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         var $mSource, $mTarget;
  31         var $monthNames = '', $rxDM, $rxMD, $rxDMY, $rxYDM, $rxMDY, $rxYMD;
  32
  33         var $regexes, $pDays, $pMonths, $pYears;
  34         var $rules, $xMonths, $preferences;
  35
  36         protected $lang;
  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 $lang Language In which language to format the date
  53          */
  54         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 = array(
 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 $lang Language|string|null In which language to format the date
 124          *              Defaults to the site content language
 125          * @return DateFormatter object
 126          */
 127         public static function &getInstance( $lang = null ) {
 128                 global $wgMemc, $wgContLang;
 129                 static $dateFormatter = false;
 130                 $lang = $lang ? wfGetLangObj( $lang ) : $wgContLang;
 131                 $key = wfMemcKey( 'dateformatter', $lang->getCode() );
 132                 if ( !$dateFormatter ) {
 133                         $dateFormatter = $wgMemc->get( $key );
 134                         if ( !$dateFormatter ) {
 135                                 $dateFormatter = new DateFormatter( $lang );
 136                                 $wgMemc->set( $key, $dateFormatter, 3600 );
 137                         }
 138                 }
 139                 return $dateFormatter;
 140         }
 141
 142         /**
 143          * @param string $preference User preference
 144          * @param string $text Text to reformat
 145          * @param array $options can contain 'linked' and/or 'match-whole'
 146          *
 147          * @return string
 148          */
 149         function reformat( $preference, $text, $options = array( 'linked' ) ) {
 150                 $linked = in_array( 'linked', $options );
 151                 $match_whole = in_array( 'match-whole', $options );
 152
 153                 if ( isset( $this->preferences[$preference] ) ) {
 154                         $preference = $this->preferences[$preference];
 155                 } else {
 156                         $preference = self::NONE;
 157                 }
 158                 for ( $i = 1; $i <= self::LAST; $i++ ) {
 159                         $this->mSource = $i;
 160                         if ( isset( $this->rules[$preference][$i] ) ) {
 161                                 # Specific rules
 162                                 $this->mTarget = $this->rules[$preference][$i];
 163                         } elseif ( isset( $this->rules[self::ALL][$i] ) ) {
 164                                 # General rules
 165                                 $this->mTarget = $this->rules[self::ALL][$i];
 166                         } elseif ( $preference ) {
 167                                 # User preference
 168                                 $this->mTarget = $preference;
 169                         } else {
 170                                 # Default
 171                                 $this->mTarget = $i;
 172                         }
 173                         $regex = $this->regexes[$i];
 174
 175                         // Horrible hack
 176                         if ( !$linked ) {
 177                                 $regex = str_replace( array( '\[\[', '\]\]' ), '', $regex );
 178                         }
 179
 180                         if ( $match_whole ) {
 181                                 // Let's hope this works
 182                                 $regex = preg_replace( '!^/!', '/^', $regex );
 183                                 $regex = str_replace( $this->regexTrail,
 184                                         '$' . $this->regexTrail, $regex );
 185                         }
 186
 187                         // Another horrible hack
 188                         $this->mLinked = $linked;
 189                         $text = preg_replace_callback( $regex, array( &$this, 'replace' ), $text );
 190                         unset( $this->mLinked );
 191                 }
 192                 return $text;
 193         }
 194
 195         /**
 196          * @param $matches
 197          * @return string
 198          */
 199         function replace( $matches ) {
 200                 # Extract information from $matches
 201                 $linked = true;
 202                 if ( isset( $this->mLinked ) ) {
 203                         $linked = $this->mLinked;
 204                 }
 205
 206                 $bits = array();
 207                 $key = $this->keys[$this->mSource];
 208                 for ( $p = 0; $p < strlen( $key ); $p++ ) {
 209                         if ( $key[$p] != ' ' ) {
 210                                 $bits[$key[$p]] = $matches[$p + 1];
 211                         }
 212                 }
 213
 214                 return $this->formatDate( $bits, $linked );
 215         }
 216
 217         /**
 218          * @param $bits array
 219          * @param $link bool
 220          * @return string
 221          */
 222         function formatDate( $bits, $link = true ) {
 223                 $format = $this->targets[$this->mTarget];
 224
 225                 if ( !$link ) {
 226                         // strip piped links
 227                         $format = preg_replace( '/\[\[[^|]+\|([^\]]+)\]\]/', '$1', $format );
 228                         // strip remaining links
 229                         $format = str_replace( array( '[[', ']]' ), '', $format );
 230                 }
 231
 232                 # Construct new date
 233                 $text = '';
 234                 $fail = false;
 235
 236                 // Pre-generate y/Y stuff because we need the year for the <span> title.
 237                 if ( !isset( $bits['y'] ) && isset( $bits['Y'] ) ) {
 238                         $bits['y'] = $this->makeIsoYear( $bits['Y'] );
 239                 }
 240                 if ( !isset( $bits['Y'] ) && isset( $bits['y'] ) ) {
 241                         $bits['Y'] = $this->makeNormalYear( $bits['y'] );
 242                 }
 243
 244                 if ( !isset( $bits['m'] ) ) {
 245                         $m = $this->makeIsoMonth( $bits['F'] );
 246                         if ( !$m || $m == '00' ) {
 247                                 $fail = true;
 248                         } else {
 249                                 $bits['m'] = $m;
 250                         }
 251                 }
 252
 253                 if ( !isset( $bits['d'] ) ) {
 254                         $bits['d'] = sprintf( '%02d', $bits['j'] );
 255                 }
 256
 257                 for ( $p = 0; $p < strlen( $format ); $p++ ) {
 258                         $char = $format[$p];
 259                         switch ( $char ) {
 260                                 case 'd': # ISO day of month
 261                                         $text .= $bits['d'];
 262                                         break;
 263                                 case 'm': # ISO month
 264                                         $text .= $bits['m'];
 265                                         break;
 266                                 case 'y': # ISO year
 267                                         $text .= $bits['y'];
 268                                         break;
 269                                 case 'j': # ordinary day of month
 270                                         if ( !isset( $bits['j'] ) ) {
 271                                                 $text .= intval( $bits['d'] );
 272                                         } else {
 273                                                 $text .= $bits['j'];
 274                                         }
 275                                         break;
 276                                 case 'F': # long month
 277                                         if ( !isset( $bits['F'] ) ) {
 278                                                 $m = intval( $bits['m'] );
 279                                                 if ( $m > 12 || $m < 1 ) {
 280                                                         $fail = true;
 281                                                 } else {
 282                                                         $text .= $this->lang->getMonthName( $m );
 283                                                 }
 284                                         } else {
 285                                                 $text .= ucfirst( $bits['F'] );
 286                                         }
 287                                         break;
 288                                 case 'Y': # ordinary (optional BC) year
 289                                         $text .= $bits['Y'];
 290                                         break;
 291                                 default:
 292                                         $text .= $char;
 293                         }
 294                 }
 295                 if ( $fail ) {
 296                         $text = $matches[0];
 297                 }
 298
 299                 $isoBits = array();
 300                 if ( isset( $bits['y'] ) ) {
 301                         $isoBits[] = $bits['y'];
 302                 }
 303                 $isoBits[] = $bits['m'];
 304                 $isoBits[] = $bits['d'];
 305                 $isoDate = implode( '-', $isoBits );
 306
 307                 // Output is not strictly HTML (it's wikitext), but <span> is whitelisted.
 308                 $text = Html::rawElement( 'span',
 309                                         array( 'class' => 'mw-formatted-date', 'title' => $isoDate ), $text );
 310
 311                 return $text;
 312         }
 313
 314         /**
 315          * @todo document
 316          * @return string
 317          */
 318         function getMonthRegex() {
 319                 $names = array();
 320                 for ( $i = 1; $i <= 12; $i++ ) {
 321                         $names[] = $this->lang->getMonthName( $i );
 322                         $names[] = $this->lang->getMonthAbbreviation( $i );
 323                 }
 324                 return implode( '|', $names );
 325         }
 326
 327         /**
 328          * Makes an ISO month, e.g. 02, from a month name
 329          * @param string $monthName month name
 330          * @return string ISO month name
 331          */
 332         function makeIsoMonth( $monthName ) {
 333                 $n = $this->xMonths[$this->lang->lc( $monthName )];
 334                 return sprintf( '%02d', $n );
 335         }
 336
 337         /**
 338          * @todo document
 339          * @param string $year Year name
 340          * @return string ISO year name
 341          */
 342         function makeIsoYear( $year ) {
 343                 # Assumes the year is in a nice format, as enforced by the regex
 344                 if ( substr( $year, -2 ) == 'BC' ) {
 345                         $num = intval( substr( $year, 0, -3 ) ) - 1;
 346                         # PHP bug note: sprintf( "%04d", -1 ) fails poorly
 347                         $text = sprintf( '-%04d', $num );
 348
 349                 } else {
 350                         $text = sprintf( '%04d', $year );
 351                 }
 352                 return $text;
 353         }
 354
 355         /**
 356          * @todo document
 357          * @return int|string
 358          */
 359         function makeNormalYear( $iso ) {
 360                 if ( $iso[0] == '-' ) {
 361                         $text = ( intval( substr( $iso, 1 ) ) + 1 ) . ' BC';
 362                 } else {
 363                         $text = intval( $iso );
 364                 }
 365                 return $text;
 366         }
 367 }