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 accoding to user preferences.
  26  * @todo preferences, OutputPage
  27  * @ingroup Parser
  28  */
  29 class DateFormatter
  30 {
  31         var $mSource, $mTarget;
  32         var $monthNames = '', $rxDM, $rxMD, $rxDMY, $rxYDM, $rxMDY, $rxYMD;
  33
  34         var $regexes, $pDays, $pMonths, $pYears;
  35         var $rules, $xMonths, $preferences;
  36
  37         const ALL = -1;
  38         const NONE = 0;
  39         const MDY = 1;
  40         const DMY = 2;
  41         const YMD = 3;
  42         const ISO1 = 4;
  43         const LASTPREF = 4;
  44         const ISO2 = 5;
  45         const YDM = 6;
  46         const DM = 7;
  47         const MD = 8;
  48         const LAST = 8;
  49
  50         /**
  51          * @todo document
  52          */
  53         function __construct() {
  54                 global $wgContLang;
  55
  56                 $this->monthNames = $this->getMonthRegex();
  57                 for ( $i=1; $i<=12; $i++ ) {
  58                         $this->xMonths[$wgContLang->lc( $wgContLang->getMonthName( $i ) )] = $i;
  59                         $this->xMonths[$wgContLang->lc( $wgContLang->getMonthAbbreviation( $i ) )] = $i;
  60                 }
  61
  62                 $this->regexTrail = '(?![a-z])/iu';
  63
  64                 # Partial regular expressions
  65                 $this->prxDM = '\[\[(\d{1,2})[ _](' . $this->monthNames . ')\]\]';
  66                 $this->prxMD = '\[\[(' . $this->monthNames . ')[ _](\d{1,2})\]\]';
  67                 $this->prxY = '\[\[(\d{1,4}([ _]BC|))\]\]';
  68                 $this->prxISO1 = '\[\[(-?\d{4})]]-\[\[(\d{2})-(\d{2})\]\]';
  69                 $this->prxISO2 = '\[\[(-?\d{4})-(\d{2})-(\d{2})\]\]';
  70
  71                 # Real regular expressions
  72                 $this->regexes[self::DMY] = "/{$this->prxDM}(?: *, *| +){$this->prxY}{$this->regexTrail}";
  73                 $this->regexes[self::YDM] = "/{$this->prxY}(?: *, *| +){$this->prxDM}{$this->regexTrail}";
  74                 $this->regexes[self::MDY] = "/{$this->prxMD}(?: *, *| +){$this->prxY}{$this->regexTrail}";
  75                 $this->regexes[self::YMD] = "/{$this->prxY}(?: *, *| +){$this->prxMD}{$this->regexTrail}";
  76                 $this->regexes[self::DM] = "/{$this->prxDM}{$this->regexTrail}";
  77                 $this->regexes[self::MD] = "/{$this->prxMD}{$this->regexTrail}";
  78                 $this->regexes[self::ISO1] = "/{$this->prxISO1}{$this->regexTrail}";
  79                 $this->regexes[self::ISO2] = "/{$this->prxISO2}{$this->regexTrail}";
  80
  81                 # Extraction keys
  82                 # See the comments in replace() for the meaning of the letters
  83                 $this->keys[self::DMY] = 'jFY';
  84                 $this->keys[self::YDM] = 'Y jF';
  85                 $this->keys[self::MDY] = 'FjY';
  86                 $this->keys[self::YMD] = 'Y Fj';
  87                 $this->keys[self::DM] = 'jF';
  88                 $this->keys[self::MD] = 'Fj';
  89                 $this->keys[self::ISO1] = 'ymd'; # y means ISO year
  90                 $this->keys[self::ISO2] = 'ymd';
  91
  92                 # Target date formats
  93                 $this->targets[self::DMY] = '[[F j|j F]] [[Y]]';
  94                 $this->targets[self::YDM] = '[[Y]], [[F j|j F]]';
  95                 $this->targets[self::MDY] = '[[F j]], [[Y]]';
  96                 $this->targets[self::YMD] = '[[Y]] [[F j]]';
  97                 $this->targets[self::DM] = '[[F j|j F]]';
  98                 $this->targets[self::MD] = '[[F j]]';
  99                 $this->targets[self::ISO1] = '[[Y|y]]-[[F j|m-d]]';
 100                 $this->targets[self::ISO2] = '[[y-m-d]]';
 101
 102                 # Rules
 103                 #            pref    source       target
 104                 $this->rules[self::DMY][self::MD]       = self::DM;
 105                 $this->rules[self::ALL][self::MD]       = self::MD;
 106                 $this->rules[self::MDY][self::DM]       = self::MD;
 107                 $this->rules[self::ALL][self::DM]       = self::DM;
 108                 $this->rules[self::NONE][self::ISO2]    = self::ISO1;
 109
 110                 $this->preferences = array(
 111                         'default' => self::NONE,
 112                         'dmy' => self::DMY,
 113                         'mdy' => self::MDY,
 114                         'ymd' => self::YMD,
 115                         'ISO 8601' => self::ISO1,
 116                 );
 117         }
 118
 119         /**
 120          * Get a DateFormatter object
 121          *
 122          * @return DateFormatter object
 123          */
 124         public static function &getInstance() {
 125                 global $wgMemc;
 126                 static $dateFormatter = false;
 127                 if ( !$dateFormatter ) {
 128                         $dateFormatter = $wgMemc->get( wfMemcKey( 'dateformatter' ) );
 129                         if ( !$dateFormatter ) {
 130                                 $dateFormatter = new DateFormatter;
 131                                 $wgMemc->set( wfMemcKey( 'dateformatter' ), $dateFormatter, 3600 );
 132                         }
 133                 }
 134                 return $dateFormatter;
 135         }
 136
 137         /**
 138          * @param $preference String: User preference
 139          * @param $text String: Text to reformat
 140          * @param $options Array: can contain 'linked' and/or 'match-whole'
 141          * @return mixed|String
 142          */
 143         function reformat( $preference, $text, $options = array('linked') ) {
 144
 145                 $linked = in_array( 'linked', $options );
 146                 $match_whole = in_array( 'match-whole', $options );
 147
 148                 if ( isset( $this->preferences[$preference] ) ) {
 149                         $preference = $this->preferences[$preference];
 150                 } else {
 151                         $preference = self::NONE;
 152                 }
 153                 for ( $i=1; $i<=self::LAST; $i++ ) {
 154                         $this->mSource = $i;
 155                         if ( isset ( $this->rules[$preference][$i] ) ) {
 156                                 # Specific rules
 157                                 $this->mTarget = $this->rules[$preference][$i];
 158                         } elseif ( isset ( $this->rules[self::ALL][$i] ) ) {
 159                                 # General rules
 160                                 $this->mTarget = $this->rules[self::ALL][$i];
 161                         } elseif ( $preference ) {
 162                                 # User preference
 163                                 $this->mTarget = $preference;
 164                         } else {
 165                                 # Default
 166                                 $this->mTarget = $i;
 167                         }
 168                         $regex = $this->regexes[$i];
 169
 170                         // Horrible hack
 171                         if (!$linked) {
 172                                 $regex = str_replace( array( '\[\[', '\]\]' ), '', $regex );
 173                         }
 174
 175                         if ($match_whole) {
 176                                 // Let's hope this works
 177                                 $regex = preg_replace( '!^/!', '/^', $regex );
 178                                 $regex = str_replace( $this->regexTrail,
 179                                         '$'.$this->regexTrail, $regex );
 180                         }
 181
 182                         // Another horrible hack
 183                         $this->mLinked = $linked;
 184                         $text = preg_replace_callback( $regex, array( &$this, 'replace' ), $text );
 185                         unset($this->mLinked);
 186                 }
 187                 return $text;
 188         }
 189
 190         /**
 191          * @param $matches
 192          * @return string
 193          */
 194         function replace( $matches ) {
 195                 # Extract information from $matches
 196                 $linked = true;
 197                 if ( isset( $this->mLinked ) )
 198                         $linked = $this->mLinked;
 199
 200                 $bits = array();
 201                 $key = $this->keys[$this->mSource];
 202                 for ( $p=0; $p < strlen($key); $p++ ) {
 203                         if ( $key[$p] != ' ' ) {
 204                                 $bits[$key[$p]] = $matches[$p+1];
 205                         }
 206                 }
 207
 208                 return $this->formatDate( $bits, $linked );
 209         }
 210
 211         function formatDate( $bits, $link = true ) {
 212                 $format = $this->targets[$this->mTarget];
 213
 214                 if (!$link) {
 215                         // strip piped links
 216                         $format = preg_replace( '/\[\[[^|]+\|([^\]]+)\]\]/', '$1', $format );
 217                         // strip remaining links
 218                         $format = str_replace( array( '[[', ']]' ), '', $format );
 219                 }
 220
 221                 # Construct new date
 222                 $text = '';
 223                 $fail = false;
 224
 225                 // Pre-generate y/Y stuff because we need the year for the <span> title.
 226                 if ( !isset( $bits['y'] ) && isset( $bits['Y'] ) )
 227                         $bits['y'] = $this->makeIsoYear( $bits['Y'] );
 228                 if ( !isset( $bits['Y'] ) && isset( $bits['y'] ) )
 229                         $bits['Y'] = $this->makeNormalYear( $bits['y'] );
 230
 231                 if ( !isset( $bits['m'] ) ) {
 232                         $m = $this->makeIsoMonth( $bits['F'] );
 233                         if ( !$m || $m == '00' ) {
 234                                 $fail = true;
 235                         } else {
 236                                 $bits['m'] = $m;
 237                         }
 238                 }
 239
 240                 if ( !isset($bits['d']) ) {
 241                         $bits['d'] = sprintf( '%02d', $bits['j'] );
 242                 }
 243
 244                 for ( $p=0; $p < strlen( $format ); $p++ ) {
 245                         $char = $format[$p];
 246                         switch ( $char ) {
 247                                 case 'd': # ISO day of month
 248                                         $text .= $bits['d'];
 249                                         break;
 250                                 case 'm': # ISO month
 251                                         $text .= $bits['m'];
 252                                         break;
 253                                 case 'y': # ISO year
 254                                         $text .= $bits['y'];
 255                                         break;
 256                                 case 'j': # ordinary day of month
 257                                         if ( !isset($bits['j']) ) {
 258                                                 $text .= intval( $bits['d'] );
 259                                         } else {
 260                                                 $text .= $bits['j'];
 261                                         }
 262                                         break;
 263                                 case 'F': # long month
 264                                         if ( !isset( $bits['F'] ) ) {
 265                                                 $m = intval($bits['m']);
 266                                                 if ( $m > 12 || $m < 1 ) {
 267                                                         $fail = true;
 268                                                 } else {
 269                                                         global $wgContLang;
 270                                                         $text .= $wgContLang->getMonthName( $m );
 271                                                 }
 272                                         } else {
 273                                                 $text .= ucfirst( $bits['F'] );
 274                                         }
 275                                         break;
 276                                 case 'Y': # ordinary (optional BC) year
 277                                         $text .= $bits['Y'];
 278                                         break;
 279                                 default:
 280                                         $text .= $char;
 281                         }
 282                 }
 283                 if ( $fail ) {
 284                         $text = $matches[0];
 285                 }
 286
 287                 $isoBits = array();
 288                 if ( isset($bits['y']) )
 289                         $isoBits[] = $bits['y'];
 290                 $isoBits[] = $bits['m'];
 291                 $isoBits[] = $bits['d'];
 292                 $isoDate = implode( '-', $isoBits );
 293
 294                 // Output is not strictly HTML (it's wikitext), but <span> is whitelisted.
 295                 $text = Html::rawElement( 'span',
 296                                         array( 'class' => 'mw-formatted-date', 'title' => $isoDate ), $text );
 297
 298                 return $text;
 299         }
 300
 301         /**
 302          * @todo document
 303          * @return string
 304          */
 305         function getMonthRegex() {
 306                 global $wgContLang;
 307                 $names = array();
 308                 for( $i = 1; $i <= 12; $i++ ) {
 309                         $names[] = $wgContLang->getMonthName( $i );
 310                         $names[] = $wgContLang->getMonthAbbreviation( $i );
 311                 }
 312                 return implode( '|', $names );
 313         }
 314
 315         /**
 316          * Makes an ISO month, e.g. 02, from a month name
 317          * @param $monthName String: month name
 318          * @return string ISO month name
 319          */
 320         function makeIsoMonth( $monthName ) {
 321                 global $wgContLang;
 322
 323                 $n = $this->xMonths[$wgContLang->lc( $monthName )];
 324                 return sprintf( '%02d', $n );
 325         }
 326
 327         /**
 328          * @todo document
 329          * @param $year String: Year name
 330          * @return string ISO year name
 331          */
 332         function makeIsoYear( $year ) {
 333                 # Assumes the year is in a nice format, as enforced by the regex
 334                 if ( substr( $year, -2 ) == 'BC' ) {
 335                         $num = intval(substr( $year, 0, -3 )) - 1;
 336                         # PHP bug note: sprintf( "%04d", -1 ) fails poorly
 337                         $text = sprintf( '-%04d', $num );
 338
 339                 } else {
 340                         $text = sprintf( '%04d', $year );
 341                 }
 342                 return $text;
 343         }
 344
 345         /**
 346          * @todo document
 347          * @return int|string
 348          */
 349         function makeNormalYear( $iso ) {
 350                 if ( $iso[0] == '-' ) {
 351                         $text = (intval( substr( $iso, 1 ) ) + 1) . ' BC';
 352                 } else {
 353                         $text = intval( $iso );
 354                 }
 355                 return $text;
 356         }
 357 }