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.
 *
 * 04/04/2005 Ported to drupal 4.6 API by Amr Mostafa
 */

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


/**
 * Implementation of hook_help()
 */
function bidi_help($section) {
  $output = "";

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

  return $output;
}

/**
 * Implementation of hook_filter_tips()
 */
function bidi_filter_tips($delta, $format, $long = false) {
  if ($long) {
    return t("You may write mixed Arabic and English freely, line direction will be computed automaticaly");
  }
  else {
    return t("You may write mixed Arabic and English freely, line direction will be computed automaticaly");
  }
}

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

/**
 * Determine text language (arabic or english).
 * This function needs to be split into a single include file so it can be used on theme and other places.
 *
 * @param $text
 *   The text needs to be.
 * @return
 *   right2left or left2right depending on the detected language.
 */
function _get_dir($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 ' dir="LTR" ';
    else
      return ' dir="RTL" ';
  }
  else {
    return "";
  }
}

/**
 * Sets HTML text line directions (rtl or ltr) as needed.
 *
 * @param $text
 *   The HTML text needs to be processed.
 *
 * @return
 *   The processed result, with line directions are set using css
 *   classes "righ2left" and "left2right".
 */
function _bidi_filter_process($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;
        // stack counter
        $j = 1;
        
        // 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
          // new nested block
          if ($text[$i+$e-1] == '<') {
            ++$j;
          }
          // end of block
          else if ($text[$i+$e-1] == '/') {
            --$j;
          }
          // skip current tag
          $e += 2;
        }
        
        $content = substr($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;
}

?>
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	* 04/04/2005 Ported to drupal 4.6 API by Amr Mostafa
10	*/
11
12	setlocale(LC_ALL, "ar_EG.UTF-8");
13
14
15	/**
16	* Implementation of hook_help()
17	*/
18	function bidi_help($section) {
19	$output = "";
20
21	switch ($section) {
22	case 'admin/modules#description':
23	$output = t("Automaticaly sets line direction");
24	break;
25	}
26
27	return $output;
28	}
29
30	/**
31	* Implementation of hook_filter_tips()
32	*/
33	function bidi_filter_tips($delta, $format, $long = false) {
34	if ($long) {
35	return t("You may write mixed Arabic and English freely, line direction will be computed automaticaly");
36	}
37	else {
38	return t("You may write mixed Arabic and English freely, line direction will be computed automaticaly");
39	}
40	}
41
42	/**
43	* Implementation of hook_filter()
44	*/
45	function bidi_filter($op, $delta = 0, $format = -1, $text = "") {
46	switch ($op) {
47	case 'list':
48	return array(0 => t('BiDi'));
49	case 'description':
50	return t("Bidi filter");
51	case "process":
52	return _bidi_filter_process($text);
53	default:
54	return $text;
55	}
56	}
57
58	/**
59	* Determine text language (arabic or english).
60	* This function needs to be split into a single include file so it can be used on theme and other places.
61	*
62	* @param $text
63	* The text needs to be.
64	* @return
65	* right2left or left2right depending on the detected language.
66	*/
67	function _get_dir($text) {
68	$text = strtolower($text);
69	$len = strlen($text);
70
71	$arabic = 0;
72	$english = 0;
73
74	// constants
75	$a = ord('a');
76	$z = ord('z');
77
78	// arabic UTF-8 letters have one of these values in the 1st byte
79	$ar1 = 0xd8;
80	$ar2 = 0xd9;
81
82	// this calculates the dominant language in the text, most bidi implementations just use the first letter (which may be more efficient)
83	for ($i = 0; $i < $len; ++$i) {
84	$bin = ord($text[$i]);
85
86	if ($bin == $ar1 \|\| $bin == $ar2) {
87	++$arabic;
88	++$i;
89	}
90	else if ($bin >= $a && $bin <= $z) {
91	++$english;
92	}
93	}
94
95	// should this return a bool instead?
96	if ($english \|\| $arabic) {
97	if ($english > $arabic)
98	return ' dir="LTR" ';
99	else
100	return ' dir="RTL" ';
101	}
102	else {
103	return "";
104	}
105	}
106
107	/**
108	* Sets HTML text line directions (rtl or ltr) as needed.
109	*
110	* @param $text
111	* The HTML text needs to be processed.
112	*
113	* @return
114	* The processed result, with line directions are set using css
115	* classes "righ2left" and "left2right".
116	*/
117	function _bidi_filter_process($text) {
118	$output = "";
119
120	$len = strlen($text);
121	$i = 0;
122
123	while ($i < $len) {
124	// find next tag
125	$e = strpos(substr($text, $i), "<");
126	if ($e === FALSE) {
127	break;
128	}
129
130	$output .= substr($text, $i, $e);
131
132	$i += $e;
133
134	// find end of next tag
135	$e = strcspn(substr($text, $i), "/> ");
136	$tag = substr($text, $i+1, $e-1);
137
138	$output .= substr($text, $i, $e);
139	$i += $e;
140
141	switch ($tag) {
142	// single tag entity
143	case "img":
144	case "br":
145	case "hr":
146	continue;
147	break;
148
149	//nested cases
150	case "ul":
151	case "ol":
152	case "dl":
153	case "div":
154	$e = 0;
155	// stack counter
156	$j = 1;
157
158	// if $j = 0 then stack is empty
159	while ($j > 0) {
160	$e += strpos(substr($text, $i+$e) , $tag);
161
162	// yeah I know I'm assuming certain sizes for tags, sue me
163	// new nested block
164	if ($text[$i+$e-1] == '<') {
165	++$j;
166	}
167	// end of block
168	else if ($text[$i+$e-1] == '/') {
169	--$j;
170	}
171	// skip current tag
172	$e += 2;
173	}
174
175	$content = substr($text, $i+1, $e);
176	$output .= _get_dir(strip_tags($content));
177	$e += strlen($tag);
178	$output.= substr($text, $i, $e);
179	$i += $e;
180	break;
181
182	// block elements these get out attention
183	case "p":
184	case "blockquote":
185	case "pre":
186	case "h1":
187	case "h2":
188	case "h3":
189	case "h4":
190	case "h5":
191	case "h6":
192
193	// nested lists are proken better think of a fix
194	case "li":
195	$close_tag = "</".$tag.">";
196	// breaks on extra whitespaces, maybe I should use regexps instead
197	$e = strpos(substr($text, $i), $close_tag);
198	$content = substr($text, $i+1, $e-1);
199	// calculate directions
200	$output.= _get_dir(strip_tags($content));
201	$e += strlen($close_tag);
202	$output.= substr($text, $i, $e);
203	$i += $e;
204	break;
205
206	// non block tags we are not handeling
207	default:
208	$close_tag = "</".$tag.">";
209	$e = strpos(substr($text, $i), $close_tag) + strlen($close_tag);
210	$output.= substr($text, $i, $e);
211	$i += $e;
212	}
213	}
214
215	// any missing content
216	$output .= substr($text, $i);
217
218	return $output;
219	}
220
221	?>