array( $this, 'getSortKeyByLetterIndex' ),
 317                         $this->getFirstLetterCount(),
 318                         'strcmp',
 319                         $sortKey );
 320
 321                 if ( $min === false ) {
 322                         // Before the first letter
 323                         return '';
 324                 }
 325                 return $this->getLetterByIndex( $min );
 326         }
 327
 328         function getFirstLetterData() {
 329                 if ( $this->firstLetterData !== null ) {
 330                         return $this->firstLetterData;
 331                 }
 332
 333                 $cache = wfGetCache( CACHE_ANYTHING );
 334                 $cacheKey = wfMemcKey( 'first-letters', $this->locale );
 335                 $cacheEntry = $cache->get( $cacheKey );
 336
 337                 if ( $cacheEntry ) {
 338                         $this->firstLetterData = $cacheEntry;
 339                         return $this->firstLetterData;
 340                 }
 341
 342                 // Generate data from serialized data file
 343
 344                 if ( isset ( self::$tailoringFirstLetters[$this->locale] ) ) {
 345                         $letters = wfGetPrecompiledData( "first-letters-root.ser" );
 346                         $letters = array_merge( $letters, self::$tailoringFirstLetters[$this->locale] );
 347                 } else {
 348                         $letters = wfGetPrecompiledData( "first-letters-{$this->locale}.ser" );
 349                         if ( $letters === false ) {
 350                                 throw new MWException( "MediaWiki does not support ICU locale " .
 351                                         "\"{$this->locale}\"" );
 352                         }
 353                 }
 354
 355                 // Sort the letters.
 356                 //
 357                 // It's impossible to have the precompiled data file properly sorted,
 358                 // because the sort order changes depending on ICU version. If the
 359                 // array is not properly sorted, the binary search will return random
 360                 // results.
 361                 //
 362                 // We also take this opportunity to remove primary collisions.
 363                 $letterMap = array();
 364                 foreach ( $letters as $letter ) {
 365                         $key = $this->getPrimarySortKey( $letter );
 366                         if ( isset( $letterMap[$key] ) ) {
 367                                 // Primary collision
 368                                 // Keep whichever one sorts first in the main collator
 369                                 if ( $this->mainCollator->compare( $letter, $letterMap[$key] ) < 0 ) {
 370                                         $letterMap[$key] = $letter;
 371                                 }
 372                         } else {
 373                                 $letterMap[$key] = $letter;
 374                         }
 375                 }
 376                 ksort( $letterMap, SORT_STRING );
 377                 $data = array(
 378                         'chars' => array_values( $letterMap ),
 379                         'keys' => array_keys( $letterMap )
 380                 );
 381
 382                 // Reduce memory usage before caching
 383                 unset( $letterMap );
 384
 385                 // Save to cache
 386                 $this->firstLetterData = $data;
 387                 $cache->set( $cacheKey, $data, 86400 * 7 /* 1 week */ );
 388                 return $data;
 389         }
 390
 391         function getLetterByIndex( $index ) {
 392                 if ( $this->firstLetterData === null ) {
 393                         $this->getFirstLetterData();
 394                 }
 395                 return $this->firstLetterData['chars'][$index];
 396         }
 397
 398         function getSortKeyByLetterIndex( $index ) {
 399                 if ( $this->firstLetterData === null ) {
 400                         $this->getFirstLetterData();
 401                 }
 402                 return $this->firstLetterData['keys'][$index];
 403         }
 404
 405         function getFirstLetterCount() {
 406                 if ( $this->firstLetterData === null ) {
 407                         $this->getFirstLetterData();
 408                 }
 409                 return count( $this->firstLetterData['chars'] );
 410         }
 411
 412         /**
 413          * Do a binary search, and return the index of the largest item that sorts
 414          * less than or equal to the target value.
 415          *
 416          * @param $valueCallback array A function to call to get the value with
 417          *     a given array index.
 418          * @param $valueCount int The number of items accessible via $valueCallback,
 419          *     indexed from 0 to $valueCount - 1
 420          * @param $comparisonCallback array A callback to compare two values, returning
 421          *     -1, 0 or 1 in the style of strcmp().
 422          * @param $target string The target value to find.
 423          *
 424          * @return int|bool The item index of the lower bound, or false if the target value
 425          *     sorts before all items.
 426          */
 427         function findLowerBound( $valueCallback, $valueCount, $comparisonCallback, $target ) {
 428                 if ( $valueCount === 0 ) {
 429                         return false;
 430                 }
 431
 432                 $min = 0;
 433                 $max = $valueCount;
 434                 do {
 435                         $mid = $min + ( ( $max - $min ) >> 1 );
 436                         $item = call_user_func( $valueCallback, $mid );
 437                         $comparison = call_user_func( $comparisonCallback, $target, $item );
 438                         if ( $comparison > 0 ) {
 439                                 $min = $mid;
 440                         } elseif ( $comparison == 0 ) {
 441                                 $min = $mid;
 442                                 break;
 443                         } else {
 444                                 $max = $mid;
 445                         }
 446                 } while ( $min < $max - 1 );
 447
 448                 if ( $min == 0 ) {
 449                         $item = call_user_func( $valueCallback, $min );
 450                         $comparison = call_user_func( $comparisonCallback, $target, $item );
 451                         if ( $comparison < 0 ) {
 452                                 // Before the first item
 453                                 return false;
 454                         }
 455                 }
 456                 return $min;
 457         }
 458
 459         static function isCjk( $codepoint ) {
 460                 foreach ( self::$cjkBlocks as $block ) {
 461                         if ( $codepoint >= $block[0] && $codepoint <= $block[1] ) {
 462                                 return true;
 463                         }
 464                 }
 465                 return false;
 466         }
 467
 468         /**
 469          * Return the version of ICU library used by PHP's intl extension,
 470          * or false when the extension is not installed of the version
 471          * can't be determined.
 472          *
 473          * The constant INTL_ICU_VERSION this function refers to isn't really
 474          * documented. It is available since PHP 5.3.7 (see PHP bug 54561).
 475          * This function will return false on older PHPs.
 476          *
 477          * @since 1.21
 478          * @return string|false
 479          */
 480         static function getICUVersion() {
 481                 return defined( 'INTL_ICU_VERSION' ) ? INTL_ICU_VERSION : false;
 482         }
 483
 484         /**
 485          * Return the version of Unicode appropriate for the version of ICU library
 486          * currently in use, or false when it can't be determined.
 487          *
 488          * @since 1.21
 489          * @return string|false
 490          */
 491         static function getUnicodeVersionForICU() {
 492                 $icuVersion = IcuCollation::getICUVersion();
 493                 if ( !$icuVersion ) {
 494                         return false;
 495                 }
 496
 497                 $versionPrefix = substr( $icuVersion, 0, 3 );
 498                 // Source: http://site.icu-project.org/download
 499                 $map = array(
 500                         '50.' => '6.2',
 501                         '49.' => '6.1',
 502                         '4.8' => '6.0',
 503                         '4.6' => '6.0',
 504                         '4.4' => '5.2',
 505                         '4.2' => '5.1',
 506                         '4.0' => '5.1',
 507                         '3.8' => '5.0',
 508                         '3.6' => '5.0',
 509                         '3.4' => '4.1',
 510                 );
 511
 512                 if ( isset( $map[$versionPrefix] ) ) {
 513                         return $map[$versionPrefix];
 514                 } else {
 515                         return false;
 516                 }
 517         }
 518 }