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 'uppercase-sv':
  47                                 return new UppercaseSvCollation;
  48                         case 'identity':
  49                                 return new IdentityCollation;
  50                         case 'uca-default':
  51                                 return new IcuCollation( 'root' );
  52                         default:
  53                                 $match = array();
  54                                 if ( preg_match( '/^uca-([a-z-]+)$/', $collationName, $match ) ) {
  55                                         return new IcuCollation( $match[1] );
  56                                 }
  57
  58                                 # Provide a mechanism for extensions to hook in.
  59                                 $collationObject = null;
  60                                 wfRunHooks( 'Collation::factory', array( $collationName, &$collationObject ) );
  61
  62                                 if ( $collationObject instanceof Collation ) {
  63                                         return $collationObject;
  64                                 }
  65
  66                                 // If all else fails...
  67                                 throw new MWException( __METHOD__.": unknown collation type \"$collationName\"" );
  68                 }
  69         }
  70
  71         /**
  72          * Given a string, convert it to a (hopefully short) key that can be used
  73          * for efficient sorting.  A binary sort according to the sortkeys
  74          * corresponds to a logical sort of the corresponding strings.  Current
  75          * code expects that a line feed character should sort before all others, but
  76          * has no other particular expectations (and that one can be changed if
  77          * necessary).
  78          *
  79          * @param string $string UTF-8 string
  80          * @return string Binary sortkey
  81          */
  82         abstract function getSortKey( $string );
  83
  84         /**
  85          * Given a string, return the logical "first letter" to be used for
  86          * grouping on category pages and so on.  This has to be coordinated
  87          * carefully with convertToSortkey(), or else the sorted list might jump
  88          * back and forth between the same "initial letters" or other pathological
  89          * behavior.  For instance, if you just return the first character, but "a"
  90          * sorts the same as "A" based on getSortKey(), then you might get a
  91          * list like
  92          *
  93          * == A ==
  94          * * [[Aardvark]]
  95          *
  96          * == a ==
  97          * * [[antelope]]
  98          *
  99          * == A ==
 100          * * [[Ape]]
 101          *
 102          * etc., assuming for the sake of argument that $wgCapitalLinks is false.
 103          *
 104          * @param string $string UTF-8 string
 105          * @return string UTF-8 string corresponding to the first letter of input
 106          */
 107         abstract function getFirstLetter( $string );
 108 }
 109
 110 class UppercaseCollation extends Collation {
 111         var $lang;
 112         function __construct() {
 113                 // Get a language object so that we can use the generic UTF-8 uppercase
 114                 // function there
 115                 $this->lang = Language::factory( 'en' );
 116         }
 117
 118         function getSortKey( $string ) {
 119                 return $this->lang->uc( $string );
 120         }
 121
 122         function getFirstLetter( $string ) {
 123                 if ( $string[0] == "\0" ) {
 124                         $string = substr( $string, 1 );
 125                 }
 126                 return $this->lang->ucfirst( $this->lang->firstChar( $string ) );
 127         }
 128 }
 129
 130 /**
 131  * Like UppercaseCollation but swaps Ä and Æ.
 132  *
 133  * This provides an ordering suitable for Swedish.
 134  * @author Lejonel
 135  */
 136 class UppercaseSvCollation extends UppercaseCollation {
 137
 138         /* Unicode code point order is ÄÅÆÖ, Swedish order is ÅÄÖ and Æ is often sorted as Ä.
 139          * Replacing Ä for Æ should give a better collation. */
 140         function getSortKey( $string ) {
 141                 $uppercase = $this->lang->uc( $string );
 142                 return strtr( $uppercase, array( 'Ä' => 'Æ', 'Æ' => 'Ä' ) );
 143         }
 144 }
 145
 146 /**
 147  * Collation class that's essentially a no-op.
 148  *
 149  * Does sorting based on binary value of the string.
 150  * Like how things were pre 1.17.
 151  */
 152 class IdentityCollation extends Collation {
 153
 154         function getSortKey( $string ) {
 155                 return $string;
 156         }
 157
 158         function getFirstLetter( $string ) {
 159                 global $wgContLang;
 160                 // Copied from UppercaseCollation.
 161                 // I'm kind of unclear on when this could happen...
 162                 if ( $string[0] == "\0" ) {
 163                         $string = substr( $string, 1 );
 164                 }
 165                 return $wgContLang->firstChar( $string );
 166         }
 167 }
 168
 169
 170 class IcuCollation extends Collation {
 171         var $primaryCollator, $mainCollator, $locale;
 172         var $firstLetterData;
 173
 174         /**
 175          * Unified CJK blocks.
 176          *
 177          * The same definition of a CJK block must be used for both Collation and
 178          * generateCollationData.php. These blocks are omitted from the first
 179          * letter data, as an optimisation measure and because the default UCA table
 180          * is pretty useless for sorting Chinese text anyway. Japanese and Korean
 181          * blocks are not included here, because they are smaller and more useful.
 182          */
 183         static $cjkBlocks = array(
 184                 array( 0x2E80, 0x2EFF ), // CJK Radicals Supplement
 185                 array( 0x2F00, 0x2FDF ), // Kangxi Radicals
 186                 array( 0x2FF0, 0x2FFF ), // Ideographic Description Characters
 187                 array( 0x3000, 0x303F ), // CJK Symbols and Punctuation
 188                 array( 0x31C0, 0x31EF ), // CJK Strokes
 189                 array( 0x3200, 0x32FF ), // Enclosed CJK Letters and Months
 190                 array( 0x3300, 0x33FF ), // CJK Compatibility
 191                 array( 0x3400, 0x4DBF ), // CJK Unified Ideographs Extension A
 192                 array( 0x4E00, 0x9FFF ), // CJK Unified Ideographs
 193                 array( 0xF900, 0xFAFF ), // CJK Compatibility Ideographs
 194                 array( 0xFE30, 0xFE4F ), // CJK Compatibility Forms
 195                 array( 0x20000, 0x2A6DF ), // CJK Unified Ideographs Extension B
 196                 array( 0x2A700, 0x2B73F ), // CJK Unified Ideographs Extension C
 197                 array( 0x2B740, 0x2B81F ), // CJK Unified Ideographs Extension D
 198                 array( 0x2F800, 0x2FA1F ), // CJK Compatibility Ideographs Supplement
 199         );
 200
 201         /**
 202          * Additional characters (or character groups) to be considered first-letters
 203          *
 204          * Generated based on the primary level of Unicode collation tailorings
 205          * available at http://developer.mimer.com/charts/tailorings.htm .
 206          *
 207          * Empty arrays are intended; this signifies that the data for the language is
 208          * available and that there are, in fact, no additional letters to consider.
 209          */
 210         static $tailoringFirstLetters = array(
 211                 // Verified by native speakers
 212                 'pl' => array( "Ą", "Ć", "Ę", "Ł", "Ń", "Ó", "Ś", "Ź", "Ż" ),
 213                 'fi' => array( "Å", "Ä", "Ö" ),
 214                 // Not verified, but likely correct
 215                 'af' => array(),
 216                 'ast' => array( "CH", "LL", "Ñ" ),
 217                 'az' => array( "Ç", "Ə", "Ğ", "İ", "Ö", "Ş", "Ü" ),
 218                 'be' => array( "Ё" ),
 219                 'bg' => array(),
 220                 'br' => array( "CH", "C'H" ),
 221                 'bs' => array( "Č", "Ć", "DŽ", "Đ", "LJ", "NJ", "Š", "Ž" ),
 222                 'ca' => array(),
 223                 'co' => array(),
 224                 'cs' => array( "Č", "CH", "Ř", "Š", "Ž" ),
 225                 'cy' => array( "CH", "DD", "FF", "NG", "LL", "PH", "RH", "TH" ),
 226                 'da' => array( "Æ", "Ø", "Å" ),
 227                 'de' => array(),
 228                 'dsb' => array( "Č", "Ć", "DŹ", "Ě", "CH", "Ł", "Ń", "Ŕ", "Š", "Ś", "Ž", "Ź" ),
 229                 'el' => array(),
 230                 'en' => array(),
 231                 'eo' => array( "Ĉ", "Ĝ", "Ĥ", "Ĵ", "Ŝ", "Ŭ" ),
 232                 'es' => array( "Ñ" ),
 233                 'et' => array( "Š", "Ž", "Õ", "Ä", "Ö", "Ü" ),
 234                 'eu' => array( "Ñ" ),
 235                 'fo' => array( "Á", "Ð", "Í", "Ó", "Ú", "Ý", "Æ", "Ø", "Å" ),
 236                 'fr' => array(),
 237                 'fur' => array( "À", "Á", "Â", "È", "Ì", "Ò", "Ù" ),
 238                 'fy' => array(),
 239                 'ga' => array(),
 240                 'gd' => array(),
 241                 'gl' => array( "CH", "LL", "Ñ" ),
 242                 'hr' => array( "Č", "Ć", "DŽ", "Đ", "LJ", "NJ", "Š", "Ž" ),
 243                 'hsb' => array( "Č", "DŹ", "Ě", "CH", "Ł", "Ń", "Ř", "Š", "Ć", "Ž" ),
 244                 'hu' => array( "CS", "DZ", "DZS", "GY", "LY", "NY", "Ö", "SZ", "TY", "Ü", "ZS" ),
 245                 'is' => array( "Á", "Ð", "É", "Í", "Ó", "Ú", "Ý", "Þ", "Æ", "Ö", "Å" ),
 246                 'it' => array(),
 247                 'kk' => array( "Ү", "І" ),
 248                 'kl' => array( "Æ", "Ø", "Å" ),
 249                 'ku' => array( "Ç", "Ê", "Î", "Ş", "Û" ),
 250                 'ky' => array( "Ё" ),
 251                 'la' => array(),
 252                 'lb' => array(),
 253                 'lt' => array( "Č", "Š", "Ž" ),
 254                 'lv' => array( "Č", "Ģ", "Ķ", "Ļ", "Ņ", "Š", "Ž" ),
 255                 'mk' => array(),
 256                 'mo' => array( "Ă", "Â", "Î", "Ş", "Ţ" ),
 257                 'mt' => array( "Ċ", "Ġ", "GĦ", "Ħ", "Ż" ),
 258                 'nl' => array(),
 259                 'no' => array( "Æ", "Ø", "Å" ),
 260                 'oc' => array(),
 261                 'pt' => array(),
 262                 'rm' => array(),
 263                 'ro' => array( "Ă", "Â", "Î", "Ş", "Ţ" ),
 264                 'ru' => array(),
 265                 'rup' => array( "Ă", "Â", "Î", "Ľ", "Ń", "Ş", "Ţ" ),
 266                 'sco' => array(),
 267                 'sk' => array( "Ä", "Č", "CH", "Ô", "Š", "Ž" ),
 268                 'sl' => array( "Č", "Š", "Ž" ),
 269                 'smn' => array( "Á", "Č", "Đ", "Ŋ", "Š", "Ŧ", "Ž", "Æ", "Ø", "Å", "Ä", "Ö" ),
 270                 'sq' => array( "Ç", "DH", "Ë", "GJ", "LL", "NJ", "RR", "SH", "TH", "XH", "ZH" ),
 271                 'sr' => array(),
 272                 'sv' => array( "Å", "Ä", "Ö" ),
 273                 'tk' => array( "Ç", "Ä", "Ž", "Ň", "Ö", "Ş", "Ü", "Ý" ),
 274                 'tl' => array( "Ñ", "NG" ), /* 'fil' in the data source */
 275                 'tr' => array( "Ç", "Ğ", "İ", "Ö", "Ş", "Ü" ),
 276                 'tt' => array( "Ә", "Ө", "Ү", "Җ", "Ң", "Һ" ),
 277                 'uk' => array( "Ґ", "Ь" ),
 278                 'uz' => array( "CH", "G'", "NG", "O'", "SH" ),
 279                 'vi' => array( "Ă", "Â", "Đ", "Ê", "Ô", "Ơ", "Ư" ),
 280         );
 281
 282         const RECORD_LENGTH = 14;
 283
 284         function __construct( $locale ) {
 285                 if ( !extension_loaded( 'intl' ) ) {
 286                         throw new MWException( 'An ICU collation was requested, ' .
 287                                 'but the intl extension is not available.' );
 288                 }
 289                 $this->locale = $locale;
 290                 $this->mainCollator = Collator::create( $locale );
 291                 if ( !$this->mainCollator ) {
 292                         throw new MWException( "Invalid ICU locale specified for collation: $locale" );
 293                 }
 294
 295                 $this->primaryCollator = Collator::create( $locale );
 296                 $this->primaryCollator->setStrength( Collator::PRIMARY );
 297         }
 298
 299         function getSortKey( $string ) {
 300                 // intl extension produces non null-terminated
 301                 // strings. Appending '' fixes it so that it doesn't generate
 302                 // a warning on each access in debug php.
 303                 wfSuppressWarnings();
 304                 $key = $this->mainCollator->getSortKey( $string ) . '';
 305                 wfRestoreWarnings();
 306                 return $key;
 307         }
 308
 309         function getPrimarySortKey( $string ) {
 310                 wfSuppressWarnings();
 311                 $key = $this->primaryCollator->getSortKey( $string ) . '';
 312                 wfRestoreWarnings();
 313                 return $key;
 314         }
 315
 316         function getFirstLetter( $string ) {
 317                 $string = strval( $string );
 318                 if ( $string === '' ) {
 319                         return '';
 320                 }
 321
 322                 // Check for CJK
 323                 $firstChar = mb_substr( $string, 0, 1, 'UTF-8' );
 324                 if ( ord( $firstChar ) > 0x7f
 325                         && self::isCjk( utf8ToCodepoint( $firstChar ) ) )
 326                 {
 327                         return $firstChar;
 328                 }
 329
 330                 $sortKey = $this->getPrimarySortKey( $string );
 331
 332                 // Do a binary search to find the correct letter to sort under
 333                 $min = $this->findLowerBound(
 334                         array( $this, 'getSortKeyByLetterIndex' ),
 335                         $this->getFirstLetterCount(),
 336                         'strcmp',
 337                         $sortKey );
 338
 339                 if ( $min === false ) {
 340                         // Before the first letter
 341                         return '';
 342                 }
 343                 return $this->getLetterByIndex( $min );
 344         }
 345
 346         function getFirstLetterData() {
 347                 if ( $this->firstLetterData !== null ) {
 348                         return $this->firstLetterData;
 349                 }
 350
 351                 $cache = wfGetCache( CACHE_ANYTHING );
 352                 $cacheKey = wfMemcKey( 'first-letters', $this->locale );
 353                 $cacheEntry = $cache->get( $cacheKey );
 354
 355                 if ( $cacheEntry ) {
 356                         $this->firstLetterData = $cacheEntry;
 357                         return $this->firstLetterData;
 358                 }
 359
 360                 // Generate data from serialized data file
 361
 362                 if ( isset ( self::$tailoringFirstLetters[$this->locale] ) ) {
 363                         $letters = wfGetPrecompiledData( "first-letters-root.ser" );
 364                         $letters = array_merge( $letters, self::$tailoringFirstLetters[$this->locale] );
 365                 } else {
 366                         $letters = wfGetPrecompiledData( "first-letters-{$this->locale}.ser" );
 367                         if ( $letters === false ) {
 368                                 throw new MWException( "MediaWiki does not support ICU locale " .
 369                                         "\"{$this->locale}\"" );
 370                         }
 371                 }
 372
 373                 // Sort the letters.
 374                 //
 375                 // It's impossible to have the precompiled data file properly sorted,
 376                 // because the sort order changes depending on ICU version. If the
 377                 // array is not properly sorted, the binary search will return random
 378                 // results.
 379                 //
 380                 // We also take this opportunity to remove primary collisions.
 381                 $letterMap = array();
 382                 foreach ( $letters as $letter ) {
 383                         $key = $this->getPrimarySortKey( $letter );
 384                         if ( isset( $letterMap[$key] ) ) {
 385                                 // Primary collision
 386                                 // Keep whichever one sorts first in the main collator
 387                                 if ( $this->mainCollator->compare( $letter, $letterMap[$key] ) < 0 ) {
 388                                         $letterMap[$key] = $letter;
 389                                 }
 390                         } else {
 391                                 $letterMap[$key] = $letter;
 392                         }
 393                 }
 394                 ksort( $letterMap, SORT_STRING );
 395                 $data = array(
 396                         'chars' => array_values( $letterMap ),
 397                         'keys' => array_keys( $letterMap )
 398                 );
 399
 400                 // Reduce memory usage before caching
 401                 unset( $letterMap );
 402
 403                 // Save to cache
 404                 $this->firstLetterData = $data;
 405                 $cache->set( $cacheKey, $data, 86400 * 7 /* 1 week */ );
 406                 return $data;
 407         }
 408
 409         function getLetterByIndex( $index ) {
 410                 if ( $this->firstLetterData === null ) {
 411                         $this->getFirstLetterData();
 412                 }
 413                 return $this->firstLetterData['chars'][$index];
 414         }
 415
 416         function getSortKeyByLetterIndex( $index ) {
 417                 if ( $this->firstLetterData === null ) {
 418                         $this->getFirstLetterData();
 419                 }
 420                 return $this->firstLetterData['keys'][$index];
 421         }
 422
 423         function getFirstLetterCount() {
 424                 if ( $this->firstLetterData === null ) {
 425                         $this->getFirstLetterData();
 426                 }
 427                 return count( $this->firstLetterData['chars'] );
 428         }
 429
 430         /**
 431          * Do a binary search, and return the index of the largest item that sorts
 432          * less than or equal to the target value.
 433          *
 434          * @param $valueCallback array A function to call to get the value with
 435          *     a given array index.
 436          * @param $valueCount int The number of items accessible via $valueCallback,
 437          *     indexed from 0 to $valueCount - 1
 438          * @param $comparisonCallback array A callback to compare two values, returning
 439          *     -1, 0 or 1 in the style of strcmp().
 440          * @param $target string The target value to find.
 441          *
 442          * @return int|bool The item index of the lower bound, or false if the target value
 443          *     sorts before all items.
 444          */
 445         function findLowerBound( $valueCallback, $valueCount, $comparisonCallback, $target ) {
 446                 if ( $valueCount === 0 ) {
 447                         return false;
 448                 }
 449
 450                 $min = 0;
 451                 $max = $valueCount;
 452                 do {
 453                         $mid = $min + ( ( $max - $min ) >> 1 );
 454                         $item = call_user_func( $valueCallback, $mid );
 455                         $comparison = call_user_func( $comparisonCallback, $target, $item );
 456                         if ( $comparison > 0 ) {
 457                                 $min = $mid;
 458                         } elseif ( $comparison == 0 ) {
 459                                 $min = $mid;
 460                                 break;
 461                         } else {
 462                                 $max = $mid;
 463                         }
 464                 } while ( $min < $max - 1 );
 465
 466                 if ( $min == 0 ) {
 467                         $item = call_user_func( $valueCallback, $min );
 468                         $comparison = call_user_func( $comparisonCallback, $target, $item );
 469                         if ( $comparison < 0 ) {
 470                                 // Before the first item
 471                                 return false;
 472                         }
 473                 }
 474                 return $min;
 475         }
 476
 477         static function isCjk( $codepoint ) {
 478                 foreach ( self::$cjkBlocks as $block ) {
 479                         if ( $codepoint >= $block[0] && $codepoint <= $block[1] ) {
 480                                 return true;
 481                         }
 482                 }
 483                 return false;
 484         }
 485
 486         /**
 487          * Return the version of ICU library used by PHP's intl extension,
 488          * or false when the extension is not installed of the version
 489          * can't be determined.
 490          *
 491          * The constant INTL_ICU_VERSION this function refers to isn't really
 492          * documented. It is available since PHP 5.3.7 (see PHP bug 54561).
 493          * This function will return false on older PHPs.
 494          *
 495          * @since 1.21
 496          * @return string|false
 497          */
 498         static function getICUVersion() {
 499                 return defined( 'INTL_ICU_VERSION' ) ? INTL_ICU_VERSION : false;
 500         }
 501
 502         /**
 503          * Return the version of Unicode appropriate for the version of ICU library
 504          * currently in use, or false when it can't be determined.
 505          *
 506          * @since 1.21
 507          * @return string|false
 508          */
 509         static function getUnicodeVersionForICU() {
 510                 $icuVersion = IcuCollation::getICUVersion();
 511                 if ( !$icuVersion ) {
 512                         return false;
 513                 }
 514
 515                 $versionPrefix = substr( $icuVersion, 0, 3 );
 516                 // Source: http://site.icu-project.org/download
 517                 $map = array(
 518                         '50.' => '6.2',
 519                         '49.' => '6.1',
 520                         '4.8' => '6.0',
 521                         '4.6' => '6.0',
 522                         '4.4' => '5.2',
 523                         '4.2' => '5.1',
 524                         '4.0' => '5.1',
 525                         '3.8' => '5.0',
 526                         '3.6' => '5.0',
 527                         '3.4' => '4.1',
 528                 );
 529
 530                 if ( isset( $map[$versionPrefix] ) ) {
 531                         return $map[$versionPrefix];
 532                 } else {
 533                         return false;
 534                 }
 535         }
 536 }