sandbox/alaa/bidi.module

<?php // -*-php-*-
/*
 * Original module by Alaa Abd El Fatah.
 * Mohammed Sameer: My Modifications Copyright (c) 2004 Mohammed Sameer, 
 *                  under the GNU GPL v2 or later.
 * Mohammed Sameer: 2004 10 30
 *                  * Ported to drupal 4.5 API.
 */

setlocale(LC_ALL, "ar_EG.UTF-8");

function bidi_help($section) {
  $output = "";

  switch ($section) {
    case 'admin/modules#description':
      $output = t("Automaticaly sets line direction");
      break;
  }

  return $output;
}


function bidi_filter($op, $delta = 0, $format = -1, $text = "") {
  switch ($op) {
  case 'list':
    return array(0 => t('BiDi'));
    //    case "name":
    //    return t("Bidi filter");
  case 'description':
    return t("Bidi filter");
  case "process":
      return _bidi_filter_process($text);
  case "settings":
      return _bidi_filter_settings($text);
  default:
      return $text;
  }
}


//get dir needs to be split into a single include file so it can be used on theme and other places
function _get_dir($text) {
  $text = strtolower($text);
  $len = strlen($text);

  $arabic = 0;
  $english = 0;
  
  //constants
  $a = ord('a');
  $z = ord('z');
  
  //arabic UTF-8 letters have one of these values in the 1st byte
  $ar1 = 0xd8;
  $ar2 = 0xd9;
    
  // this calculates the dominant language in the text, most bidi implementations just use the first letter (which may be more efficient)
  for ($i = 0; $i < $len; ++$i) {
    $bin = ord($text[$i]);

    if ($bin == $ar1 || $bin == $ar2) {
      ++$arabic;
      ++$i;
    } else if ($bin >= $a && $bin <= $z) {
      ++$english;
    }
  }

  //should this return a bool instead?
  if ($english || $arabic) {
    if ($english > $arabic)	
      return ' class="left2right"';
    else
      return ' class="right2left"';
  } else {
    return "";
  }
}


function _bidi_filter_process($text) {
  $output = "";

  $len = strlen($text);
  $i = 0;
  

  while ($i < $len) {
    //find next tag
    $e = strpos(substr($text, $i), "<");
    if ($e === FALSE) break;
    
    $output.=substr($text, $i, $e);
    
    $i += $e;
    
    //find end of next tag
    $e = strcspn(substr($text, $i), "/> ");    
    $tag = substr($text, $i+1, $e-1);

    $output.=substr($text, $i, $e);
    $i += $e;
    
    
    switch ($tag) {
      //single tag entity
    case "img":
    case "br":
    case "hr":
      continue;
      break;

      //nested cases
    case "ul":
    case "ol":
    case "dl":
    case "div":
      $e = 0;
      $j = 1; //stack counter

      // if $j = 0 then stack is empty
      while ($j > 0) {	
	$e += strpos(substr($text, $i+$e) , $tag);
	
	//yeah I know I'm assuming certain sizes for tags, sue me
	if ($text[$i+$e-1] == '<') { //new nested block
	  ++$j;
	} else if ($text[$i+$e-1] == '/') { //end of block
	  --$j;
	}
	$e += 2; //skip current tag
      }      
      $content = substr($text, $i+1, $e); 
      $output.=_get_dir(strip_tags($content));
      $e += strlen($tag);
      $output.= substr($text, $i, $e);
      $i += $e;      
      break;
      

      //block elements these get out attention
    case "p":
    case "blockquote":
    case "pre":
    case "h1":
    case "h2":
    case "h3":
    case "h4":
    case "h5":
    case "h6":

      //nested lists are proken better think of a fix
    case "li":
      $close_tag = "</".$tag.">";
      //breaks on extra whitespaces, maybe I should use regexps instead
      $e = strpos(substr($text, $i), $close_tag);
      $content = substr($text, $i+1, $e-1);
      //calculate directions
      $output.= _get_dir(strip_tags($content));
      $e += strlen($close_tag);
      $output.= substr($text, $i, $e);
      $i += $e;
      
      break;
      
      //non block tags we are not handeling
    default:
      $close_tag = "</".$tag.">";
      $e = strpos(substr($text, $i), $close_tag) + strlen($close_tag);
      $output.= substr($text, $i, $e);
      $i += $e;
    }

  }
    
  //any missing content
  $output .= substr($text, $i);
  
  
  return $output;  
}


function _bidi_filter_settings() {
  return form_group(t("Bidi filter"), t("automatic bidi"));
}

function bidi_filter_tips() {
  if ($long)
    {

    }
  else {
  return t("You may write mixed Arabic and English freely, line direction will be computed automaticaly");
    }
}

?>
1	alaa	1.1	<?php // --php--
2			/*
3			* Original module by Alaa Abd El Fatah.
4			* Mohammed Sameer: My Modifications Copyright (c) 2004 Mohammed Sameer,
5			* under the GNU GPL v2 or later.
6			* Mohammed Sameer: 2004 10 30
7			* * Ported to drupal 4.5 API.
8			*/
9
10			setlocale(LC_ALL, "ar_EG.UTF-8");
11
12			function bidi_help($section) {
13			$output = "";
14
15			switch ($section) {
16			case 'admin/modules#description':
17			$output = t("Automaticaly sets line direction");
18			break;
19			}
20
21			return $output;
22			}
23
24
25			function bidi_filter($op, $delta = 0, $format = -1, $text = "") {
26			switch ($op) {
27			case 'list':
28			return array(0 => t('BiDi'));
29			// case "name":
30			// return t("Bidi filter");
31			case 'description':
32			return t("Bidi filter");
33			case "process":
34			return _bidi_filter_process($text);
35			case "settings":
36			return _bidi_filter_settings($text);
37			default:
38			return $text;
39			}
40			}
41
42
43			//get dir needs to be split into a single include file so it can be used on theme and other places
44			function _get_dir($text) {
45			$text = strtolower($text);
46			$len = strlen($text);
47
48			$arabic = 0;
49			$english = 0;
50
51			//constants
52			$a = ord('a');
53			$z = ord('z');
54
55			//arabic UTF-8 letters have one of these values in the 1st byte
56			$ar1 = 0xd8;
57			$ar2 = 0xd9;
58
59			// this calculates the dominant language in the text, most bidi implementations just use the first letter (which may be more efficient)
60			for ($i = 0; $i < $len; ++$i) {
61			$bin = ord($text[$i]);
62
63			if ($bin == $ar1 \|\| $bin == $ar2) {
64			++$arabic;
65			++$i;
66			} else if ($bin >= $a && $bin <= $z) {
67			++$english;
68			}
69			}
70
71			//should this return a bool instead?
72			if ($english \|\| $arabic) {
73			if ($english > $arabic)
74			return ' class="left2right"';
75			else
76			return ' class="right2left"';
77			} else {
78			return "";
79			}
80			}
81
82
83			function _bidi_filter_process($text) {
84			$output = "";
85
86			$len = strlen($text);
87			$i = 0;
88
89
90			while ($i < $len) {
91			//find next tag
92			$e = strpos(substr($text, $i), "<");
93			if ($e === FALSE) break;
94
95			$output.=substr($text, $i, $e);
96
97			$i += $e;
98
99			//find end of next tag
100			$e = strcspn(substr($text, $i), "/> ");
101			$tag = substr($text, $i+1, $e-1);
102
103			$output.=substr($text, $i, $e);
104			$i += $e;
105
106
107			switch ($tag) {
108			//single tag entity
109			case "img":
110			case "br":
111			case "hr":
112			continue;
113			break;
114
115			//nested cases
116			case "ul":
117			case "ol":
118			case "dl":
119			case "div":
120			$e = 0;
121			$j = 1; //stack counter
122
123			// if $j = 0 then stack is empty
124			while ($j > 0) {
125			$e += strpos(substr($text, $i+$e) , $tag);
126
127			//yeah I know I'm assuming certain sizes for tags, sue me
128			if ($text[$i+$e-1] == '<') { //new nested block
129			++$j;
130			} else if ($text[$i+$e-1] == '/') { //end of block
131			--$j;
132			}
133			$e += 2; //skip current tag
134			}
135			$content = substr($text, $i+1, $e);
136			$output.=_get_dir(strip_tags($content));
137			$e += strlen($tag);
138			$output.= substr($text, $i, $e);
139			$i += $e;
140			break;
141
142
143			//block elements these get out attention
144			case "p":
145			case "blockquote":
146			case "pre":
147			case "h1":
148			case "h2":
149			case "h3":
150			case "h4":
151			case "h5":
152			case "h6":
153
154			//nested lists are proken better think of a fix
155			case "li":
156			$close_tag = "</".$tag.">";
157			//breaks on extra whitespaces, maybe I should use regexps instead
158			$e = strpos(substr($text, $i), $close_tag);
159			$content = substr($text, $i+1, $e-1);
160			//calculate directions
161			$output.= _get_dir(strip_tags($content));
162			$e += strlen($close_tag);
163			$output.= substr($text, $i, $e);
164			$i += $e;
165
166			break;
167
168			//non block tags we are not handeling
169			default:
170			$close_tag = "</".$tag.">";
171			$e = strpos(substr($text, $i), $close_tag) + strlen($close_tag);
172			$output.= substr($text, $i, $e);
173			$i += $e;
174			}
175
176			}
177
178			//any missing content
179			$output .= substr($text, $i);
180
181
182			return $output;
183			}
184
185
186			function _bidi_filter_settings() {
187			return form_group(t("Bidi filter"), t("automatic bidi"));
188			}
189
190			function bidi_filter_tips() {
191			if ($long)
192			{
193
194			}
195			else {
196			return t("You may write mixed Arabic and English freely, line direction will be computed automaticaly");
197			}
198			}
199
200			?>