languages/ConverterRule.php

   1 <?php
   2 /**
   3  * This program is free software; you can redistribute it and/or modify
   4  * it under the terms of the GNU General Public License as published by
   5  * the Free Software Foundation; either version 2 of the License, or
   6  * (at your option) any later version.
   7  *
   8  * This program is distributed in the hope that it will be useful,
   9  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  10  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
  11  * GNU General Public License for more details.
  12  *
  13  * You should have received a copy of the GNU General Public License along
  14  * with this program; if not, write to the Free Software Foundation, Inc.,
  15  * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
  16  * http://www.gnu.org/copyleft/gpl.html
  17  *
  18  * @file
  19  * @ingroup Language
  20  */
  21
  22 /**
  23  * Parser for rules of language conversion , parse rules in -{ }- tag.
  24  * @ingroup Language
  25  * @author fdcn <fdcn64@gmail.com>, PhiLiP <philip.npc@gmail.com>
  26  */
  27 class ConverterRule {
  28         public $mText; // original text in -{text}-
  29         public $mConverter; // LanguageConverter object
  30         public $mRuleDisplay = '';
  31         public $mRuleTitle = false;
  32         public $mRules = '';// string : the text of the rules
  33         public $mRulesAction = 'none';
  34         public $mFlags = array();
  35         public $mVariantFlags = array();
  36         public $mConvTable = array();
  37         public $mBidtable = array();// array of the translation in each variant
  38         public $mUnidtable = array();// array of the translation in each variant
  39
  40         /**
  41          * Constructor
  42          *
  43          * @param string $text The text between -{ and }-
  44          * @param LanguageConverter $converter
  45          */
  46         public function __construct( $text, $converter ) {
  47                 $this->mText = $text;
  48                 $this->mConverter = $converter;
  49         }
  50
  51         /**
  52          * Check if variants array in convert array.
  53          *
  54          * @param array|string $variants Variant language code
  55          * @return string Translated text
  56          */
  57         public function getTextInBidtable( $variants ) {
  58                 $variants = (array)$variants;
  59                 if ( !$variants ) {
  60                         return false;
  61                 }
  62                 foreach ( $variants as $variant ) {
  63                         if ( isset( $this->mBidtable[$variant] ) ) {
  64                                 return $this->mBidtable[$variant];
  65                         }
  66                 }
  67                 return false;
  68         }
  69
  70         /**
  71          * Parse flags with syntax -{FLAG| ... }-
  72          * @private
  73          */
  74         function parseFlags() {
  75                 $text = $this->mText;
  76                 $flags = array();
  77                 $variantFlags = array();
  78
  79                 $sepPos = strpos( $text, '|' );
  80                 if ( $sepPos !== false ) {
  81                         $validFlags = $this->mConverter->mFlags;
  82                         $f = StringUtils::explode( ';', substr( $text, 0, $sepPos ) );
  83                         foreach ( $f as $ff ) {
  84                                 $ff = trim( $ff );
  85                                 if ( isset( $validFlags[$ff] ) ) {
  86                                         $flags[$validFlags[$ff]] = true;
  87                                 }
  88                         }
  89                         $text = strval( substr( $text, $sepPos + 1 ) );
  90                 }
  91
  92                 if ( !$flags ) {
  93                         $flags['S'] = true;
  94                 } elseif ( isset( $flags['R'] ) ) {
  95                         $flags = array( 'R' => true );// remove other flags
  96                 } elseif ( isset( $flags['N'] ) ) {
  97                         $flags = array( 'N' => true );// remove other flags
  98                 } elseif ( isset( $flags['-'] ) ) {
  99                         $flags = array( '-' => true );// remove other flags
 100                 } elseif ( count( $flags ) == 1 && isset( $flags['T'] ) ) {
 101                         $flags['H'] = true;
 102                 } elseif ( isset( $flags['H'] ) ) {
 103                         // replace A flag, and remove other flags except T
 104                         $temp = array( '+' => true, 'H' => true );
 105                         if ( isset( $flags['T'] ) ) {
 106                                 $temp['T'] = true;
 107                         }
 108                         if ( isset( $flags['D'] ) ) {
 109                                 $temp['D'] = true;
 110                         }
 111                         $flags = $temp;
 112                 } else {
 113                         if ( isset( $flags['A'] ) ) {
 114                                 $flags['+'] = true;
 115                                 $flags['S'] = true;
 116                         }
 117                         if ( isset( $flags['D'] ) ) {
 118                                 unset( $flags['S'] );
 119                         }
 120                         // try to find flags like "zh-hans", "zh-hant"
 121                         // allow syntaxes like "-{zh-hans;zh-hant|XXXX}-"
 122                         $variantFlags = array_intersect( array_keys( $flags ), $this->mConverter->mVariants );
 123                         if ( $variantFlags ) {
 124                                 $variantFlags = array_flip( $variantFlags );
 125                                 $flags = array();
 126                         }
 127                 }
 128                 $this->mVariantFlags = $variantFlags;
 129                 $this->mRules = $text;
 130                 $this->mFlags = $flags;
 131         }
 132
 133         /**
 134          * Generate conversion table.
 135          * @private
 136          */
 137         function parseRules() {
 138                 $rules = $this->mRules;
 139                 $bidtable = array();
 140                 $unidtable = array();
 141                 $variants = $this->mConverter->mVariants;
 142                 $varsep_pattern = $this->mConverter->getVarSeparatorPattern();
 143
 144                 // Split according to $varsep_pattern, but ignore semicolons from HTML entities
 145                 $rules = preg_replace( '/(&[#a-zA-Z0-9]+);/', "$1\x01", $rules );
 146                 $choice = preg_split( $varsep_pattern, $rules );
 147                 $choice = str_replace( "\x01", ';', $choice );
 148
 149                 foreach ( $choice as $c ) {
 150                         $v = explode( ':', $c, 2 );
 151                         if ( count( $v ) != 2 ) {
 152                                 // syntax error, skip
 153                                 continue;
 154                         }
 155                         $to = trim( $v[1] );
 156                         $v = trim( $v[0] );
 157                         $u = explode( '=>', $v, 2 );
 158                         // if $to is empty (which is also used as $from in bidtable),
 159                         // strtr() could return a wrong result.
 160                         if ( count( $u ) == 1 && $to !== '' && in_array( $v, $variants ) ) {
 161                                 $bidtable[$v] = $to;
 162                         } elseif ( count( $u ) == 2 ) {
 163                                 $from = trim( $u[0] );
 164                                 $v = trim( $u[1] );
 165                                 // if $from is empty, strtr() could return a wrong result.
 166                                 if ( array_key_exists( $v, $unidtable )
 167                                         && !is_array( $unidtable[$v] )
 168                                         && $from !== ''
 169                                         && in_array( $v, $variants ) ) {
 170                                         $unidtable[$v] = array( $from => $to );
 171                                 } elseif ( $from !== '' && in_array( $v, $variants ) ) {
 172                                         $unidtable[$v][$from] = $to;
 173                                 }
 174                         }
 175                         // syntax error, pass
 176                         if ( !isset( $this->mConverter->mVariantNames[$v] ) ) {
 177                                 $bidtable = array();
 178                                 $unidtable = array();
 179                                 break;
 180                         }
 181                 }
 182                 $this->mBidtable = $bidtable;
 183                 $this->mUnidtable = $unidtable;
 184         }
 185
 186         /**
 187          * @private
 188          *
 189          * @return string
 190          */
 191         function getRulesDesc() {
 192                 $codesep = $this->mConverter->mDescCodeSep;
 193                 $varsep = $this->mConverter->mDescVarSep;
 194                 $text = '';
 195                 foreach ( $this->mBidtable as $k => $v ) {
 196                         $text .= $this->mConverter->mVariantNames[$k] . "$codesep$v$varsep";
 197                 }
 198                 foreach ( $this->mUnidtable as $k => $a ) {
 199                         foreach ( $a as $from => $to ) {
 200                                 $text .= $from . '⇒' . $this->mConverter->mVariantNames[$k] .
 201                                         "$codesep$to$varsep";
 202                         }
 203                 }
 204                 return $text;
 205         }
 206
 207         /**
 208          * Parse rules conversion.
 209          * @private
 210          *
 211          * @param string $variant
 212          *
 213          * @return string
 214          */
 215         function getRuleConvertedStr( $variant ) {
 216                 $bidtable = $this->mBidtable;
 217                 $unidtable = $this->mUnidtable;
 218
 219                 if ( count( $bidtable ) + count( $unidtable ) == 0 ) {
 220                         return $this->mRules;
 221                 } else {
 222                         // display current variant in bidirectional array
 223                         $disp = $this->getTextInBidtable( $variant );
 224                         // or display current variant in fallbacks
 225                         if ( $disp === false ) {
 226                                 $disp = $this->getTextInBidtable(
 227                                         $this->mConverter->getVariantFallbacks( $variant ) );
 228                         }
 229                         // or display current variant in unidirectional array
 230                         if ( $disp === false && array_key_exists( $variant, $unidtable ) ) {
 231                                 $disp = array_values( $unidtable[$variant] );
 232                                 $disp = $disp[0];
 233                         }
 234                         // or display frist text under disable manual convert
 235                         if ( $disp === false && $this->mConverter->mManualLevel[$variant] == 'disable' ) {
 236                                 if ( count( $bidtable ) > 0 ) {
 237                                         $disp = array_values( $bidtable );
 238                                         $disp = $disp[0];
 239                                 } else {
 240                                         $disp = array_values( $unidtable );
 241                                         $disp = array_values( $disp[0] );
 242                                         $disp = $disp[0];
 243                                 }
 244                         }
 245                         return $disp;
 246                 }
 247         }
 248
 249         /**
 250          * Similar to getRuleConvertedStr(), but this prefers to use original
 251          * page title if $variant === $this->mConverter->mMainLanguageCode
 252          * and may return false in this case (so this title conversion rule
 253          * will be ignored and the original title is shown).
 254          *
 255          * @since 1.22
 256          * @param string $variant The variant code to display page title in
 257          * @return string|bool The converted title or false if just page name
 258          */
 259         function getRuleConvertedTitle( $variant ) {
 260                 if ( $variant === $this->mConverter->mMainLanguageCode ) {
 261                         // If a string targeting exactly this variant is set,
 262                         // use it. Otherwise, just return false, so the real
 263                         // page name can be shown (and because variant === main,
 264                         // there'll be no further automatic conversion).
 265                         $disp = $this->getTextInBidtable( $variant );
 266                         if ( $disp ) {
 267                                 return $disp;
 268                         }
 269                         if ( array_key_exists( $variant, $this->mUnidtable ) ) {
 270                                 $disp = array_values( $this->mUnidtable[$variant] );
 271                                 $disp = $disp[0];
 272                         }
 273                         // Assigned above or still false.
 274                         return $disp;
 275                 } else {
 276                         return $this->getRuleConvertedStr( $variant );
 277                 }
 278         }
 279
 280         /**
 281          * Generate conversion table for all text.
 282          * @private
 283          */
 284         function generateConvTable() {
 285                 // Special case optimisation
 286                 if ( !$this->mBidtable && !$this->mUnidtable ) {
 287                         $this->mConvTable = array();
 288                         return;
 289                 }
 290
 291                 $bidtable = $this->mBidtable;
 292                 $unidtable = $this->mUnidtable;
 293                 $manLevel = $this->mConverter->mManualLevel;
 294
 295                 $vmarked = array();
 296                 foreach ( $this->mConverter->mVariants as $v ) {
 297                         /* for bidirectional array
 298                                 fill in the missing variants, if any,
 299                                 with fallbacks */
 300                         if ( !isset( $bidtable[$v] ) ) {
 301                                 $variantFallbacks =
 302                                         $this->mConverter->getVariantFallbacks( $v );
 303                                 $vf = $this->getTextInBidtable( $variantFallbacks );
 304                                 if ( $vf ) {
 305                                         $bidtable[$v] = $vf;
 306                                 }
 307                         }
 308
 309                         if ( isset( $bidtable[$v] ) ) {
 310                                 foreach ( $vmarked as $vo ) {
 311                                         // use syntax: -{A|zh:WordZh;zh-tw:WordTw}-
 312                                         // or -{H|zh:WordZh;zh-tw:WordTw}-
 313                                         // or -{-|zh:WordZh;zh-tw:WordTw}-
 314                                         // to introduce a custom mapping between
 315                                         // words WordZh and WordTw in the whole text
 316                                         if ( $manLevel[$v] == 'bidirectional' ) {
 317                                                 $this->mConvTable[$v][$bidtable[$vo]] = $bidtable[$v];
 318                                         }
 319                                         if ( $manLevel[$vo] == 'bidirectional' ) {
 320                                                 $this->mConvTable[$vo][$bidtable[$v]] = $bidtable[$vo];
 321                                         }
 322                                 }
 323                                 $vmarked[] = $v;
 324                         }
 325                         /* for unidirectional array fill to convert tables */
 326                         if ( ( $manLevel[$v] == 'bidirectional' || $manLevel[$v] == 'unidirectional' )
 327                                 && isset( $unidtable[$v] )
 328                         ) {
 329                                 if ( isset( $this->mConvTable[$v] ) ) {
 330                                         $this->mConvTable[$v] = $unidtable[$v] + $this->mConvTable[$v];
 331                                 } else {
 332                                         $this->mConvTable[$v] = $unidtable[$v];
 333                                 }
 334                         }
 335                 }
 336         }
 337
 338         /**
 339          * Parse rules and flags.
 340          * @param string $variant Variant language code
 341          */
 342         public function parse( $variant = null ) {
 343                 if ( !$variant ) {
 344                         $variant = $this->mConverter->getPreferredVariant();
 345                 }
 346
 347                 $this->parseFlags();
 348                 $flags = $this->mFlags;
 349
 350                 // convert to specified variant
 351                 // syntax: -{zh-hans;zh-hant[;...]|<text to convert>}-
 352                 if ( $this->mVariantFlags ) {
 353                         // check if current variant in flags
 354                         if ( isset( $this->mVariantFlags[$variant] ) ) {
 355                                 // then convert <text to convert> to current language
 356                                 $this->mRules = $this->mConverter->autoConvert( $this->mRules,
 357                                         $variant );
 358                         } else {
 359                                 // if current variant no in flags,
 360                                 // then we check its fallback variants.
 361                                 $variantFallbacks =
 362                                         $this->mConverter->getVariantFallbacks( $variant );
 363                                 if ( is_array( $variantFallbacks ) ) {
 364                                         foreach ( $variantFallbacks as $variantFallback ) {
 365                                                 // if current variant's fallback exist in flags
 366                                                 if ( isset( $this->mVariantFlags[$variantFallback] ) ) {
 367                                                         // then convert <text to convert> to fallback language
 368                                                         $this->mRules =
 369                                                                 $this->mConverter->autoConvert( $this->mRules,
 370                                                                         $variantFallback );
 371                                                         break;
 372                                                 }
 373                                         }
 374                                 }
 375                         }
 376                         $this->mFlags = $flags = array( 'R' => true );
 377                 }
 378
 379                 if ( !isset( $flags['R'] ) && !isset( $flags['N'] ) ) {
 380                         // decode => HTML entities modified by Sanitizer::removeHTMLtags
 381                         $this->mRules = str_replace( '=&gt;', '=>', $this->mRules );
 382                         $this->parseRules();
 383                 }
 384                 $rules = $this->mRules;
 385
 386                 if ( !$this->mBidtable && !$this->mUnidtable ) {
 387                         if ( isset( $flags['+'] ) || isset( $flags['-'] ) ) {
 388                                 // fill all variants if text in -{A/H/-|text}- is non-empty but without rules
 389                                 if ( $rules !== '' ) {
 390                                         foreach ( $this->mConverter->mVariants as $v ) {
 391                                                 $this->mBidtable[$v] = $rules;
 392                                         }
 393                                 }
 394                         } elseif ( !isset( $flags['N'] ) && !isset( $flags['T'] ) ) {
 395                                 $this->mFlags = $flags = array( 'R' => true );
 396                         }
 397                 }
 398
 399                 $this->mRuleDisplay = false;
 400                 foreach ( $flags as $flag => $unused ) {
 401                         switch ( $flag ) {
 402                                 case 'R':
 403                                         // if we don't do content convert, still strip the -{}- tags
 404                                         $this->mRuleDisplay = $rules;
 405                                         break;
 406                                 case 'N':
 407                                         // process N flag: output current variant name
 408                                         $ruleVar = trim( $rules );
 409                                         if ( isset( $this->mConverter->mVariantNames[$ruleVar] ) ) {
 410                                                 $this->mRuleDisplay = $this->mConverter->mVariantNames[$ruleVar];
 411                                         } else {
 412                                                 $this->mRuleDisplay = '';
 413                                         }
 414                                         break;
 415                                 case 'D':
 416                                         // process D flag: output rules description
 417                                         $this->mRuleDisplay = $this->getRulesDesc();
 418                                         break;
 419                                 case 'H':
 420                                         // process H,- flag or T only: output nothing
 421                                         $this->mRuleDisplay = '';
 422                                         break;
 423                                 case '-':
 424                                         $this->mRulesAction = 'remove';
 425                                         $this->mRuleDisplay = '';
 426                                         break;
 427                                 case '+':
 428                                         $this->mRulesAction = 'add';
 429                                         $this->mRuleDisplay = '';
 430                                         break;
 431                                 case 'S':
 432                                         $this->mRuleDisplay = $this->getRuleConvertedStr( $variant );
 433                                         break;
 434                                 case 'T':
 435                                         $this->mRuleTitle = $this->getRuleConvertedTitle( $variant );
 436                                         $this->mRuleDisplay = '';
 437                                         break;
 438                                 default:
 439                                         // ignore unknown flags (but see error case below)
 440                         }
 441                 }
 442                 if ( $this->mRuleDisplay === false ) {
 443                         $this->mRuleDisplay = '<span class="error">'
 444                                 . wfMessage( 'converter-manual-rule-error' )->inContentLanguage()->escaped()
 445                                 . '</span>';
 446                 }
 447
 448                 $this->generateConvTable();
 449         }
 450
 451         /**
 452          * Checks if there are conversion rules.
 453          * @return bool
 454          */
 455         public function hasRules() {
 456                 return $this->mRules !== '';
 457         }
 458
 459         /**
 460          * Get display text on markup -{...}-
 461          * @return string
 462          */
 463         public function getDisplay() {
 464                 return $this->mRuleDisplay;
 465         }
 466
 467         /**
 468          * Get converted title.
 469          * @return string
 470          */
 471         public function getTitle() {
 472                 return $this->mRuleTitle;
 473         }
 474
 475         /**
 476          * Return how deal with conversion rules.
 477          * @return string
 478          */
 479         public function getRulesAction() {
 480                 return $this->mRulesAction;
 481         }
 482
 483         /**
 484          * Get conversion table. (bidirectional and unidirectional
 485          * conversion table)
 486          * @return array
 487          */
 488         public function getConvTable() {
 489                 return $this->mConvTable;
 490         }
 491
 492         /**
 493          * Get conversion rules string.
 494          * @return string
 495          */
 496         public function getRules() {
 497                 return $this->mRules;
 498         }
 499
 500         /**
 501          * Get conversion flags.
 502          * @return array
 503          */
 504         public function getFlags() {
 505                 return $this->mFlags;
 506         }
 507 }