Project

General

Profile

« Previous | Next » 

Revision 869

Added by thorn about 16 years ago

search: added search_lang for use in search-form. Using DE, the search will search ä=ae,... (alternate spelling of german umlauts)
removed undocumented word-boundary search
search_path is now anchored to the beginning of link ("link LIKE '$path%'" instead of "link LIKE '%$path%'")
added key 'ext_charset' to search-extension to query external databases

View differences:

trunk/CHANGELOG
14 14
29-Oct-2008 Thomas Hornik
15 15
# News: total number of news wasn't calculated correctly
16 16
! Captcha: removed IFRAME for Calculation-as-text captcha
17
- removed useless init_utf8funcs()-function
17
- removed useless init_utf8funcs()-function
18
+ search: added search_lang for use in search-form. Using DE, the search will search ä=ae,... (alternate spelling of german umlauts)
19
- search: removed undocumented word-boundary search
20
! search: search_path is now anchored to the beginning of link ("link LIKE '$path%'" instead of "link LIKE '%$path%'")
21
+ search: added key 'ext_charset' to search-extension to query external databases
18 22
25-Oct-2008 Christian Sommer
19 23
!	updated basic templates according the template guidelines
20 24
24-Oct-2008 Dietrich Roland Pehlke
trunk/wb/framework/frontend.functions.php
77 77
	array_walk($arr_string, create_function('&$v,$k','$v = preg_quote($v, \'/\');'));
78 78
	$search_string = implode("|", $arr_string);
79 79
	$string = strtr($search_string, $string_ul_umlauts);
80
	// special-feature: '|' means word-boundary (\b). Searching for 'the|' will find 'the', but not 'thema'.
81
	$string = strtr($string, array('\\|'=>'\b'));
82 80
	// the highlighting
83 81
	// match $string, but not inside <style>...</style>, <script>...</script>, <!--...--> or HTML-Tags
84 82
	// split $string into pieces - "cut away" styles, scripts, comments, HTML-tags and eMail-addresses
trunk/wb/search/search.php
117 117
	}
118 118
}
119 119

  
120
// Get search language
121
$search_lang = '';
122
if(isset($_REQUEST['search_lang'])) {
123
	$search_lang = $_REQUEST['search_lang'];
124
	if(!preg_match('~^[A-Z]{2}$~', $search_lang))
125
		$search_lang = LANGUAGE;
126
} else {
127
	$search_lang = LANGUAGE;
128
}
129

  
120 130
// Get the path to search into. Normally left blank
131
// ATTN: since wb2.7.1 the path is evaluated as SQL: LIKE "/path%" - which will find "/path.php", "/path/info.php", ...; But not "/de/path.php"
132
// Add a '%' in front of each path to get SQL: LIKE "%/path%"
121 133
/* possible values:
122 134
 * - a single path: "/en/" - search only pages whose link contains 'path' ("/en/machinery/bender-x09")
123
 * - a bunch of alternative pathes: "/en/,/machinery/,docs/" - alternatives paths, seperated by comma
124
 * - a bunch of paths to exclude: "-/about,/info,/jp/,/light" - search all, exclude these.
135
 * - a single path not to search into: "-/help" - search all, exclude /help...
136
 * - a bunch of alternative pathes: "/en/,%/machinery/,/docs/" - alternatives paths, seperated by comma
137
 * - a bunch of paths to exclude: "-/about,%/info,/jp/,/light" - search all, exclude these.
125 138
 * These different styles can't be mixed.
126 139
 */
127
$search_path_SQL = "";
128
$search_path = "";
140
// ATTN: in wb2.7.0 "/en/" matched all links with "/en/" somewhere in the link: "/info/en/intro.php", "/en/info.php", ...
141
// since wb2.7.1 "/en/" matches only links _starting_  with "/en/": "/en/intro/info.php"
142
// use "%/en/" (or "%/en/, %/info", ...) to get the old behavior
143
$search_path_SQL = '';
144
$search_path = '';
129 145
if(isset($_REQUEST['search_path'])) {
130
	$search_path = $wb->add_slashes($_REQUEST['search_path']);
131
	if(!preg_match('~^[-a-zA-Z0-9_,/ ]+$~', $search_path))
146
	$search_path = addslashes(htmlspecialchars(strip_tags($wb->strip_slashes($_REQUEST['search_path']))));
147
	if(!preg_match('~^%?[-a-zA-Z0-9_,/ ]+$~', $search_path))
132 148
		$search_path = '';
133 149
	if($search_path != '') {
134
		$search_path_SQL = "AND ( ";
135
		$not = "";
136
		$op = "OR";
150
		$search_path_SQL = 'AND ( ';
151
		$not = '';
152
		$op = 'OR';
137 153
		if($search_path[0] == '-') {
138
			$not = "NOT";
139
			$op = "AND";
154
			$not = 'NOT';
155
			$op = 'AND';
140 156
			$paths = explode(',', substr($search_path, 1) );
141 157
		} else {
142 158
			$paths = explode(',',$search_path);
......
144 160
		$i=0;
145 161
		foreach($paths as $p) {
146 162
			if($i++ > 0) {
147
				$search_path_SQL .= " $op";
163
				$search_path_SQL .= ' $op';
148 164
			}
149
			$search_path_SQL .= " link $not LIKE '%$p%'";			
165
			$search_path_SQL .= " link $not LIKE '".$p."%'";			
150 166
		}
151
		$search_path_SQL .= " )";
167
		$search_path_SQL .= ' )';
152 168
	}
153 169
}
154 170

  
......
167 183
$search_normal_string = '';
168 184
$search_entities_string = ''; // for SQL's LIKE
169 185
$search_display_string = ''; // for displaying
170
$search_url_string = ''; // for $_GET
186
$search_url_string = ''; // for $_GET -- ATTN: unquoted! Will become urldecoded later
171 187
$string = '';
172 188
if(isset($_REQUEST['string'])) {
173
	if($match!='exact') {
189
	if($match!='exact') { // $string will be cleaned below
174 190
		$string=str_replace(',', '', $_REQUEST['string']);
175 191
	} else {
176
		$string=$_REQUEST['string']; // $string will be cleaned below
192
		$string=$_REQUEST['string'];
177 193
	}
178 194
	// redo possible magic quotes
179 195
	$string = $wb->strip_slashes($string);
......
187 203
	$search_entities_string = str_replace('\\\\', '\\\\\\\\', $search_entities_string);
188 204
	// convert string to utf-8
189 205
	$string = entities_to_umlauts($string, 'UTF-8');
190
	// quote ' " and /  -we need quoted / for regex
191 206
	$search_url_string = $string;
192 207
	$string = preg_quote($string);
208
	// quote ' " and /  -we need quoted / for regex
193 209
	$search_normal_string = str_replace(array('\'','"','/'), array('\\\'','\"','\/'), $string);
194 210
}
195 211
// make arrays from the search_..._strings above
212
if($match == 'exact')
213
	$search_url_array[] = $search_url_string;
214
else
196 215
$search_url_array = explode(' ', $search_url_string);
197 216
$search_normal_array = array();
198 217
$search_entities_array = array();
......
220 239
$search_words = array();
221 240
foreach($search_normal_array AS $str) {
222 241
	$str = strtr($str, $string_ul_umlauts);
223
	// special-feature: '|' means word-boundary (\b). Searching for 'the|' will find the, but not thema.
224
	// this doesn't(?) work correctly for unicode-chars: '|test' will work, but '|über' not.
225
	$str = strtr($str, array('\\|'=>'\b'));
226 242
	$search_words[] = $str;
227 243
}
228 244

  
trunk/wb/search/search_convert.php
32 32
	header('Location: ../index.php');
33 33
	exit(0);
34 34
}
35
if(!isset($search_lang)) $search_lang = LANGUAGE;
35 36

  
36 37
//umlauts to '(upper|lower)' for preg_match()
37 38
//this is UTF-8-encoded
38 39
$string_ul_umlauts = array(
40
	"i" => "(?:i|\xc4\xb0|I)",
41
	"I" => "(?:I|\xc4\xb1|i)",
42
	"k" => "(?:k|\xe2\x84\xaa|K)",
43
	"S" => "(?:S|\xc5\xbf|s)",
39 44
	"\xc2\xb5" => "(?:\xc2\xb5|\xce\x9c)",
45
	"\xc3\x9f" => "(?:\xc3\x9f|SS|ss)", // german ß
40 46
	"\xc3\xa0" => "(?:\xc3\xa0|\xc3\x80)",
41 47
	"\xc3\xa1" => "(?:\xc3\xa1|\xc3\x81)",
42 48
	"\xc3\xa2" => "(?:\xc3\xa2|\xc3\x82)",
......
1460 1466
	"\xf0\x90\x90\xa3" => "(?:\xf0\x90\x90\xa3|\xf0\x90\x91\x8b)",
1461 1467
	"\xf0\x90\x90\xa4" => "(?:\xf0\x90\x90\xa4|\xf0\x90\x91\x8c)",
1462 1468
	"\xf0\x90\x90\xa5" => "(?:\xf0\x90\x90\xa5|\xf0\x90\x91\x8d)"
1463
	//"i" => "(?:i|\xc4\xb0|I)",
1464
	//"I" => "(?:I|\xc4\xb1|i)",
1465
	//"k" => "(?:k|\xe2\x84\xaa|K)",
1466
	//"S" => "(?:S|\xc5\xbf|s)"
1467 1469
);
1468 1470

  
1471

  
1472
if($search_lang=='DE') {
1473
	$string_ul_umlauts["\xc3\x9f"] = "(?:\xc3\x9f|SS|ss)"; // german ß
1474
	$string_ul_umlauts["\xc3\xa4"] = "(?:\xc3\xa4|\xc3\x84|ae|Ae)"; // german ä
1475
	$string_ul_umlauts["\xc3\xb6"] = "(?:\xc3\xb6|\xc3\x96|oe|Oe)"; // german ö
1476
	$string_ul_umlauts["\xc3\xbc"] = "(?:\xc3\xbc|\xc3\x9c|ue|Ue)"; // german ü
1477
	$string_ul_umlauts["\xc3\x84"] = "(?:\xc3\x84|\xc3\xa4|Ae|ae)"; // german Ä
1478
	$string_ul_umlauts["\xc3\x96"] = "(?:\xc3\x96|\xc3\xb6|Oe|oe)"; // german Ö
1479
	$string_ul_umlauts["\xc3\x9c"] = "(?:\xc3\x9c|\xc3\xbc|Ue|ue)"; // german Ü
1480
}
1481

  
1469 1482
?>
trunk/wb/search/search_modext.php
101 101
	}
102 102
	$regex='/(?:^|\b|['.$str1.'])([^'.$str1.']{0,200}?'.$word.'[^'.$str2.']{0,200}(?:['.$str2.']|\b|$))/isu';
103 103
	if(version_compare(PHP_VERSION, '4.3.3', '>=') &&
104
	strpos(strtoupper(PHP_OS), 'WIN')!==0) { // this may crash windows server, so skip if on windows
104
	   strpos(strtoupper(PHP_OS), 'WIN')!==0
105
	) { // this may crash windows server, so skip if on windows
105 106
		// jump from match to match, get excerpt, stop if $max_excerpt_num is reached
106 107
		$last_end = 0; $offset = 0;
107 108
		while(preg_match('/'.$word.'/Sisu', $text, $match_array, PREG_OFFSET_CAPTURE, $last_end)) {
......
115 116
					if(count($excerpt_array) >= $max_excerpt_num)
116 117
						break;
117 118
				}
118
			} else { // problem - preg_match failed: can't find a start- or stop-sign
119
			} else { // problem: preg_match failed - can't find a start- or stop-sign
119 120
				$last_end += 201; // jump forward and try again
120 121
			}
121 122
		}
122
	} else { // compatile, but may be very slow with many large pages
123
	} else { // compatible, but may be very slow with large pages
123 124
		if(preg_match_all($regex, $text, $match_array)) {
124 125
			foreach($match_array[1] AS $string) {
125 126
				if(!preg_match('/\b[a-zA-Z0-9._%+-]+@[a-zA-Z0-9.-]+\./', $string)) // skip excerpts with email-addresses
......
173 174
		if($match && is_array($match[0])) {
174 175
			$x=$match[0][1]; // position of first match
175 176
			// is there an anchor nearby?
176
			if(preg_match_all('/<(?:[^>]+id|\s*a[^>]+name)\s*=\s*"(.*)"/SiU', substr($text,0,$x), $match, PREG_OFFSET_CAPTURE)) {
177
			if(preg_match_all('/<(?:[^>]+id|\s*a[^>]+name)\s*=\s*"(.*)"/iU', substr($text,0,$x), $match, PREG_OFFSET_CAPTURE)) {
177 178
				$anchor='';
178 179
				foreach($match[1] AS $array) {
179 180
					if($array[1] > $x) {
......
234 235
	if(!isset($mod_pic_link))           $mod_pic_link = "";
235 236
	if(!isset($mod_no_highlight))       $mod_no_highlight = false;
236 237
	if(!isset($func_enable_flush))      $func_enable_flush = false; // set this in db: wb_search.cfg_enable_flush [READ THE DOC BEFORE]
238
	if(isset($mod_ext_charset) && $mod_ext_charset!='utf-8') $mod_ext_charset = 'utf-8'; // only utf-8 is allowed, yet. For other charset see DOCU
237 239
	if($mod_text == "") // nothing to do
238 240
		{ return false; }
239 241
	if($mod_no_highlight) // no highlighting
......
243 245
	$mod_text = preg_replace('#<(br( /)?|dt|/dd|/?(h[1-6]|tr|table|p|li|ul|pre|code|div|hr))[^>]*>#Si', '.', $mod_text);
244 246
	$mod_text = preg_replace('/\s+/', ' ', $mod_text);
245 247
	$mod_text = preg_replace('/ \./', '.', $mod_text);
248
	if(isset($mod_ext_charset)) { // data from external database may have a different charset
249
		require_once(WB_PATH.'/framework/functions-utf8.php');
250
		$mod_text = charset_to_utf8($mod_text, $mod_ext_charset);
251
	} else {
246 252
	$mod_text = entities_to_umlauts($mod_text, 'UTF-8');
253
	}
247 254
	$anchor_text = $mod_text; // make an copy containing html-tags
248 255
	$mod_text = strip_tags($mod_text);
249 256
	$mod_text = str_replace(array('&gt;','&lt;','&amp;','&quot;','&#39;','&apos;','&nbsp;'), array('>','<','&','"','\'','\'',"\xC2\xA0"), $mod_text);
257
	$mod_text = '.'.trim($mod_text).'.';
250 258
	// Do a fast scan over $mod_text first. This will speedup things a lot.
251 259
	if($func_search_match == 'all') {
252 260
		if(!is_all_matched($mod_text, $func_search_words))

Also available in: Unified diff