/**
* 开始执行分析
* @parem bool optimize 是否对结果进行优化
* @return bool
*/
static private function StartAnalysis($optimize = TRUE) {
if (!self::$isLoadDic) {
self::$LoadDict();
}
self::$simpleResult = self::$finallyResult = array();
self::$sourceString .= chr(0) . chr(32);
$slen = strlen(self::$sourceString);
$sbcArr = array();
$j = 0;
//全角与半角字符对照表
for ($i = 0xFF00; $i < 0xFF5F; $i++) {
$scb = 0x20 + $j;
$j++;
$sbcArr[$i] = $scb;
}
//对字符串进行粗分
$onstr = '';
$lastc = 1; //1 中/韩/日文, 2 英文/数字/符号('.', '@', '#', '+'), 3 ANSI符号 4 纯数字 5 非ANSI符号或不支持字符
$s = 0;
$ansiWordMatch = "[0-9a-z@#%\+\.-]";
$notNumberMatch = "[a-z@#%\+]";
for ($i = 0; $i < $slen; $i++) {
$c = self::$sourceString[$i] . self::$sourceString[++$i];
$cn = hexdec(bin2hex($c));
$cn = isset($sbcArr[$cn]) ? $sbcArr[$cn] : $cn;
//ANSI字符
if ($cn < 0x80) {
if (preg_match('/' . $ansiWordMatch . '/i', chr($cn))) {
if ($lastc != 2 && $onstr != '') {
self::$simpleResult[$s]['w'] = $onstr;
self::$simpleResult[$s]['t'] = $lastc;
self::_deep_analysis($onstr, $lastc, $s, $optimize);
$s++;
$onstr = '';
}
$lastc = 2;
$onstr .= chr(0) . chr($cn);
} else {
if ($onstr != '') {
self::$simpleResult[$s]['w'] = $onstr;
if ($lastc == 2) {
if (!preg_match('/' . $notNumberMatch . '/i', iconv(UCS2, 'utf-8', $onstr)))
$lastc = 4;
}
self::$simpleResult[$s]['t'] = $lastc;
if ($lastc != 4)
self::_deep_analysis($onstr, $lastc, $s, $optimize);
$s++;
}
$onstr = '';
$lastc = 3;
if ($cn < 31) {
continue;
} else {
self::$simpleResult[$s]['w'] = chr(0) . chr($cn);
self::$simpleResult[$s]['t'] = 3;
$s++;
}
}
}
//普通字符
else {
//正常文字
if (($cn > 0x3FFF && $cn < 0x9FA6) || ($cn > 0xF8FF && $cn < 0xFA2D)
|| ($cn > 0xABFF && $cn < 0xD7A4) || ($cn > 0x3040 && $cn < 0x312B)) {
if ($lastc != 1 && $onstr != '') {
self::$simpleResult[$s]['w'] = $onstr;
if ($lastc == 2) {
if (!preg_match('/' . $notNumberMatch . '/i', iconv(UCS2, 'utf-8', $onstr)))
$lastc = 4;
}
self::$simpleResult[$s]['t'] = $lastc;
if ($lastc != 4)
self::_deep_analysis($onstr, $lastc, $s, $optimize);
$s++;
$onstr = '';
}
$lastc = 1;
$onstr .= $c;
}
//特殊符号
else {
if ($onstr != '') {
self::$simpleResult[$s]['w'] = $onstr;
if ($lastc == 2) {
if (!preg_match('/' . $notNumberMatch . '/i', iconv(UCS2, 'utf-8', $onstr)))
$lastc = 4;
}
self::$simpleResult[$s]['t'] = $lastc;
if ($lastc != 4)
self::_deep_analysis($onstr, $lastc, $s, $optimize);
$s++;
}
//检测书名
if ($cn == 0x300A) {
$tmpw = '';
$n = 1;
$isok = FALSE;
$ew = chr(0x30) . chr(0x0B);
while (TRUE) {
if (!isset(self::$sourceString[$i + $n]) && !isset(self::$sourceString[$i + $n + 1]))
break;
$w = self::$sourceString[$i + $n] . self::$sourceString[$i + $n + 1];
if ($w == $ew) {
self::$simpleResult[$s]['w'] = $c;
self::$simpleResult[$s]['t'] = 5;
$s++;
self::$simpleResult[$s]['w'] = $tmpw;
self::$newWords[$tmpw] = 1;
if (!isset(self::$newWords[$tmpw])) {
self::$foundWordStr .= self::$_out_string_encoding($tmpw) . '/nb, ';
self::$SetWordInfos($tmpw, array('c' => 1, 'm' => 'nb'));
}
self::$simpleResult[$s]['t'] = 13;
$s++;
//最大切分模式对书名继续分词
if (self::$differMax) {
self::$simpleResult[$s]['w'] = $tmpw;
self::$simpleResult[$s]['t'] = 21;
self::_deep_analysis($tmpw, $lastc, $s, $optimize);
$s++;
}
self::$simpleResult[$s]['w'] = $ew;
self::$simpleResult[$s]['t'] = 5;
$s++;
$i = $i + $n + 1;
$isok = TRUE;
$onstr = '';
$lastc = 5;
break;
} else {
$n = $n + 2;
$tmpw .= $w;
if (strlen($tmpw) > 60) {
break;
}
}
}//while
if (!$isok) {
self::$simpleResult[$s]['w'] = $c;
self::$simpleResult[$s]['t'] = 5;
$s++;
$onstr = '';
$lastc = 5;
}
continue;
}
$onstr = '';
$lastc = 5;
if ($cn == 0x3000) {
continue;
} else {
self::$simpleResult[$s]['w'] = $c;
self::$simpleResult[$s]['t'] = 5;
$s++;
}
}//2byte symbol
}//end 2byte char
}//end for
//处理分词后的结果
self::_sort_finally_result();
}
