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         protected $mSource;
  31
  32         protected $mTarget;
  33
  34         /** @var string */
  35         protected $monthNames = '';
  36
  37         /** @todo Are these unused? */
  38         private $rxDM;
  39         private $rxMD;
  40         private $rxDMY;
  41         private $rxYDM;
  42         private $rxMDY;
  43         private $rxYMD;
  44
  45         /** @var array */
  46         protected $regexes;
  47
  48         /** @todo Are these unused? */
  49         private $pDays;
  50         private $pMonths;
  51         private $pYears;
  52
  53         /** @var array */
  54         protected $rules;
  55
  56         /** @var array */
  57         protected $xMonths;
  58
  59         /** @var array */
  60         protected $preferences;
  61
  62         /** @var bool */
  63         protected $mLinked;
  64
  65         protected $lang;
  66
  67         const ALL = -1;
  68         const NONE = 0;
  69         const MDY = 1;
  70         const DMY = 2;
  71         const YMD = 3;
  72         const ISO1 = 4;
  73         const LASTPREF = 4;
  74         const ISO2 = 5;
  75         const YDM = 6;
  76         const DM = 7;
  77         const MD = 8;
  78         const LAST = 8;
  79
  80         /**
  81          * @param Language $lang In which language to format the date
  82          */
  83         function __construct( Language $lang ) {
  84                 $this->lang = $lang;
  85
  86                 $this->monthNames = $this->getMonthRegex();
  87                 for ( $i = 1; $i <= 12; $i++ ) {
  88                         $this->xMonths[$this->lang->lc( $this->lang->getMonthName( $i ) )] = $i;
  89                         $this->xMonths[$this->lang->lc( $this->lang->getMonthAbbreviation( $i ) )] = $i;
  90                 }
  91
  92                 $this->regexTrail = '(?![a-z])/iu';
  93
  94                 # Partial regular expressions
  95                 $this->prxDM = '\[\[(\d{1,2})[ _](' . $this->monthNames . ')\]\]';
  96                 $this->prxMD = '\[\[(' . $this->monthNames . ')[ _](\d{1,2})\]\]';
  97                 $this->prxY = '\[\[(\d{1,4}([ _]BC|))\]\]';
  98                 $this->prxISO1 = '\[\[(-?\d{4})]]-\[\[(\d{2})-(\d{2})\]\]';
  99                 $this->prxISO2 = '\[\[(-?\d{4})-(\d{2})-(\d{2})\]\]';
 100
 101                 # Real regular expressions
 102                 $this->regexes[self::DMY] = "/{$this->prxDM}(?: *, *| +){$this->prxY}{$this->regexTrail}";
 103                 $this->regexes[self::YDM] = "/{$this->prxY}(?: *, *| +){$this->prxDM}{$this->regexTrail}";
 104                 $this->regexes[self::MDY] = "/{$this->prxMD}(?: *, *| +){$this->prxY}{$this->regexTrail}";
 105                 $this->regexes[self::YMD] = "/{$this->prxY}(?: *, *| +){$this->prxMD}{$this->regexTrail}";
 106                 $this->regexes[self::DM] = "/{$this->prxDM}{$this->regexTrail}";
 107                 $this->regexes[self::MD] = "/{$this->prxMD}{$this->regexTrail}";
 108                 $this->regexes[self::ISO1] = "/{$this->prxISO1}{$this->regexTrail}";
 109                 $this->regexes[self::ISO2] = "/{$this->prxISO2}{$this->regexTrail}";
 110
 111                 # Extraction keys
 112                 # See the comments in replace() for the meaning of the letters
 113                 $this->keys[self::DMY] = 'jFY';
 114                 $this->keys[self::YDM] = 'Y jF';
 115                 $this->keys[self::MDY] = 'FjY';
 116                 $this->keys[self::YMD] = 'Y Fj';
 117                 $this->keys[self::DM] = 'jF';
 118                 $this->keys[self::MD] = 'Fj';
 119                 $this->keys[self::ISO1] = 'ymd'; # y means ISO year
 120                 $this->keys[self::ISO2] = 'ymd';
 121
 122                 # Target date formats
 123                 $this->targets[self::DMY] = '[[F j|j F]] [[Y]]';
 124                 $this->targets[self::YDM] = '[[Y]], [[F j|j F]]';
 125                 $this->targets[self::MDY] = '[[F j]], [[Y]]';
 126                 $this->targets[self::YMD] = '[[Y]] [[F j]]';
 127                 $this->targets[self::DM] = '[[F j|j F]]';
 128                 $this->targets[self::MD] = '[[F j]]';
 129                 $this->targets[self::ISO1] = '[[Y|y]]-[[F j|m-d]]';
 130                 $this->targets[self::ISO2] = '[[y-m-d]]';
 131
 132                 # Rules
 133                 #            pref    source       target
 134                 $this->rules[self::DMY][self::MD] = self::DM;
 135                 $this->rules[self::ALL][self::MD] = self::MD;
 136                 $this->rules[self::MDY][self::DM] = self::MD;
 137                 $this->rules[self::ALL][self::DM] = self::DM;
 138                 $this->rules[self::NONE][self::ISO2] = self::ISO1;
 139
 140                 $this->preferences = array(
 141                         'default' => self::NONE,
 142                         'dmy' => self::DMY,
 143                         'mdy' => self::MDY,
 144                         'ymd' => self::YMD,
 145                         'ISO 8601' => self::ISO1,
 146                 );
 147         }
 148
 149         /**
 150          * Get a DateFormatter object
 151          *
 152          * @param Language|string|null $lang In which language to format the date
 153          *              Defaults to the site content language
 154          * @return DateFormatter object
 155          */
 156         public static function &getInstance( $lang = null ) {
 157                 global $wgMemc, $wgContLang;
 158                 static $dateFormatter = false;
 159                 $lang = $lang ? wfGetLangObj( $lang ) : $wgContLang;
 160                 $key = wfMemcKey( 'dateformatter', $lang->getCode() );
 161                 if ( !$dateFormatter ) {
 162                         $dateFormatter = $wgMemc->get( $key );
 163                         if ( !$dateFormatter ) {
 164                                 $dateFormatter = new DateFormatter( $lang );
 165                                 $wgMemc->set( $key, $dateFormatter, 3600 );
 166                         }
 167                 }
 168                 return $dateFormatter;
 169         }
 170
 171         /**
 172          * @param string $preference User preference
 173          * @param string $text Text to reformat
 174          * @param array $options Array can contain 'linked' and/or 'match-whole'
 175          *
 176          * @return string
 177          */
 178         function reformat( $preference, $text, $options = array( 'linked' ) ) {
 179                 $linked = in_array( 'linked', $options );
 180                 $match_whole = in_array( 'match-whole', $options );
 181
 182                 if ( isset( $this->preferences[$preference] ) ) {
 183                         $preference = $this->preferences[$preference];
 184                 } else {
 185                         $preference = self::NONE;
 186                 }
 187                 for ( $i = 1; $i <= self::LAST; $i++ ) {
 188                         $this->mSource = $i;
 189                         if ( isset( $this->rules[$preference][$i] ) ) {
 190                                 # Specific rules
 191                                 $this->mTarget = $this->rules[$preference][$i];
 192                         } elseif ( isset( $this->rules[self::ALL][$i] ) ) {
 193                                 # General rules
 194                                 $this->mTarget = $this->rules[self::ALL][$i];
 195                         } elseif ( $preference ) {
 196                                 # User preference
 197                                 $this->mTarget = $preference;
 198                         } else {
 199                                 # Default
 200                                 $this->mTarget = $i;
 201                         }
 202                         $regex = $this->regexes[$i];
 203
 204                         // Horrible hack
 205                         if ( !$linked ) {
 206                                 $regex = str_replace( array( '\[\[', '\]\]' ), '', $regex );
 207                         }
 208
 209                         if ( $match_whole ) {
 210                                 // Let's hope this works
 211                                 $regex = preg_replace( '!^/!', '/^', $regex );
 212                                 $regex = str_replace( $this->regexTrail,
 213                                         '$' . $this->regexTrail, $regex );
 214                         }
 215
 216                         // Another horrible hack
 217                         $this->mLinked = $linked;
 218                         $text = preg_replace_callback( $regex, array( &$this, 'replace' ), $text );
 219                         unset( $this->mLinked );
 220                 }
 221                 return $text;
 222         }
 223
 224         /**
 225          * @param array $matches
 226          * @return string
 227          */
 228         function replace( $matches ) {
 229                 # Extract information from $matches
 230                 $linked = true;
 231                 if ( isset( $this->mLinked ) ) {
 232                         $linked = $this->mLinked;
 233                 }
 234
 235                 $bits = array();
 236                 $key = $this->keys[$this->mSource];
 237                 $keyLength = strlen( $key );
 238                 for ( $p = 0; $p < $keyLength; $p++ ) {
 239                         if ( $key[$p] != ' ' ) {
 240                                 $bits[$key[$p]] = $matches[$p + 1];
 241                         }
 242                 }
 243
 244                 return $this->formatDate( $bits, $linked );
 245         }
 246
 247         /**
 248          * @param array $bits
 249          * @param bool $link
 250          * @return string
 251          */
 252         function formatDate( $bits, $link = true ) {
 253                 $format = $this->targets[$this->mTarget];
 254
 255                 if ( !$link ) {
 256                         // strip piped links
 257                         $format = preg_replace( '/\[\[[^|]+\|([^\]]+)\]\]/', '$1', $format );
 258                         // strip remaining links
 259                         $format = str_replace( array( '[[', ']]' ), '', $format );
 260                 }
 261
 262                 # Construct new date
 263                 $text = '';
 264                 $fail = false;
 265
 266                 // Pre-generate y/Y stuff because we need the year for the <span> title.
 267                 if ( !isset( $bits['y'] ) && isset( $bits['Y'] ) ) {
 268                         $bits['y'] = $this->makeIsoYear( $bits['Y'] );
 269                 }
 270                 if ( !isset( $bits['Y'] ) && isset( $bits['y'] ) ) {
 271                         $bits['Y'] = $this->makeNormalYear( $bits['y'] );
 272                 }
 273
 274                 if ( !isset( $bits['m'] ) ) {
 275                         $m = $this->makeIsoMonth( $bits['F'] );
 276                         if ( !$m || $m == '00' ) {
 277                                 $fail = true;
 278                         } else {
 279                                 $bits['m'] = $m;
 280                         }
 281                 }
 282
 283                 if ( !isset( $bits['d'] ) ) {
 284                         $bits['d'] = sprintf( '%02d', $bits['j'] );
 285                 }
 286
 287                 $formatLength = strlen( $format );
 288                 for ( $p = 0; $p < $formatLength; $p++ ) {
 289                         $char = $format[$p];
 290                         switch ( $char ) {
 291                                 case 'd': # ISO day of month
 292                                         $text .= $bits['d'];
 293                                         break;
 294                                 case 'm': # ISO month
 295                                         $text .= $bits['m'];
 296                                         break;
 297                                 case 'y': # ISO year
 298                                         $text .= $bits['y'];
 299                                         break;
 300                                 case 'j': # ordinary day of month
 301                                         if ( !isset( $bits['j'] ) ) {
 302                                                 $text .= intval( $bits['d'] );
 303                                         } else {
 304                                                 $text .= $bits['j'];
 305                                         }
 306                                         break;
 307                                 case 'F': # long month
 308                                         if ( !isset( $bits['F'] ) ) {
 309                                                 $m = intval( $bits['m'] );
 310                                                 if ( $m > 12 || $m < 1 ) {
 311                                                         $fail = true;
 312                                                 } else {
 313                                                         $text .= $this->lang->getMonthName( $m );
 314                                                 }
 315                                         } else {
 316                                                 $text .= ucfirst( $bits['F'] );
 317                                         }
 318                                         break;
 319                                 case 'Y': # ordinary (optional BC) year
 320                                         $text .= $bits['Y'];
 321                                         break;
 322                                 default:
 323                                         $text .= $char;
 324                         }
 325                 }
 326                 if ( $fail ) {
 327                         /** @todo FIXME: $matches doesn't exist here, what's expected? */
 328                         $text = $matches[0];
 329                 }
 330
 331                 $isoBits = array();
 332                 if ( isset( $bits['y'] ) ) {
 333                         $isoBits[] = $bits['y'];
 334                 }
 335                 $isoBits[] = $bits['m'];
 336                 $isoBits[] = $bits['d'];
 337                 $isoDate = implode( '-', $isoBits );
 338
 339                 // Output is not strictly HTML (it's wikitext), but <span> is whitelisted.
 340                 $text = Html::rawElement( 'span',
 341                                         array( 'class' => 'mw-formatted-date', 'title' => $isoDate ), $text );
 342
 343                 return $text;
 344         }
 345
 346         /**
 347          * @todo document
 348          * @return string
 349          */
 350         function getMonthRegex() {
 351                 $names = array();
 352                 for ( $i = 1; $i <= 12; $i++ ) {
 353                         $names[] = $this->lang->getMonthName( $i );
 354                         $names[] = $this->lang->getMonthAbbreviation( $i );
 355                 }
 356                 return implode( '|', $names );
 357         }
 358
 359         /**
 360          * Makes an ISO month, e.g. 02, from a month name
 361          * @param string $monthName Month name
 362          * @return string ISO month name
 363          */
 364         function makeIsoMonth( $monthName ) {
 365                 $n = $this->xMonths[$this->lang->lc( $monthName )];
 366                 return sprintf( '%02d', $n );
 367         }
 368
 369         /**
 370          * @todo document
 371          * @param string $year Year name
 372          * @return string ISO year name
 373          */
 374         function makeIsoYear( $year ) {
 375                 # Assumes the year is in a nice format, as enforced by the regex
 376                 if ( substr( $year, -2 ) == 'BC' ) {
 377                         $num = intval( substr( $year, 0, -3 ) ) - 1;
 378                         # PHP bug note: sprintf( "%04d", -1 ) fails poorly
 379                         $text = sprintf( '-%04d', $num );
 380
 381                 } else {
 382                         $text = sprintf( '%04d', $year );
 383                 }
 384                 return $text;
 385         }
 386
 387         /**
 388          * @todo document
 389          * @param string $iso
 390          * @return int|string
 391          */
 392         function makeNormalYear( $iso ) {
 393                 if ( $iso[0] == '-' ) {
 394                         $text = ( intval( substr( $iso, 1 ) ) + 1 ) . ' BC';
 395                 } else {
 396                         $text = intval( $iso );
 397                 }
 398                 return $text;
 399         }
 400 }