includes/parser/DateFormatter.php

   1 <?php
   2 /**
   3  * Date formatter
   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  * @ingroup Parser
  22  */
  23
  24 /**
  25  * Date formatter, recognises dates in plain text and formats them according to user preferences.
  26  * @todo preferences, OutputPage
  27  * @ingroup Parser
  28  */
  29 class DateFormatter {
  30         private $mSource, $mTarget;
  31         private $monthNames = '';
  32
  33         private $regexes;
  34         private $rules, $xMonths, $preferences;
  35
  36         private $lang, $mLinked;
  37
  38         /** @var string[] */
  39         private $keys;
  40
  41         /** @var string[] */
  42         private $targets;
  43
  44         const ALL = -1;
  45         const NONE = 0;
  46         const MDY = 1;
  47         const DMY = 2;
  48         const YMD = 3;
  49         const ISO1 = 4;
  50         const LASTPREF = 4;
  51         const ISO2 = 5;
  52         const YDM = 6;
  53         const DM = 7;
  54         const MD = 8;
  55         const LAST = 8;
  56
  57         /**
  58          * @param Language $lang In which language to format the date
  59          */
  60         public function __construct( Language $lang ) {
  61                 $this->lang = $lang;
  62
  63                 $this->monthNames = $this->getMonthRegex();
  64                 for ( $i = 1; $i <= 12; $i++ ) {
  65                         $this->xMonths[$this->lang->lc( $this->lang->getMonthName( $i ) )] = $i;
  66                         $this->xMonths[$this->lang->lc( $this->lang->getMonthAbbreviation( $i ) )] = $i;
  67                 }
  68
  69                 $this->regexTrail = '(?![a-z])/iu';
  70
  71                 # Partial regular expressions
  72                 $this->prxDM = '\[\[(\d{1,2})[ _](' . $this->monthNames . ')\]\]';
  73                 $this->prxMD = '\[\[(' . $this->monthNames . ')[ _](\d{1,2})\]\]';
  74                 $this->prxY = '\[\[(\d{1,4}([ _]BC|))\]\]';
  75                 $this->prxISO1 = '\[\[(-?\d{4})]]-\[\[(\d{2})-(\d{2})\]\]';
  76                 $this->prxISO2 = '\[\[(-?\d{4})-(\d{2})-(\d{2})\]\]';
  77
  78                 # Real regular expressions
  79                 $this->regexes[self::DMY] = "/{$this->prxDM}(?: *, *| +){$this->prxY}{$this->regexTrail}";
  80                 $this->regexes[self::YDM] = "/{$this->prxY}(?: *, *| +){$this->prxDM}{$this->regexTrail}";
  81                 $this->regexes[self::MDY] = "/{$this->prxMD}(?: *, *| +){$this->prxY}{$this->regexTrail}";
  82                 $this->regexes[self::YMD] = "/{$this->prxY}(?: *, *| +){$this->prxMD}{$this->regexTrail}";
  83                 $this->regexes[self::DM] = "/{$this->prxDM}{$this->regexTrail}";
  84                 $this->regexes[self::MD] = "/{$this->prxMD}{$this->regexTrail}";
  85                 $this->regexes[self::ISO1] = "/{$this->prxISO1}{$this->regexTrail}";
  86                 $this->regexes[self::ISO2] = "/{$this->prxISO2}{$this->regexTrail}";
  87
  88                 # Extraction keys
  89                 # See the comments in replace() for the meaning of the letters
  90                 $this->keys[self::DMY] = 'jFY';
  91                 $this->keys[self::YDM] = 'Y jF';
  92                 $this->keys[self::MDY] = 'FjY';
  93                 $this->keys[self::YMD] = 'Y Fj';
  94                 $this->keys[self::DM] = 'jF';
  95                 $this->keys[self::MD] = 'Fj';
  96                 $this->keys[self::ISO1] = 'ymd'; # y means ISO year
  97                 $this->keys[self::ISO2] = 'ymd';
  98
  99                 # Target date formats
 100                 $this->targets[self::DMY] = '[[F j|j F]] [[Y]]';
 101                 $this->targets[self::YDM] = '[[Y]], [[F j|j F]]';
 102                 $this->targets[self::MDY] = '[[F j]], [[Y]]';
 103                 $this->targets[self::YMD] = '[[Y]] [[F j]]';
 104                 $this->targets[self::DM] = '[[F j|j F]]';
 105                 $this->targets[self::MD] = '[[F j]]';
 106                 $this->targets[self::ISO1] = '[[Y|y]]-[[F j|m-d]]';
 107                 $this->targets[self::ISO2] = '[[y-m-d]]';
 108
 109                 # Rules
 110                 #            pref    source       target
 111                 $this->rules[self::DMY][self::MD] = self::DM;
 112                 $this->rules[self::ALL][self::MD] = self::MD;
 113                 $this->rules[self::MDY][self::DM] = self::MD;
 114                 $this->rules[self::ALL][self::DM] = self::DM;
 115                 $this->rules[self::NONE][self::ISO2] = self::ISO1;
 116
 117                 $this->preferences = [
 118                         'default' => self::NONE,
 119                         'dmy' => self::DMY,
 120                         'mdy' => self::MDY,
 121                         'ymd' => self::YMD,
 122                         'ISO 8601' => self::ISO1,
 123                 ];
 124         }
 125
 126         /**
 127          * Get a DateFormatter object
 128          *
 129          * @param Language|string|null $lang In which language to format the date
 130          *              Defaults to the site content language
 131          * @return DateFormatter
 132          */
 133         public static function getInstance( $lang = null ) {
 134                 global $wgContLang, $wgMainCacheType;
 135
 136                 $lang = $lang ? wfGetLangObj( $lang ) : $wgContLang;
 137                 $cache = ObjectCache::getLocalServerInstance( $wgMainCacheType );
 138
 139                 static $dateFormatter = false;
 140                 if ( !$dateFormatter ) {
 141                         $dateFormatter = $cache->getWithSetCallback(
 142                                 $cache->makeKey( 'dateformatter', $lang->getCode() ),
 143                                 $cache::TTL_HOUR,
 144                                 function () use ( $lang ) {
 145                                         return new DateFormatter( $lang );
 146                                 }
 147                         );
 148                 }
 149
 150                 return $dateFormatter;
 151         }
 152
 153         /**
 154          * @param string $preference User preference
 155          * @param string $text Text to reformat
 156          * @param array $options Array can contain 'linked' and/or 'match-whole'
 157          *
 158          * @return string
 159          */
 160         public function reformat( $preference, $text, $options = [ 'linked' ] ) {
 161                 $linked = in_array( 'linked', $options );
 162                 $match_whole = in_array( 'match-whole', $options );
 163
 164                 if ( isset( $this->preferences[$preference] ) ) {
 165                         $preference = $this->preferences[$preference];
 166                 } else {
 167                         $preference = self::NONE;
 168                 }
 169                 for ( $i = 1; $i <= self::LAST; $i++ ) {
 170                         $this->mSource = $i;
 171                         if ( isset( $this->rules[$preference][$i] ) ) {
 172                                 # Specific rules
 173                                 $this->mTarget = $this->rules[$preference][$i];
 174                         } elseif ( isset( $this->rules[self::ALL][$i] ) ) {
 175                                 # General rules
 176                                 $this->mTarget = $this->rules[self::ALL][$i];
 177                         } elseif ( $preference ) {
 178                                 # User preference
 179                                 $this->mTarget = $preference;
 180                         } else {
 181                                 # Default
 182                                 $this->mTarget = $i;
 183                         }
 184                         $regex = $this->regexes[$i];
 185
 186                         // Horrible hack
 187                         if ( !$linked ) {
 188                                 $regex = str_replace( [ '\[\[', '\]\]' ], '', $regex );
 189                         }
 190
 191                         if ( $match_whole ) {
 192                                 // Let's hope this works
 193                                 $regex = preg_replace( '!^/!', '/^', $regex );
 194                                 $regex = str_replace( $this->regexTrail,
 195                                         '$' . $this->regexTrail, $regex );
 196                         }
 197
 198                         // Another horrible hack
 199                         $this->mLinked = $linked;
 200                         $text = preg_replace_callback( $regex, [ &$this, 'replace' ], $text );
 201                         unset( $this->mLinked );
 202                 }
 203                 return $text;
 204         }
 205
 206         /**
 207          * Regexp replacement callback
 208          *
 209          * @param array $matches
 210          * @return string
 211          */
 212         private function replace( $matches ) {
 213                 # Extract information from $matches
 214                 $linked = true;
 215                 if ( isset( $this->mLinked ) ) {
 216                         $linked = $this->mLinked;
 217                 }
 218
 219                 $bits = [];
 220                 $key = $this->keys[$this->mSource];
 221                 $keyLength = strlen( $key );
 222                 for ( $p = 0; $p < $keyLength; $p++ ) {
 223                         if ( $key[$p] != ' ' ) {
 224                                 $bits[$key[$p]] = $matches[$p + 1];
 225                         }
 226                 }
 227
 228                 return $this->formatDate( $bits, $matches[0], $linked );
 229         }
 230
 231         /**
 232          * @param array $bits
 233          * @param string $orig Original input string, to be returned
 234          *  on formatting failure.
 235          * @param bool $link
 236          * @return string
 237          */
 238         private function formatDate( $bits, $orig, $link = true ) {
 239                 $format = $this->targets[$this->mTarget];
 240
 241                 if ( !$link ) {
 242                         // strip piped links
 243                         $format = preg_replace( '/\[\[[^|]+\|([^\]]+)\]\]/', '$1', $format );
 244                         // strip remaining links
 245                         $format = str_replace( [ '[[', ']]' ], '', $format );
 246                 }
 247
 248                 # Construct new date
 249                 $text = '';
 250                 $fail = false;
 251
 252                 // Pre-generate y/Y stuff because we need the year for the <span> title.
 253                 if ( !isset( $bits['y'] ) && isset( $bits['Y'] ) ) {
 254                         $bits['y'] = $this->makeIsoYear( $bits['Y'] );
 255                 }
 256                 if ( !isset( $bits['Y'] ) && isset( $bits['y'] ) ) {
 257                         $bits['Y'] = $this->makeNormalYear( $bits['y'] );
 258                 }
 259
 260                 if ( !isset( $bits['m'] ) ) {
 261                         $m = $this->makeIsoMonth( $bits['F'] );
 262                         if ( !$m || $m == '00' ) {
 263                                 $fail = true;
 264                         } else {
 265                                 $bits['m'] = $m;
 266                         }
 267                 }
 268
 269                 if ( !isset( $bits['d'] ) ) {
 270                         $bits['d'] = sprintf( '%02d', $bits['j'] );
 271                 }
 272
 273                 $formatLength = strlen( $format );
 274                 for ( $p = 0; $p < $formatLength; $p++ ) {
 275                         $char = $format[$p];
 276                         switch ( $char ) {
 277                                 case 'd': # ISO day of month
 278                                         $text .= $bits['d'];
 279                                         break;
 280                                 case 'm': # ISO month
 281                                         $text .= $bits['m'];
 282                                         break;
 283                                 case 'y': # ISO year
 284                                         $text .= $bits['y'];
 285                                         break;
 286                                 case 'j': # ordinary day of month
 287                                         if ( !isset( $bits['j'] ) ) {
 288                                                 $text .= intval( $bits['d'] );
 289                                         } else {
 290                                                 $text .= $bits['j'];
 291                                         }
 292                                         break;
 293                                 case 'F': # long month
 294                                         if ( !isset( $bits['F'] ) ) {
 295                                                 $m = intval( $bits['m'] );
 296                                                 if ( $m > 12 || $m < 1 ) {
 297                                                         $fail = true;
 298                                                 } else {
 299                                                         $text .= $this->lang->getMonthName( $m );
 300                                                 }
 301                                         } else {
 302                                                 $text .= ucfirst( $bits['F'] );
 303                                         }
 304                                         break;
 305                                 case 'Y': # ordinary (optional BC) year
 306                                         $text .= $bits['Y'];
 307                                         break;
 308                                 default:
 309                                         $text .= $char;
 310                         }
 311                 }
 312                 if ( $fail ) {
 313                         // This occurs when parsing a date with day or month outside the bounds
 314                         // of possibilities.
 315                         $text = $orig;
 316                 }
 317
 318                 $isoBits = [];
 319                 if ( isset( $bits['y'] ) ) {
 320                         $isoBits[] = $bits['y'];
 321                 }
 322                 $isoBits[] = $bits['m'];
 323                 $isoBits[] = $bits['d'];
 324                 $isoDate = implode( '-', $isoBits );
 325
 326                 // Output is not strictly HTML (it's wikitext), but <span> is whitelisted.
 327                 $text = Html::rawElement( 'span',
 328                                         [ 'class' => 'mw-formatted-date', 'title' => $isoDate ], $text );
 329
 330                 return $text;
 331         }
 332
 333         /**
 334          * Return a regex that can be used to find month names in string
 335          * @return string regex to find the months with
 336          */
 337         private function getMonthRegex() {
 338                 $names = [];
 339                 for ( $i = 1; $i <= 12; $i++ ) {
 340                         $names[] = $this->lang->getMonthName( $i );
 341                         $names[] = $this->lang->getMonthAbbreviation( $i );
 342                 }
 343                 return implode( '|', $names );
 344         }
 345
 346         /**
 347          * Makes an ISO month, e.g. 02, from a month name
 348          * @param string $monthName Month name
 349          * @return string ISO month name
 350          */
 351         private function makeIsoMonth( $monthName ) {
 352                 $n = $this->xMonths[$this->lang->lc( $monthName )];
 353                 return sprintf( '%02d', $n );
 354         }
 355
 356         /**
 357          * Make an ISO year from a year name, for instance: '-1199' from '1200 BC'
 358          * @param string $year Year name
 359          * @return string ISO year name
 360          */
 361         private function makeIsoYear( $year ) {
 362                 # Assumes the year is in a nice format, as enforced by the regex
 363                 if ( substr( $year, -2 ) == 'BC' ) {
 364                         $num = intval( substr( $year, 0, -3 ) ) - 1;
 365                         # PHP bug note: sprintf( "%04d", -1 ) fails poorly
 366                         $text = sprintf( '-%04d', $num );
 367
 368                 } else {
 369                         $text = sprintf( '%04d', $year );
 370                 }
 371                 return $text;
 372         }
 373
 374         /**
 375          * Make a year one from an ISO year, for instance: '400 BC' from '-0399'.
 376          * @param string $iso ISO year
 377          * @return int|string int representing year number in case of AD dates, or string containing
 378          *   year number and 'BC' at the end otherwise.
 379          */
 380         private function makeNormalYear( $iso ) {
 381                 if ( $iso[0] == '-' ) {
 382                         $text = ( intval( substr( $iso, 1 ) ) + 1 ) . ' BC';
 383                 } else {
 384                         $text = intval( $iso );
 385                 }
 386                 return $text;
 387         }
 388 }