plugins/DirectionDetector/DirectionDetectorPlugin.php

   1 <?php
   2
   3 /**
   4  * DirectionDetector plugin, detects notices with RTL content & sets RTL
   5  * style for them.
   6  *
   7  * This program is free software: you can redistribute it and/or modify
   8  * it under the terms of the GNU General Public License as published by
   9  * the Free Software Foundation, either version 3 of the License, or
  10  * (at your option) any later version.
  11  *
  12  * This program is distributed in the hope that it will be useful,
  13  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  14  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
  15  * GNU General Public License for more details.
  16  *
  17  * You should have received a copy of the GNU General Public License
  18  * along with this program.     If not, see <http://www.gnu.org/licenses/>.
  19  *
  20  * @category     Plugin
  21  * @package      StatusNet
  22  * @author       Behrooz shabani (everplays) - <behrooz@rock.com>
  23  * @copyright    2009-2010 Behrooz shabani
  24  * @license      http://www.fsf.org/licensing/licenses/agpl-3.0.html AGPL 3.0
  25  *
  26  */
  27
  28 if (!defined('STATUSNET')) {
  29                 exit(1);
  30 }
  31
  32 define('DIRECTIONDETECTORPLUGIN_VERSION', '0.1.1');
  33
  34 class DirectionDetectorPlugin extends Plugin {
  35         /**
  36          * SN plugin API, here we will make changes on rendered column
  37          *
  38          * @param object $notice notice is going to be saved
  39          */
  40         public function onStartNoticeSave(&$notice){
  41                 if(self::isRTL($notice->content))
  42                         $notice->rendered = '<span class="rtl">'.$notice->rendered.'</span>';
  43                 return true;
  44         }
  45
  46         /**
  47          * SN plugin API, here we will add css needed for modifiyed rendered
  48          *
  49          * @param
  50          */
  51         public function onEndShowStatusNetStyles(&$xml){
  52                 $xml->element('style', array('type' => 'text/css'), 'span.rtl {display:block;direction:rtl;text-align:right;float:right;width:490px;} .notice .author {float:left}');
  53         }
  54         /**
  55          * checks that passed string is a RTL language or not
  56          *
  57          * @param string $str string to be checked
  58          */
  59         public static function isRTL($str){
  60                 self::getClearText($str);
  61                 if( is_array($cc = self::utf8ToUnicode(mb_substr($str, 0, 1, 'utf-8'))) )
  62                         $cc = $cc[0];
  63                 else
  64                         return false;
  65                 if($cc>=1536 && $cc<=1791) // arabic, persian, urdu, kurdish, ...
  66                         return true;
  67                 if($cc>=65136 && $cc<=65279) // arabic peresent 2
  68                         return true;
  69                 if($cc>=64336 && $cc<=65023) // arabic peresent 1
  70                         return true;
  71                 if($cc>=1424 && $cc<=1535) // hebrew
  72                         return true;
  73                 if($cc>=64256 && $cc<=64335) // hebrew peresent
  74                         return true;
  75                 if($cc>=1792 && $cc<=1871) // Syriac
  76                         return true;
  77                 if($cc>=1920 && $cc<=1983) // Thaana
  78                         return true;
  79                 if($cc>=1984 && $cc<=2047) // NKo
  80                         return true;
  81                 if($cc>=11568 && $cc<=11647) // Tifinagh
  82                         return true;
  83                 return false;
  84         }
  85
  86         /**
  87          * clears text from replys, tags, groups, reteets & whitespaces
  88          *
  89          * @param string &$str string to be cleared
  90          */
  91         private static function getClearText(&$str){
  92                 $str = preg_replace('/@[^ ]+|![^ ]+|#[^ ]+/u', '', $str); // reply, tag, group
  93                 $str = preg_replace('/^RT[: ]{1}| RT | RT: |^RD[: ]{1}| RD | RD: |[♺♻:]/u', '', $str); // redent, retweet
  94                 $str = preg_replace("/[ \r\t\n]+/", ' ', trim($str)); // remove spaces
  95         }
  96
  97         /**
  98          * Takes an UTF-8 string and returns an array of ints representing the
  99          * Unicode characters. Astral planes are supported ie. the ints in the
 100          * output can be > 0xFFFF. O$ccurrances of the BOM are ignored. Surrogates
 101          * are not allowed. ### modified ### returns first character code
 102          *
 103          * Returns false if the input string isn't a valid UTF-8 octet sequence.
 104          */
 105         private static function utf8ToUnicode(&$str){
 106                 $mState = 0;       // cached expected number of octets after the current octet
 107                                    // until the beginning of the next UTF8 character sequence
 108                 $mUcs4  = 0;     // cached Unicode character
 109                 $mBytes = 1;       // cached expected number of octets in the current sequence
 110                 $out = array();
 111                 $len = strlen($str);
 112
 113                 for($i = 0; $i < $len; $i++) {
 114                         $in = ord($str{$i});
 115                         if (0 == $mState) {
 116                                 // When mState is zero we expect either a US-ASCII character or a
 117                                 // multi-octet sequence.
 118                                 if (0 == (0x80 & ($in))) {
 119                                         // US-ASCII, pass straight through.
 120                                         $out[] = $in;
 121                                         $mBytes = 1;
 122                                 } elseif (0xC0 == (0xE0 & ($in))) {
 123                                         // First octet of 2 octet sequence
 124                                         $mUcs4 = ($in);
 125                                         $mUcs4 = ($mUcs4 & 0x1F) << 6;
 126                                         $mState = 1;
 127                                         $mBytes = 2;
 128                                 } elseif (0xE0 == (0xF0 & ($in))) {
 129                                         // First octet of 3 octet sequence
 130                                         $mUcs4 = ($in);
 131                                         $mUcs4 = ($mUcs4 & 0x0F) << 12;
 132                                         $mState = 2;
 133                                         $mBytes = 3;
 134                                 } elseif (0xF0 == (0xF8 & ($in))) {
 135                                         // First octet of 4 octet sequence
 136                                         $mUcs4 = ($in);
 137                                         $mUcs4 = ($mUcs4 & 0x07) << 18;
 138                                         $mState = 3;
 139                                         $mBytes = 4;
 140                                 } elseif (0xF8 == (0xFC & ($in))) {
 141                                         /* First octet of 5 octet sequence.
 142                                          *
 143                                          * This is illegal because the encoded codepoint must be either
 144                                          * (a) not the shortest form or
 145                                          * (b) outside the Unicode range of 0-0x10FFFF.
 146                                          * Rather than trying to resynchronize, we will carry on until the end
 147                                          * of the sequence and let the later error handling code catch it.
 148                                          */
 149                                         $mUcs4 = ($in);
 150                                         $mUcs4 = ($mUcs4 & 0x03) << 24;
 151                                         $mState = 4;
 152                                         $mBytes = 5;
 153                                 } elseif (0xFC == (0xFE & ($in))) {
 154                                         // First octet of 6 octet sequence, see comments for 5 octet sequence.
 155                                         $mUcs4 = ($in);
 156                                         $mUcs4 = ($mUcs4 & 1) << 30;
 157                                         $mState = 5;
 158                                         $mBytes = 6;
 159                                 } else {
 160                                         /* Current octet is neither in the US-ASCII range nor a legal first
 161                                          * octet of a multi-octet sequence.
 162                                          */
 163                                         return false;
 164                                 }
 165                         } else {
 166                                 // When mState is non-zero, we expect a continuation of the multi-octet
 167                                 // sequence
 168                                 if (0x80 == (0xC0 & ($in))) {
 169                                         // Legal continuation.
 170                                         $shift = ($mState - 1) * 6;
 171                                         $tmp = $in;
 172                                         $tmp = ($tmp & 0x0000003F) << $shift;
 173                                         $mUcs4 |= $tmp;
 174                                         if (0 == --$mState) {
 175                                                 /* End of the multi-octet sequence. mUcs4 now contains the final
 176                                                  * Unicode codepoint to be output
 177                                                  *
 178                                                  * Check for illegal sequences and codepoints.
 179                                                  */
 180                                                 // From Unicode 3.1, non-shortest form is illegal
 181                                                 if      (
 182                                                                 ((2 == $mBytes) && ($mUcs4 < 0x0080)) ||
 183                                                                 ((3 == $mBytes) && ($mUcs4 < 0x0800)) ||
 184                                                                 ((4 == $mBytes) && ($mUcs4 < 0x10000)) ||
 185                                                                 (4 < $mBytes) ||
 186                                                                 // From Unicode 3.2, surrogate characters are illegal
 187                                                                 (($mUcs4 & 0xFFFFF800) == 0xD800) ||
 188                                                                 // Codepoints outside the Unicode range are illegal
 189                                                                 ($mUcs4 > 0x10FFFF)
 190                                                         ){
 191                                                         return false;
 192                                                 }
 193                                                 if (0xFEFF != $mUcs4) {
 194                                                         $out[] = $mUcs4;
 195                                                 }
 196                                                 //initialize UTF8 cache
 197                                                 $mState = 0;
 198                                                 $mUcs4  = 0;
 199                                                 $mBytes = 1;
 200                                         }
 201                                 } else {
 202                                         /* ((0xC0 & (*in) != 0x80) && (mState != 0))
 203                                          *
 204                                          * Incomplete multi-octet sequence.
 205                                          */
 206                                         return false;
 207                                 }
 208                         }
 209                 }
 210                 return $out;
 211         }
 212
 213         /**
 214          * plugin details
 215          */
 216         function onPluginVersion(&$versions){
 217                 $versions[] = array(
 218                         'name' => 'Direction detector',
 219                         'version' => DIRECTIONDETECTORPLUGIN_VERSION,
 220                         'author' => 'behrooz shabani',
 221                         'rawdescription' => _m('shows notices with right-to-left content in correct direction.')
 222                 );
 223                 return true;
 224         }
 225 }
 226
 227 /*
 228 // Example:
 229 var_dump(DirectionDetectorPlugin::isRTL('RT @everplays ♺: دادگاه به دليل عدم حضور وکلای متهمان بنا بر اصل ١٣٥ قانون اساسی غير قانونی است')); // true
 230 */