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 and formats them according to a specified preference.
  28  *
  29  * This class was originally introduced to detect and transform dates in free text. It is now
  30  * only used by the {{#dateformat}} parser function. This is a very rudimentary date formatter;
  31  * Language::sprintfDate() has many more features and is the correct choice for most new code.
  32  * The main advantage of this date formatter is that it is able to format incomplete dates with an
  33  * unspecified year.
  34  *
  35  * @ingroup Parser
  36  */
  37 class DateFormatter {
  38         /** @var string[] Date format regexes indexed the class constants */
  39         private $regexes;
  40
  41         /**
  42          * @var int[][] Array of special rules. The first key is the preference ID
  43          * (one of the class constants), the second key is the detected source
  44          * format, and the value is the ID of the target format that will be used
  45          * in that case.
  46          */
  47         private $rules = [];
  48
  49         /**
  50          * @var int[] Month numbers by lowercase name
  51          */
  52         private $xMonths = [];
  53
  54         /**
  55          * @var string[] Month names by number
  56          */
  57         private $monthNames = [];
  58
  59         /**
  60          * @var int[] A map of descriptive preference text to internal format ID
  61          */
  62         private $preferenceIDs;
  63
  64         /** @var string[] Format strings similar to those used by date(), indexed by ID */
  65         private $targetFormats;
  66
  67         /** Used as a preference ID for rules that apply regardless of preference */
  68         const ALL = -1;
  69
  70         /** No preference: the date may be left in the same format as the input */
  71         const NONE = 0;
  72
  73         /** e.g. January 15, 2001 */
  74         const MDY = 1;
  75
  76         /** e.g. 15 January 2001 */
  77         const DMY = 2;
  78
  79         /** e.g. 2001 January 15 */
  80         const YMD = 3;
  81
  82         /** e.g. 2001-01-15 */
  83         const ISO = 4;
  84
  85         /** The highest ID that is a valid user preference */
  86         const LASTPREF = 4;
  87
  88         /** e.g. 2001, 15 January */
  89         const YDM = 5;
  90
  91         /** e.g. 15 January */
  92         const DM = 6;
  93
  94         /** e.g. January 15 */
  95         const MD = 7;
  96
  97         /** The highest ID that is a valid target format */
  98         const LAST = 7;
  99
 100         /**
 101          * @param Language $lang In which language to format the date
 102          */
 103         public function __construct( Language $lang ) {
 104                 $monthRegexParts = [];
 105                 for ( $i = 1; $i <= 12; $i++ ) {
 106                         $monthName = $lang->getMonthName( $i );
 107                         $monthAbbrev = $lang->getMonthAbbreviation( $i );
 108                         $this->monthNames[$i] = $monthName;
 109                         $monthRegexParts[] = preg_quote( $monthName, '/' );
 110                         $monthRegexParts[] = preg_quote( $monthAbbrev, '/' );
 111                         $this->xMonths[mb_strtolower( $monthName )] = $i;
 112                         $this->xMonths[mb_strtolower( $monthAbbrev )] = $i;
 113                 }
 114
 115                 // Partial regular expressions
 116                 $monthNames = implode( '|', $monthRegexParts );
 117                 $dm = "(?<day>\d{1,2})[ _](?<monthName>{$monthNames})";
 118                 $md = "(?<monthName>{$monthNames})[ _](?<day>\d{1,2})";
 119                 $y = '(?<year>\d{1,4}([ _]BC|))';
 120                 $iso = '(?<isoYear>-?\d{4})-(?<isoMonth>\d{2})-(?<isoDay>\d{2})';
 121
 122                 $this->regexes = [
 123                         self::DMY => "/^{$dm}(?: *, *| +){$y}$/iu",
 124                         self::YDM => "/^{$y}(?: *, *| +){$dm}$/iu",
 125                         self::MDY => "/^{$md}(?: *, *| +){$y}$/iu",
 126                         self::YMD => "/^{$y}(?: *, *| +){$md}$/iu",
 127                         self::DM => "/^{$dm}$/iu",
 128                         self::MD => "/^{$md}$/iu",
 129                         self::ISO => "/^{$iso}$/iu",
 130                 ];
 131
 132                 // Target date formats
 133                 $this->targetFormats = [
 134                         self::DMY => 'j F Y',
 135                         self::YDM => 'Y, j F',
 136                         self::MDY => 'F j, Y',
 137                         self::YMD => 'Y F j',
 138                         self::DM => 'j F',
 139                         self::MD => 'F j',
 140                         self::ISO => 'y-m-d',
 141                 ];
 142
 143                 // Rules
 144                 //           pref       source      target
 145                 $this->rules[self::DMY][self::MD] = self::DM;
 146                 $this->rules[self::ALL][self::MD] = self::MD;
 147                 $this->rules[self::MDY][self::DM] = self::MD;
 148                 $this->rules[self::ALL][self::DM] = self::DM;
 149                 $this->rules[self::NONE][self::ISO] = self::ISO;
 150
 151                 $this->preferenceIDs = [
 152                         'default' => self::NONE,
 153                         'dmy' => self::DMY,
 154                         'mdy' => self::MDY,
 155                         'ymd' => self::YMD,
 156                         'ISO 8601' => self::ISO,
 157                 ];
 158         }
 159
 160         /**
 161          * Get a DateFormatter object
 162          *
 163          * @deprecated since 1.33 use MediaWikiServices::getDateFormatterFactory()
 164          *
 165          * @param Language|null $lang In which language to format the date
 166          *     Defaults to the site content language
 167          * @return DateFormatter
 168          */
 169         public static function getInstance( Language $lang = null ) {
 170                 $lang = $lang ?? MediaWikiServices::getInstance()->getContentLanguage();
 171                 return MediaWikiServices::getInstance()->getDateFormatterFactory()->get( $lang );
 172         }
 173
 174         /**
 175          * @param string $preference User preference, must be one of "default",
 176          *   "dmy", "mdy", "ymd" or "ISO 8601".
 177          * @param string $text Text to reformat
 178          * @param array $options Ignored. Since 1.33, 'match-whole' is implied, and
 179          *  'linked' has been removed.
 180          *
 181          * @return string
 182          */
 183         public function reformat( $preference, $text, $options = [] ) {
 184                 if ( isset( $this->preferenceIDs[$preference] ) ) {
 185                         $preference = $this->preferenceIDs[$preference];
 186                 } else {
 187                         $preference = self::NONE;
 188                 }
 189                 for ( $source = 1; $source <= self::LAST; $source++ ) {
 190                         if ( isset( $this->rules[$preference][$source] ) ) {
 191                                 # Specific rules
 192                                 $target = $this->rules[$preference][$source];
 193                         } elseif ( isset( $this->rules[self::ALL][$source] ) ) {
 194                                 # General rules
 195                                 $target = $this->rules[self::ALL][$source];
 196                         } elseif ( $preference ) {
 197                                 # User preference
 198                                 $target = $preference;
 199                         } else {
 200                                 # Default
 201                                 $target = $source;
 202                         }
 203                         $regex = $this->regexes[$source];
 204
 205                         $text = preg_replace_callback( $regex,
 206                                 function ( $match ) use ( $target ) {
 207                                         $format = $this->targetFormats[$target];
 208
 209                                         $text = '';
 210
 211                                         // Pre-generate y/Y stuff because we need the year for the <span> title.
 212                                         if ( !isset( $match['isoYear'] ) && isset( $match['year'] ) ) {
 213                                                 $match['isoYear'] = $this->makeIsoYear( $match['year'] );
 214                                         }
 215                                         if ( !isset( $match['year'] ) && isset( $match['isoYear'] ) ) {
 216                                                 $match['year'] = $this->makeNormalYear( $match['isoYear'] );
 217                                         }
 218
 219                                         if ( !isset( $match['isoMonth'] ) ) {
 220                                                 $m = $this->makeIsoMonth( $match['monthName'] );
 221                                                 if ( $m === false ) {
 222                                                         // Fail
 223                                                         return $match[0];
 224                                                 } else {
 225                                                         $match['isoMonth'] = $m;
 226                                                 }
 227                                         }
 228
 229                                         if ( !isset( $match['isoDay'] ) ) {
 230                                                 $match['isoDay'] = sprintf( '%02d', $match['day'] );
 231                                         }
 232
 233                                         $formatLength = strlen( $format );
 234                                         for ( $p = 0; $p < $formatLength; $p++ ) {
 235                                                 $char = $format[$p];
 236                                                 switch ( $char ) {
 237                                                         case 'd': // ISO day of month
 238                                                                 $text .= $match['isoDay'];
 239                                                                 break;
 240                                                         case 'm': // ISO month
 241                                                                 $text .= $match['isoMonth'];
 242                                                                 break;
 243                                                         case 'y': // ISO year
 244                                                                 $text .= $match['isoYear'];
 245                                                                 break;
 246                                                         case 'j': // ordinary day of month
 247                                                                 if ( !isset( $match['day'] ) ) {
 248                                                                         $text .= intval( $match['isoDay'] );
 249                                                                 } else {
 250                                                                         $text .= $match['day'];
 251                                                                 }
 252                                                                 break;
 253                                                         case 'F': // long month
 254                                                                 $m = intval( $match['isoMonth'] );
 255                                                                 if ( $m > 12 || $m < 1 ) {
 256                                                                         // Fail
 257                                                                         return $match[0];
 258                                                                 } else {
 259                                                                         $text .= $this->monthNames[$m];
 260                                                                 }
 261                                                                 break;
 262                                                         case 'Y': // ordinary (optional BC) year
 263                                                                 $text .= $match['year'];
 264                                                                 break;
 265                                                         default:
 266                                                                 $text .= $char;
 267                                                 }
 268                                         }
 269
 270                                         $isoBits = [];
 271                                         if ( isset( $match['isoYear'] ) ) {
 272                                                 $isoBits[] = $match['isoYear'];
 273                                         }
 274                                         $isoBits[] = $match['isoMonth'];
 275                                         $isoBits[] = $match['isoDay'];
 276                                         $isoDate = implode( '-', $isoBits );
 277
 278                                         // Output is not strictly HTML (it's wikitext), but <span> is whitelisted.
 279                                         $text = Html::rawElement( 'span',
 280                                                 [ 'class' => 'mw-formatted-date', 'title' => $isoDate ], $text );
 281
 282                                         return $text;
 283                                 }, $text
 284                         );
 285                 }
 286                 return $text;
 287         }
 288
 289         /**
 290          * Makes an ISO month, e.g. 02, from a month name
 291          * @param string $monthName Month name
 292          * @return string|false ISO month name, or false if the input was invalid
 293          */
 294         private function makeIsoMonth( $monthName ) {
 295                 $isoMonth = $this->xMonths[mb_strtolower( $monthName )] ?? false;
 296                 if ( $isoMonth === false ) {
 297                         return false;
 298                 }
 299                 return sprintf( '%02d', $isoMonth );
 300         }
 301
 302         /**
 303          * Make an ISO year from a year name, for instance: '-1199' from '1200 BC'
 304          * @param string $year Year name
 305          * @return string ISO year name
 306          */
 307         private function makeIsoYear( $year ) {
 308                 // Assumes the year is in a nice format, as enforced by the regex
 309                 if ( substr( $year, -2 ) == 'BC' ) {
 310                         $num = intval( substr( $year, 0, -3 ) ) - 1;
 311                         // PHP bug note: sprintf( "%04d", -1 ) fails poorly
 312                         $text = sprintf( '-%04d', $num );
 313                 } else {
 314                         $text = sprintf( '%04d', $year );
 315                 }
 316                 return $text;
 317         }
 318
 319         /**
 320          * Make a year from an ISO year, for instance: '400 BC' from '-0399'.
 321          * @param string $iso ISO year
 322          * @return int|string int representing year number in case of AD dates, or string containing
 323          *   year number and 'BC' at the end otherwise.
 324          */
 325         private function makeNormalYear( $iso ) {
 326                 if ( $iso[0] == '-' ) {
 327                         $text = ( intval( substr( $iso, 1 ) ) + 1 ) . ' BC';
 328                 } else {
 329                         $text = intval( $iso );
 330                 }
 331                 return $text;
 332         }
 333 }