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