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 namespace MediaWiki\Parser;
  25
  26 use MediaWiki\Html\Html;
  27 use MediaWiki\Language\Language;
  28 use MediaWiki\MediaWikiServices;
  29
  30 /**
  31  * Date formatter. Recognises dates and formats them according to a specified preference.
  32  *
  33  * This class was originally introduced to detect and transform dates in free text. It is now
  34  * only used by the {{#dateformat}} parser function. This is a very rudimentary date formatter;
  35  * Language::sprintfDate() has many more features and is the correct choice for most new code.
  36  * The main advantage of this date formatter is that it is able to format incomplete dates with an
  37  * unspecified year.
  38  *
  39  * @ingroup Parser
  40  */
  41 class DateFormatter {
  42         /** @var string[] Date format regexes indexed the class constants */
  43         private $regexes;
  44
  45         /**
  46          * @var int[][] Array of special rules. The first key is the preference ID
  47          * (one of the class constants), the second key is the detected source
  48          * format, and the value is the ID of the target format that will be used
  49          * in that case.
  50          */
  51         private const RULES = [
  52                 self::ALL => [
  53                         self::MD => self::MD,
  54                         self::DM => self::DM,
  55                 ],
  56                 self::NONE => [
  57                         self::ISO => self::ISO,
  58                 ],
  59                 self::MDY => [
  60                         self::DM => self::MD,
  61                 ],
  62                 self::DMY => [
  63                         self::MD => self::DM,
  64                 ],
  65         ];
  66
  67         /**
  68          * @var array<string,int> Month numbers by lowercase name
  69          */
  70         private $xMonths = [];
  71
  72         /**
  73          * @var array<int,string> Month names by number
  74          */
  75         private $monthNames = [];
  76
  77         /**
  78          * @var int[] A map of descriptive preference text to internal format ID
  79          */
  80         private const PREFERENCE_IDS = [
  81                 'default' => self::NONE,
  82                 'dmy' => self::DMY,
  83                 'mdy' => self::MDY,
  84                 'ymd' => self::YMD,
  85                 'ISO 8601' => self::ISO,
  86         ];
  87
  88         /** @var string[] Format strings similar to those used by date(), indexed by ID */
  89         private const TARGET_FORMATS = [
  90                 self::MDY => 'F j, Y',
  91                 self::DMY => 'j F Y',
  92                 self::YMD => 'Y F j',
  93                 self::ISO => 'y-m-d',
  94                 self::YDM => 'Y, j F',
  95                 self::DM => 'j F',
  96                 self::MD => 'F j',
  97         ];
  98
  99         /** Used as a preference ID for rules that apply regardless of preference */
 100         private const ALL = -1;
 101
 102         /** No preference: the date may be left in the same format as the input */
 103         private const NONE = 0;
 104
 105         /** e.g. January 15, 2001 */
 106         private const MDY = 1;
 107
 108         /** e.g. 15 January 2001 */
 109         private const DMY = 2;
 110
 111         /** e.g. 2001 January 15 */
 112         private const YMD = 3;
 113
 114         /** e.g. 2001-01-15 */
 115         private const ISO = 4;
 116
 117         /** e.g. 2001, 15 January */
 118         private const YDM = 5;
 119
 120         /** e.g. 15 January */
 121         private const DM = 6;
 122
 123         /** e.g. January 15 */
 124         private const MD = 7;
 125
 126         /**
 127          * @param Language $lang In which language to format the date
 128          */
 129         public function __construct( Language $lang ) {
 130                 $monthRegexParts = [];
 131                 for ( $i = 1; $i <= 12; $i++ ) {
 132                         $monthName = $lang->getMonthName( $i );
 133                         $monthAbbrev = $lang->getMonthAbbreviation( $i );
 134                         $this->monthNames[$i] = $monthName;
 135                         $monthRegexParts[] = preg_quote( $monthName, '/' );
 136                         $monthRegexParts[] = preg_quote( $monthAbbrev, '/' );
 137                         $this->xMonths[mb_strtolower( $monthName )] = $i;
 138                         $this->xMonths[mb_strtolower( $monthAbbrev )] = $i;
 139                 }
 140
 141                 // Partial regular expressions
 142                 $monthNames = implode( '|', $monthRegexParts );
 143                 $dm = "(?<day>\d{1,2})[ _](?<monthName>{$monthNames})";
 144                 $md = "(?<monthName>{$monthNames})[ _](?<day>\d{1,2})";
 145                 $y = '(?<year>\d{1,4}([ _]BC|))';
 146                 $iso = '(?<isoYear>-?\d{4})-(?<isoMonth>\d{2})-(?<isoDay>\d{2})';
 147
 148                 $this->regexes = [
 149                         self::DMY => "/^{$dm}(?: *, *| +){$y}$/iu",
 150                         self::YDM => "/^{$y}(?: *, *| +){$dm}$/iu",
 151                         self::MDY => "/^{$md}(?: *, *| +){$y}$/iu",
 152                         self::YMD => "/^{$y}(?: *, *| +){$md}$/iu",
 153                         self::DM => "/^{$dm}$/iu",
 154                         self::MD => "/^{$md}$/iu",
 155                         self::ISO => "/^{$iso}$/iu",
 156                 ];
 157         }
 158
 159         /**
 160          * Get a DateFormatter object
 161          *
 162          * @deprecated since 1.33 use MediaWikiServices::getDateFormatterFactory()
 163          *
 164          * @param Language|null $lang In which language to format the date
 165          *     Defaults to the site content language
 166          * @return DateFormatter
 167          */
 168         public static function getInstance( ?Language $lang = null ) {
 169                 $lang ??= MediaWikiServices::getInstance()->getContentLanguage();
 170                 return MediaWikiServices::getInstance()->getDateFormatterFactory()->get( $lang );
 171         }
 172
 173         /**
 174          * @param string $preference User preference, must be one of "default",
 175          *   "dmy", "mdy", "ymd" or "ISO 8601".
 176          * @param string $text Text to reformat
 177          * @param array $options Ignored. Since 1.33, 'match-whole' is implied, and
 178          *  'linked' has been removed.
 179          *
 180          * @return string
 181          */
 182         public function reformat( $preference, $text, $options = [] ) {
 183                 $userFormatId = self::PREFERENCE_IDS[$preference] ?? self::NONE;
 184                 foreach ( self::TARGET_FORMATS as $source => $_ ) {
 185                         if ( isset( self::RULES[$userFormatId][$source] ) ) {
 186                                 # Specific rules
 187                                 $target = self::RULES[$userFormatId][$source];
 188                         } elseif ( isset( self::RULES[self::ALL][$source] ) ) {
 189                                 # General rules
 190                                 $target = self::RULES[self::ALL][$source];
 191                         } elseif ( $userFormatId ) {
 192                                 # User preference
 193                                 $target = $userFormatId;
 194                         } else {
 195                                 # Default
 196                                 $target = $source;
 197                         }
 198                         $format = self::TARGET_FORMATS[$target];
 199                         $regex = $this->regexes[$source];
 200
 201                         $text = preg_replace_callback( $regex,
 202                                 function ( $match ) use ( $format ) {
 203                                         $text = '';
 204
 205                                         // Pre-generate y/Y stuff because we need the year for the <span> title.
 206                                         if ( !isset( $match['isoYear'] ) && isset( $match['year'] ) ) {
 207                                                 $match['isoYear'] = $this->makeIsoYear( $match['year'] );
 208                                         }
 209                                         if ( !isset( $match['year'] ) && isset( $match['isoYear'] ) ) {
 210                                                 $match['year'] = $this->makeNormalYear( $match['isoYear'] );
 211                                         }
 212
 213                                         if ( !isset( $match['isoMonth'] ) ) {
 214                                                 $m = $this->makeIsoMonth( $match['monthName'] );
 215                                                 if ( $m === null ) {
 216                                                         // Fail
 217                                                         return $match[0];
 218                                                 }
 219                                                 $match['isoMonth'] = $m;
 220                                         }
 221
 222                                         if ( !isset( $match['isoDay'] ) ) {
 223                                                 $match['isoDay'] = sprintf( '%02d', $match['day'] );
 224                                         }
 225
 226                                         $formatLength = strlen( $format );
 227                                         for ( $p = 0; $p < $formatLength; $p++ ) {
 228                                                 $char = $format[$p];
 229                                                 switch ( $char ) {
 230                                                         case 'd': // ISO day of month
 231                                                                 $text .= $match['isoDay'];
 232                                                                 break;
 233                                                         case 'm': // ISO month
 234                                                                 $text .= $match['isoMonth'];
 235                                                                 break;
 236                                                         case 'y': // ISO year
 237                                                                 // @phan-suppress-next-line PhanTypePossiblyInvalidDimOffset False positive
 238                                                                 $text .= $match['isoYear'];
 239                                                                 break;
 240                                                         case 'j': // ordinary day of month
 241                                                                 if ( !isset( $match['day'] ) ) {
 242                                                                         $text .= intval( $match['isoDay'] );
 243                                                                 } else {
 244                                                                         $text .= $match['day'];
 245                                                                 }
 246                                                                 break;
 247                                                         case 'F': // long month
 248                                                                 $m = intval( $match['isoMonth'] );
 249                                                                 if ( $m > 12 || $m < 1 ) {
 250                                                                         // Fail
 251                                                                         return $match[0];
 252                                                                 }
 253                                                                 $text .= $this->monthNames[$m];
 254                                                                 break;
 255                                                         case 'Y': // ordinary (optional BC) year
 256                                                                 // @phan-suppress-next-line PhanTypePossiblyInvalidDimOffset False positive
 257                                                                 $text .= $match['year'];
 258                                                                 break;
 259                                                         default:
 260                                                                 $text .= $char;
 261                                                 }
 262                                         }
 263
 264                                         $isoBits = [];
 265                                         if ( isset( $match['isoYear'] ) ) {
 266                                                 $isoBits[] = $match['isoYear'];
 267                                         }
 268                                         $isoBits[] = $match['isoMonth'];
 269                                         $isoBits[] = $match['isoDay'];
 270                                         $isoDate = implode( '-', $isoBits );
 271
 272                                         // Output is not strictly HTML (it's wikitext), but <span> is allowed.
 273                                         return Html::rawElement( 'span',
 274                                                 [ 'class' => 'mw-formatted-date', 'title' => $isoDate ], $text );
 275                                 }, $text
 276                         );
 277                 }
 278                 return $text;
 279         }
 280
 281         /**
 282          * @param string $monthName
 283          * @return string|null 2-digit month number, e.g. "02", or null if the input was invalid
 284          */
 285         private function makeIsoMonth( $monthName ) {
 286                 $number = $this->xMonths[mb_strtolower( $monthName )] ?? null;
 287                 return $number !== null ? sprintf( '%02d', $number ) : null;
 288         }
 289
 290         /**
 291          * Make an ISO year from a year name, for instance: '-1199' from '1200 BC'
 292          * @param string $year Year name
 293          * @return string ISO year name
 294          */
 295         private function makeIsoYear( $year ) {
 296                 // Assumes the year is in a nice format, as enforced by the regex
 297                 if ( substr( $year, -2 ) == 'BC' ) {
 298                         $num = intval( substr( $year, 0, -3 ) ) - 1;
 299                         // PHP bug note: sprintf( "%04d", -1 ) fails poorly
 300                         $text = sprintf( '-%04d', $num );
 301                 } else {
 302                         $text = sprintf( '%04d', $year );
 303                 }
 304                 return $text;
 305         }
 306
 307         /**
 308          * Make a year from an ISO year, for instance: '400 BC' from '-0399'.
 309          * @param string $iso ISO year
 310          * @return int|string int representing year number in case of AD dates, or string containing
 311          *   year number and 'BC' at the end otherwise.
 312          */
 313         private function makeNormalYear( $iso ) {
 314                 if ( $iso <= 0 ) {
 315                         $text = ( intval( substr( $iso, 1 ) ) + 1 ) . ' BC';
 316                 } else {
 317                         $text = intval( $iso );
 318                 }
 319                 return $text;
 320         }
 321 }
 322
 323 /** @deprecated class alias since 1.43 */
 324 class_alias( DateFormatter::class, 'DateFormatter' );