3 * Client for querying zhdaemon
10 var $mHost, $mPort, $mFP, $mConnected;
17 function ZhClient($host, $port) {
20 $this->mConnected
= $this->connect();
24 * Check if connection to zhdaemon is successful
28 function isconnected() {
29 return $this->mConnected
;
33 * Establish conncetion
39 $this->mFP
= fsockopen($this->mHost
, $this->mPort
, $errno, $errstr, 30);
48 * Query the daemon and return the result
52 function query($request) {
53 if(!$this->mConnected
)
56 fwrite($this->mFP
, $request);
58 $result=fgets($this->mFP
, 1024);
60 list($status, $len) = explode(" ", $result);
61 if($status == 'ERROR') {
62 //$len is actually the error code...
63 print "zhdaemon error $len<br />\n";
68 while(!feof($this->mFP
) && $bytesread<$len) {
69 $str= fread($this->mFP
, $len-$bytesread);
70 $bytesread +
= strlen($str);
73 //data should be of length $len. otherwise something is wrong
74 if(strlen($data) != $len)
80 * Convert the input to a different language variant
82 * @param string $text input text
83 * @param string $tolang language variant
84 * @return string the converted text
87 function convert($text, $tolang) {
89 $q = "CONV $tolang $len\n$text";
90 $result = $this->query($q);
97 * Convert the input to all possible variants
99 * @param string $text input text
100 * @return array langcode => converted_string
103 function convertToAllVariants($text) {
104 $len = strlen($text);
105 $q = "CONV ALL $len\n$text";
106 $result = $this->query($q);
109 list($infoline, $data) = explode('|', $result);
110 $info = explode(";", $infoline);
113 foreach($info as $code => $len) {
114 $ret[strtolower($code)] = substr($data, $i, $len);
120 * Perform word segmentation
122 * @param string $text input text
123 * @return string segmented text
126 function segment($text) {
127 $len = strlen($text);
128 $q = "SEG $len\n$text";
129 $result = $this->query($q);
130 if(!$result) {// fallback to character based segmentation
131 $result = ZhClientFake
::segment($text);
137 * Close the connection
148 function ZhClientFake() {
149 global $wgMemc, $wgDBname;
150 $this->zh2TW
= $wgMemc->get($key1 = "$wgDBname:zhConvert:tw");
151 $this->zh2CN
= $wgMemc->get($key2 = "$wgDBname:zhConvert:cn");
152 $this->zh2SG
= $wgMemc->get($key3 = "$wgDBname:zhConvert:sg");
153 $this->zh2HK
= $wgMemc->get($key4 = "$wgDBname:zhConvert:hk");
154 if(empty($this->zh2TW
) ||
empty($this->zh2CN
) ||
empty($this->zh2SG
) ||
empty($this->zh2HK
)) {
155 require("includes/ZhConversion.php");
156 $this->zh2TW
= $zh2TW;
157 $this->zh2CN
= $zh2CN;
158 $this->zh2HK
= $zh2HK;
159 $this->zh2SG
= $zh2SG;
160 $wgMemc->set($key1, $this->zh2TW
);
161 $wgMemc->set($key2, $this->zh2CN
);
162 $wgMemc->set($key3, $this->zh2SG
);
163 $wgMemc->set($key4, $this->zh2HK
);
167 function isconnected() {
176 function zh2tw($text) {
177 return strtr($text, $this->zh2TW
);
185 function zh2cn($text) {
186 return strtr($text, $this->zh2CN
);
194 function zh2sg($text) {
195 return strtr(strtr($text, $this->zh2CN
), $this->zh2SG
);
203 function zh2hk($text) {
204 return strtr(strtr($text, $this->zh2TW
), $this->zh2HK
);
208 * Convert the input to a different language variant
210 * @param string $text input text
211 * @param string $tolang language variant
212 * @return string the converted text
215 function convert($text, $tolang) {
219 $t = $this->zh2cn($text);
222 $t = $this->zh2tw($text);
225 $t = $this->zh2sg($text);
228 $t = $this->zh2hk($text);
236 function convertToAllVariants($text) {
238 $ret['zh-cn'] = $this->zh2cn($text);
239 $ret['zh-tw'] = $this->zh2tw($text);
240 $ret['zh-sg'] = $this->zh2sg($text);
241 $ret['zh-hk'] = $this->zh2hk($text);
246 * Perform "fake" word segmentation, i.e. treating each character as a word
248 * @param string $text input text
249 * @return string segmented text
252 function segment($text) {
253 /* adapted from LanguageZh_cn.stripForSearch()
254 here we will first separate the single characters,
255 and let the caller conver it to hex
257 if( function_exists( 'mb_strtolower' ) ) {
259 "/([\\xc0-\\xff][\\x80-\\xbf]*)/e",
261 mb_strtolower( $text ) );
263 global $wikiLowerChars;
265 "/([\\xc0-\\xff][\\x80-\\xbf]*)/e",
266 "' ' . strtr( \"\$1\", \$wikiLowerChars )",
272 * Close the fake connection