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\Html\Html;
  25 use MediaWiki\MediaWikiServices;
  26
  27 /**
  28  * Date formatter. Recognises dates and formats them according to a specified preference.
  29  *
  30  * This class was originally introduced to detect and transform dates in free text. It is now
  31  * only used by the {{#dateformat}} parser function. This is a very rudimentary date formatter;
  32  * Language::sprintfDate() has many more features and is the correct choice for most new code.
  33  * The main advantage of this date formatter is that it is able to format incomplete dates with an
  34  * unspecified year.
  35  *
  36  * @ingroup Parser
  37  */
  38 class DateFormatter {
  39         /** @var string[] Date format regexes indexed the class constants */
  40         private $regexes;
  41
  42         /**
  43          * @var int[][] Array of special rules. The first key is the preference ID
  44          * (one of the class constants), the second key is the detected source
  45          * format, and the value is the ID of the target format that will be used
  46          * in that case.
  47          */
  48         private const RULES = [
  49                 self::ALL => [
  50                         self::MD => self::MD,
  51                         self::DM => self::DM,
  52                 ],
  53                 self::NONE => [
  54                         self::ISO => self::ISO,
  55                 ],
  56                 self::MDY => [
  57                         self::DM => self::MD,
  58                 ],
  59                 self::DMY => [
  60                         self::MD => self::DM,
  61                 ],
  62         ];
  63
  64         /**
  65          * @var array<string,int> Month numbers by lowercase name
  66          */
  67         private $xMonths = [];
  68
  69         /**
  70          * @var array<int,string> Month names by number
  71          */
  72         private $monthNames = [];
  73
  74         /**
  75          * @var int[] A map of descriptive preference text to internal format ID
  76          */
  77         private const PREFERENCE_IDS = [
  78                 'default' => self::NONE,
  79                 'dmy' => self::DMY,
  80                 'mdy' => self::MDY,
  81                 'ymd' => self::YMD,
  82                 'ISO 8601' => self::ISO,
  83         ];
  84
  85         /** @var string[] Format strings similar to those used by date(), indexed by ID */
  86         private const TARGET_FORMATS = [
  87                 self::MDY => 'F j, Y',
  88                 self::DMY => 'j F Y',
  89                 self::YMD => 'Y F j',
  90                 self::ISO => 'y-m-d',
  91                 self::YDM => 'Y, j F',
  92                 self::DM => 'j F',
  93                 self::MD => 'F j',
  94         ];
  95
  96         /** Used as a preference ID for rules that apply regardless of preference */
  97         private const ALL = -1;
  98
  99         /** No preference: the date may be left in the same format as the input */
 100         private const NONE = 0;
 101
 102         /** e.g. January 15, 2001 */
 103         private const MDY = 1;
 104
 105         /** e.g. 15 January 2001 */
 106         private const DMY = 2;
 107
 108         /** e.g. 2001 January 15 */
 109         private const YMD = 3;
 110
 111         /** e.g. 2001-01-15 */
 112         private const ISO = 4;
 113
 114         /** e.g. 2001, 15 January */
 115         private const YDM = 5;
 116
 117         /** e.g. 15 January */
 118         private const DM = 6;
 119
 120         /** e.g. January 15 */
 121         private const MD = 7;
 122
 123         /**
 124          * @param Language $lang In which language to format the date
 125          */
 126         public function __construct( Language $lang ) {
 127                 $monthRegexParts = [];
 128                 for ( $i = 1; $i <= 12; $i++ ) {
 129                         $monthName = $lang->getMonthName( $i );
 130                         $monthAbbrev = $lang->getMonthAbbreviation( $i );
 131                         $this->monthNames[$i] = $monthName;
 132                         $monthRegexParts[] = preg_quote( $monthName, '/' );
 133                         $monthRegexParts[] = preg_quote( $monthAbbrev, '/' );
 134                         $this->xMonths[mb_strtolower( $monthName )] = $i;
 135                         $this->xMonths[mb_strtolower( $monthAbbrev )] = $i;
 136                 }
 137
 138                 // Partial regular expressions
 139                 $monthNames = implode( '|', $monthRegexParts );
 140                 $dm = "(?<day>\d{1,2})[ _](?<monthName>{$monthNames})";
 141                 $md = "(?<monthName>{$monthNames})[ _](?<day>\d{1,2})";
 142                 $y = '(?<year>\d{1,4}([ _]BC|))';
 143                 $iso = '(?<isoYear>-?\d{4})-(?<isoMonth>\d{2})-(?<isoDay>\d{2})';
 144
 145                 $this->regexes = [
 146                         self::DMY => "/^{$dm}(?: *, *| +){$y}$/iu",
 147                         self::YDM => "/^{$y}(?: *, *| +){$dm}$/iu",
 148                         self::MDY => "/^{$md}(?: *, *| +){$y}$/iu",
 149                         self::YMD => "/^{$y}(?: *, *| +){$md}$/iu",
 150                         self::DM => "/^{$dm}$/iu",
 151                         self::MD => "/^{$md}$/iu",
 152                         self::ISO => "/^{$iso}$/iu",
 153                 ];
 154         }
 155
 156         /**
 157          * Get a DateFormatter object
 158          *
 159          * @deprecated since 1.33 use MediaWikiServices::getDateFormatterFactory()
 160          *
 161          * @param Language|null $lang In which language to format the date
 162          *     Defaults to the site content language
 163          * @return DateFormatter
 164          */
 165         public static function getInstance( Language $lang = null ) {
 166                 $lang ??= MediaWikiServices::getInstance()->getContentLanguage();
 167                 return MediaWikiServices::getInstance()->getDateFormatterFactory()->get( $lang );
 168         }
 169
 170         /**
 171          * @param string $preference User preference, must be one of "default",
 172          *   "dmy", "mdy", "ymd" or "ISO 8601".
 173          * @param string $text Text to reformat
 174          * @param array $options Ignored. Since 1.33, 'match-whole' is implied, and
 175          *  'linked' has been removed.
 176          *
 177          * @return string
 178          */
 179         public function reformat( $preference, $text, $options = [] ) {
 180                 $userFormatId = self::PREFERENCE_IDS[$preference] ?? self::NONE;
 181                 foreach ( self::TARGET_FORMATS as $source => $_ ) {
 182                         if ( isset( self::RULES[$userFormatId][$source] ) ) {
 183                                 # Specific rules
 184                                 $target = self::RULES[$userFormatId][$source];
 185                         } elseif ( isset( self::RULES[self::ALL][$source] ) ) {
 186                                 # General rules
 187                                 $target = self::RULES[self::ALL][$source];
 188                         } elseif ( $userFormatId ) {
 189                                 # User preference
 190                                 $target = $userFormatId;
 191                         } else {
 192                                 # Default
 193                                 $target = $source;
 194                         }
 195                         $format = self::TARGET_FORMATS[$target];
 196                         $regex = $this->regexes[$source];
 197
 198                         $text = preg_replace_callback( $regex,
 199                                 function ( $match ) use ( $format ) {
 200                                         $text = '';
 201
 202                                         // Pre-generate y/Y stuff because we need the year for the <span> title.
 203                                         if ( !isset( $match['isoYear'] ) && isset( $match['year'] ) ) {
 204                                                 $match['isoYear'] = $this->makeIsoYear( $match['year'] );
 205                                         }
 206                                         if ( !isset( $match['year'] ) && isset( $match['isoYear'] ) ) {
 207                                                 $match['year'] = $this->makeNormalYear( $match['isoYear'] );
 208                                         }
 209
 210                                         if ( !isset( $match['isoMonth'] ) ) {
 211                                                 $m = $this->makeIsoMonth( $match['monthName'] );
 212                                                 if ( $m === null ) {
 213                                                         // Fail
 214                                                         return $match[0];
 215                                                 }
 216                                                 $match['isoMonth'] = $m;
 217                                         }
 218
 219                                         if ( !isset( $match['isoDay'] ) ) {
 220                                                 $match['isoDay'] = sprintf( '%02d', $match['day'] );
 221                                         }
 222
 223                                         $formatLength = strlen( $format );
 224                                         for ( $p = 0; $p < $formatLength; $p++ ) {
 225                                                 $char = $format[$p];
 226                                                 switch ( $char ) {
 227                                                         case 'd': // ISO day of month
 228                                                                 $text .= $match['isoDay'];
 229                                                                 break;
 230                                                         case 'm': // ISO month
 231                                                                 $text .= $match['isoMonth'];
 232                                                                 break;
 233                                                         case 'y': // ISO year
 234                                                                 // @phan-suppress-next-line PhanTypePossiblyInvalidDimOffset False positive
 235                                                                 $text .= $match['isoYear'];
 236                                                                 break;
 237                                                         case 'j': // ordinary day of month
 238                                                                 if ( !isset( $match['day'] ) ) {
 239                                                                         $text .= intval( $match['isoDay'] );
 240                                                                 } else {
 241                                                                         $text .= $match['day'];
 242                                                                 }
 243                                                                 break;
 244                                                         case 'F': // long month
 245                                                                 $m = intval( $match['isoMonth'] );
 246                                                                 if ( $m > 12 || $m < 1 ) {
 247                                                                         // Fail
 248                                                                         return $match[0];
 249                                                                 }
 250                                                                 $text .= $this->monthNames[$m];
 251                                                                 break;
 252                                                         case 'Y': // ordinary (optional BC) year
 253                                                                 // @phan-suppress-next-line PhanTypePossiblyInvalidDimOffset False positive
 254                                                                 $text .= $match['year'];
 255                                                                 break;
 256                                                         default:
 257                                                                 $text .= $char;
 258                                                 }
 259                                         }
 260
 261                                         $isoBits = [];
 262                                         if ( isset( $match['isoYear'] ) ) {
 263                                                 $isoBits[] = $match['isoYear'];
 264                                         }
 265                                         $isoBits[] = $match['isoMonth'];
 266                                         $isoBits[] = $match['isoDay'];
 267                                         $isoDate = implode( '-', $isoBits );
 268
 269                                         // Output is not strictly HTML (it's wikitext), but <span> is allowed.
 270                                         return Html::rawElement( 'span',
 271                                                 [ 'class' => 'mw-formatted-date', 'title' => $isoDate ], $text );
 272                                 }, $text
 273                         );
 274                 }
 275                 return $text;
 276         }
 277
 278         /**
 279          * @param string $monthName
 280          * @return string|null 2-digit month number, e.g. "02", or null if the input was invalid
 281          */
 282         private function makeIsoMonth( $monthName ) {
 283                 $number = $this->xMonths[mb_strtolower( $monthName )] ?? null;
 284                 return $number !== null ? sprintf( '%02d', $number ) : null;
 285         }
 286
 287         /**
 288          * Make an ISO year from a year name, for instance: '-1199' from '1200 BC'
 289          * @param string $year Year name
 290          * @return string ISO year name
 291          */
 292         private function makeIsoYear( $year ) {
 293                 // Assumes the year is in a nice format, as enforced by the regex
 294                 if ( substr( $year, -2 ) == 'BC' ) {
 295                         $num = intval( substr( $year, 0, -3 ) ) - 1;
 296                         // PHP bug note: sprintf( "%04d", -1 ) fails poorly
 297                         $text = sprintf( '-%04d', $num );
 298                 } else {
 299                         $text = sprintf( '%04d', $year );
 300                 }
 301                 return $text;
 302         }
 303
 304         /**
 305          * Make a year from an ISO year, for instance: '400 BC' from '-0399'.
 306          * @param string $iso ISO year
 307          * @return int|string int representing year number in case of AD dates, or string containing
 308          *   year number and 'BC' at the end otherwise.
 309          */
 310         private function makeNormalYear( $iso ) {
 311                 if ( $iso <= 0 ) {
 312                         $text = ( intval( substr( $iso, 1 ) ) + 1 ) . ' BC';
 313                 } else {
 314                         $text = intval( $iso );
 315                 }
 316                 return $text;
 317         }
 318 }