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