class.chinese.php

聪明星企业wap网站系统 Version 1.0

	    $tempcontent = @file($position);

		if (!$tempcontent) {
		    echo "打开文件失败！";
			exit;
		}

		// 将数组的所有内容转换为字符串
		$this->SourceText = implode("",$tempcontent);

		$this->SourceText = eregi_replace( "charset=".$this->config['SourceLang'] , "charset=".$this->config['TargetLang'] , $this->SourceText);


//		ereg(href="css/dir.css"
	} // 结束 OpenFile 函数

	/**
	 * 设置变量的值
	 *
	 * 详细说明
	 * @形参
	 * @起始      1.0
	 * @最后修改  1.0
	 * @访问      公开
	 * @返回值    无
	 * @throws
	 */
	function setvar( $parameter , $value )
	{
		if(!trim($parameter))
			return $parameter;

		$this->config[$parameter] = $value;

	} // 结束 setvar 函数

	/**
	 * 将简体、繁体中文的 UNICODE 编码转换为 UTF8 字符
	 *
	 * 详细说明
	 * @形参      数字 $c 简体中文汉字的UNICODE编码的10进制
	 * @起始      1.1
	 * @最后修改  1.2
	 * @访问      内部
	 * @返回      字符串
	 * @throws    
	 */
	function CHSUtoUTF8($c)
	{
		$str="";

		if ($c < 0x80) {
			$str.=$c;
		}

		else if ($c < 0x800) {
			$str.=(0xC0 | $c>>6);
			$str.=(0x80 | $c & 0x3F);
		}

		else if ($c < 0x10000) {
			$str.=(0xE0 | $c>>12);
			$str.=(0x80 | $c>>6 & 0x3F);
			$str.=(0x80 | $c & 0x3F);
		}

		else if ($c < 0x200000) {
			$str.=(0xF0 | $c>>18);
			$str.=(0x80 | $c>>12 & 0x3F);
			$str.=(0x80 | $c>>6 & 0x3F);
			$str.=(0x80 | $c & 0x3F);
		}

		return $str;
	} // 结束 CHSUtoUTF8 函数
	
	/**
	 * 简体、繁体中文 <-> UTF8 互相转换的函数
	 *
	 * 详细说明
	 * @形参      
	 * @起始      1.1
	 * @最后修改  1.5
	 * @访问      内部
	 * @返回      字符串
	 * @throws    
	 */
	function CHStoUTF8(){

		if ($this->config["SourceLang"]=="BIG5" || $this->config["SourceLang"]=="GB2312") {
			$ret="";

			while($this->SourceText != ""){

				if(ord(substr($this->SourceText,0,1))>127){

					if ($this->config["SourceLang"]=="BIG5") {
						$utf8=$this->CHSUtoUTF8(hexdec($this->unicode_table[hexdec(bin2hex(substr($this->SourceText,0,2)))]));
					}
					if ($this->config["SourceLang"]=="GB2312") {
						$utf8=$this->CHSUtoUTF8(hexdec($this->unicode_table[hexdec(bin2hex(substr($this->SourceText,0,2)))-0x8080]));
					}
					for($i=0;$i<strlen($utf8);$i+=3)
						$ret.=chr(substr($utf8,$i,3));

					$this->SourceText=substr($this->SourceText,2,strlen($this->SourceText));
				}
				
				else{
					$ret.=substr($this->SourceText,0,1);
					$this->SourceText=substr($this->SourceText,1,strlen($this->SourceText));
				}
			}
			$this->unicode_table = array();
			$this->SourceText = "";
			return $ret;
		}

		if ($this->config["SourceLang"]=="UTF8") {
			$out = "";
			$len = strlen($this->SourceText);
			$i = 0;
			while($i < $len) {
				$c = ord( substr( $this->SourceText, $i++, 1 ) );
				switch($c >> 4)
				{ 
					case 0: case 1: case 2: case 3: case 4: case 5: case 6: case 7:
						// 0xxxxxxx
						$out .= substr( $this->SourceText, $i-1, 1 );
					break;
					case 12: case 13:
						// 110x xxxx   10xx xxxx
						$char2 = ord( substr( $this->SourceText, $i++, 1 ) );
						$char3 = $this->unicode_table[(($c & 0x1F) << 6) | ($char2 & 0x3F)];

						if ($this->config["TargetLang"]=="GB2312")
							$out .= $this->_hex2bin( dechex(  $char3 + 0x8080 ) );

						if ($this->config["TargetLang"]=="BIG5")
							$out .= $this->_hex2bin( $char3 );
					break;
					case 14:
						// 1110 xxxx  10xx xxxx  10xx xxxx
						$char2 = ord( substr( $this->SourceText, $i++, 1 ) );
						$char3 = ord( substr( $this->SourceText, $i++, 1 ) );
						$char4 = $this->unicode_table[(($c & 0x0F) << 12) | (($char2 & 0x3F) << 6) | (($char3 & 0x3F) << 0)];

						if ($this->config["TargetLang"]=="GB2312")
							$out .= $this->_hex2bin( dechex ( $char4 + 0x8080 ) );

						if ($this->config["TargetLang"]=="BIG5")
							$out .= $this->_hex2bin( $char4 );
					break;
				}
			}

			// 返回结果
			return $out;
		}
	} // 结束 CHStoUTF8 函数

	/**
	 * 简体、繁体中文转换为 UNICODE编码
	 *
	 * 详细说明
	 * @形参      
	 * @起始      1.2
	 * @最后修改  1.2
	 * @访问      内部
	 * @返回      字符串
	 * @throws    
	 */
	function CHStoUNICODE()
	{

		$utf="";

		while($this->SourceText != "")
		{
			if (ord(substr($this->SourceText,0,1))>127)
			{

				if ($this->config["SourceLang"]=="GB2312")
					$utf.="&#x".$this->unicode_table[hexdec(bin2hex(substr($this->SourceText,0,2)))-0x8080].";";

				if ($this->config["SourceLang"]=="BIG5")
					$utf.="&#x".$this->unicode_table[hexdec(bin2hex(substr($this->SourceText,0,2)))].";";

				$this->SourceText=substr($this->SourceText,2,strlen($this->SourceText));
			}
			else
			{
				$utf.=substr($this->SourceText,0,1);
				$this->SourceText=substr($this->SourceText,1,strlen($this->SourceText));
			}
		}
		return $utf;
	} // 结束 CHStoUNICODE 函数

	/**
	 * 简体中文 <-> 繁体中文 互相转换的函数
	 *
	 * 详细说明
	 * @起始      1.0
	 * @访问      内部
	 * @返回值    经过编码的utf8字符
	 * @throws
	 */
	function GB2312toBIG5()
	{
		// 获取等待转换的字符串的总长度
		$max=strlen($this->SourceText)-1;

		for($i=0;$i<$max;$i++){

			$h=ord($this->SourceText[$i]);

			if($h>=160){

				$l=ord($this->SourceText[$i+1]);

				if($h==161 && $l==64){
					$gb="  ";
				}
				else{
					fseek($this->ctf,($h-160)*510+($l-1)*2);
					$gb=fread($this->ctf,2);
				}

				$this->SourceText[$i]=$gb[0];
				$this->SourceText[$i+1]=$gb[1];
				$i++;
			}
		}
		fclose($this->ctf);

		// 将转换后的结果赋予 $result;
		$result = $this->SourceText;

		// 清空 $thisSourceText
		$this->SourceText = "";

		// 返回转换结果
		return $result;
	} // 结束 GB2312toBIG5 函数

	/**
	 * 根据所得到的编码搜寻拼音
	 *
	 * 详细说明
	 * @起始      1.0
	 * @最后修改  1.0
	 * @访问      内部
	 * @返回值    字符串
	 * @throws
	 */
	function PinYinSearch($num){

		if($num>0&&$num<160){
			return chr($num);
		}

		elseif($num<-20319||$num>-10247){
			return "";
		}

		else{

			for($i=count($this->pinyin_table)-1;$i>=0;$i--){
				if($this->pinyin_table[$i][1]<=$num)
					break;
			}

			return $this->pinyin_table[$i][0];
		}
	} // 结束 PinYinSearch 函数

	/**
	 * 简体、繁体中文 -> 拼音 转换
	 *
	 * 详细说明
	 * @起始      1.0
	 * @最后修改  1.3
	 * @访问      内部
	 * @返回值    字符串，每个拼音用空格分开
	 * @throws
	 */
	function CHStoPinYin(){
		if ( $this->config['SourceLang']=="BIG5" ) {
			$this->ctf = fopen($this->config['codetable_dir'].$this->config['BIG5toGB_table'], "r");
			if (is_null($this->ctf)) {
				echo "打开打开转换表文件失败！";
				exit;
			}

			$this->SourceText = $this->GB2312toBIG5();
			$this->config['TargetLang'] = "PinYin";
		}

		$ret = array();
		$ri = 0;
		for($i=0;$i<strlen($this->SourceText);$i++){

			$p=ord(substr($this->SourceText,$i,1));

			if($p>160){
				$q=ord(substr($this->SourceText,++$i,1));
				$p=$p*256+$q-65536;
			}

			$ret[$ri]=$this->PinYinSearch($p);
			$ri = $ri + 1;
		}

		// 清空 $this->SourceText
		$this->SourceText = "";

		$this->pinyin_table = array();

		// 返回转换后的结果
		return implode(" ", $ret);
	} // 结束 CHStoPinYin 函数

	/**
	 * 输出转换结果
	 *
	 * 详细说明
	 * @形参
	 * @起始      1.0
	 * @最后修改  1.2
	 * @访问      公开
	 * @返回      字符换
	 * @throws
	 */
	function ConvertIT()
	{
		// 判断是否为中文繁、简转换
		if ( ($this->config['SourceLang']=="GB2312" || $this->config['SourceLang']=="BIG5") && ($this->config['TargetLang']=="GB2312" || $this->config['TargetLang']=="BIG5") ) {
			return $this->GB2312toBIG5();
		}

		// 判断是否为简体中文与拼音转换
		if ( ($this->config['SourceLang']=="GB2312" || $this->config['SourceLang']=="BIG5") && $this->config['TargetLang']=="PinYin" ) {
			return $this->CHStoPinYin();
		}

		// 判断是否为简体、繁体中文与UTF8转换
		if ( ($this->config['SourceLang']=="GB2312" || $this->config['SourceLang']=="BIG5" || $this->config['SourceLang']=="UTF8") && ($this->config['TargetLang']=="UTF8" || $this->config['TargetLang']=="GB2312" || $this->config['TargetLang']=="BIG5") ) {
			return $this->CHStoUTF8();
		}

		// 判断是否为简体、繁体中文与UNICODE转换
		if ( ($this->config['SourceLang']=="GB2312" || $this->config['SourceLang']=="BIG5") && $this->config['TargetLang']=="UNICODE" ) {
			return $this->CHStoUNICODE();
		}

	} // 结束 ConvertIT 函数

} // 结束类库

/**
*/

?>