You can not select more than 25 topics
Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
1137 lines
37 KiB
1137 lines
37 KiB
<?php
|
|
/*
|
|
* 居于Unicode编码词典的php分词器
|
|
* 1、只适用于php5，必要函数 iconv
|
|
* 2、本程序是使用RMM逆向匹配算法进行分词的，词库需要特别编译，本类里提供了 MakeDict() 方法
|
|
* 3、简单操作流程： SetSource -> StartAnalysis -> Get***Result
|
|
* 4、对主词典使用特殊格式进行编码, 不需要载入词典到内存操作
|
|
*
|
|
* Copyright IT柏拉图 QQ: 2500875 Email: 2500875#qq.com
|
|
*
|
|
* @version 2.0
|
|
*
|
|
*/
|
|
|
|
//常量定义
|
|
define('_SP_', chr(0xFF).chr(0xFE));
|
|
define('UCS2', 'ucs-2be');
|
|
class phpAnalysis
|
|
{
|
|
|
|
//hash算法选项
|
|
public $mask_value = 0xFFFF;
|
|
|
|
//输入和输出的字符编码（只允许 utf-8、gbk/gb2312/gb18030、big5 三种类型）
|
|
public $sourceCharSet = 'utf-8';
|
|
public $targetCharSet = 'utf-8';
|
|
|
|
//生成的分词结果数据类型 1 为全部， 2为 词典词汇及单个中日韩简繁字符及英文， 3 为词典词汇及英文
|
|
public $resultType = 1;
|
|
|
|
//句子长度小于这个数值时不拆分，notSplitLen = n(个汉字) * 2 + 1
|
|
public $notSplitLen = 5;
|
|
|
|
//把英文单词全部转小写
|
|
public $toLower = false;
|
|
|
|
//使用最大切分模式对二元词进行消岐
|
|
public $differMax = false;
|
|
|
|
//尝试合并单字
|
|
public $unitWord = true;
|
|
|
|
//初始化类时直接加载词典
|
|
public static $loadInit = true;
|
|
|
|
//使用热门词优先模式进行消岐
|
|
public $differFreq = false;
|
|
|
|
//被转换为unicode的源字符串
|
|
private $sourceString = '';
|
|
|
|
//附加词典
|
|
public $addonDic = array();
|
|
public $addonDicFile = 'dict/words_addons.dic';
|
|
|
|
//主词典
|
|
public $dicStr = '';
|
|
public $mainDic = array();
|
|
public $mainDicHand = false;
|
|
public $mainDicInfos = array();
|
|
public $mainDicFile = 'dict/base_dic_full.dic';
|
|
//是否直接载入词典（选是载入速度较慢，但解析较快；选否载入较快，但解析较慢，需要时才会载入特定的词条）
|
|
private $isLoadAll = false;
|
|
|
|
//主词典词语最大长度 x / 2
|
|
private $dicWordMax = 14;
|
|
//粗分后的数组（通常是截取句子等用途）
|
|
private $simpleResult = array();
|
|
//最终结果(用空格分开的词汇列表)
|
|
private $finallyResult = '';
|
|
|
|
//是否已经载入词典
|
|
public $isLoadDic = false;
|
|
//系统识别或合并的新词
|
|
public $newWords = array();
|
|
public $foundWordStr = '';
|
|
//词库载入时间
|
|
public $loadTime = 0;
|
|
|
|
/**
|
|
* 构造函数
|
|
* @param $source_charset
|
|
* @param $target_charset
|
|
* @param $load_alldic
|
|
* @param $source
|
|
*
|
|
* @return void
|
|
*/
|
|
public function __construct($source_charset='utf-8', $target_charset='utf-8', $load_all=true, $source='')
|
|
{
|
|
$this->addonDicFile = dirname(__FILE__).'/'.$this->addonDicFile;
|
|
$this->mainDicFile = dirname(__FILE__).'/'.$this->mainDicFile;
|
|
$this->SetSource( $source, $source_charset, $target_charset );
|
|
$this->isLoadAll = $load_all;
|
|
if(self::$loadInit) $this->LoadDict();
|
|
}
|
|
|
|
/**
|
|
* 析构函数
|
|
*/
|
|
function __destruct()
|
|
{
|
|
if( $this->mainDicHand !== false )
|
|
{
|
|
@fclose( $this->mainDicHand );
|
|
}
|
|
}
|
|
|
|
/**
|
|
* 根据字符串计算key索引
|
|
* @param $key
|
|
* @return short int
|
|
*/
|
|
private function _get_index( $key )
|
|
{
|
|
$l = strlen($key);
|
|
$h = 0x238f13af;
|
|
while ($l--)
|
|
{
|
|
$h += ($h << 5);
|
|
$h ^= ord($key[$l]);
|
|
$h &= 0x7fffffff;
|
|
}
|
|
return ($h % $this->mask_value);
|
|
}
|
|
|
|
/**
|
|
* 从文件获得词
|
|
* @param $key
|
|
* @param $type (类型 word 或 key_groups)
|
|
* @return short int
|
|
*/
|
|
public function GetWordInfos( $key, $type='word' )
|
|
{
|
|
if( !$this->mainDicHand )
|
|
{
|
|
$this->mainDicHand = fopen($this->mainDicFile, 'r');
|
|
}
|
|
$p = 0;
|
|
$keynum = $this->_get_index( $key );
|
|
if( isset($this->mainDicInfos[ $keynum ]) )
|
|
{
|
|
$data = $this->mainDicInfos[ $keynum ];
|
|
}
|
|
else
|
|
{
|
|
//rewind( $this->mainDicHand );
|
|
$move_pos = $keynum * 8;
|
|
fseek($this->mainDicHand, $move_pos, SEEK_SET);
|
|
$dat = fread($this->mainDicHand, 8);
|
|
$arr = unpack('I1s/n1l/n1c', $dat);
|
|
if( $arr['l'] == 0 )
|
|
{
|
|
return false;
|
|
}
|
|
fseek($this->mainDicHand, $arr['s'], SEEK_SET);
|
|
$data = @unserialize(fread($this->mainDicHand, $arr['l']));
|
|
$this->mainDicInfos[ $keynum ] = $data;
|
|
}
|
|
if( !is_array($data) || !isset($data[$key]) )
|
|
{
|
|
return false;
|
|
}
|
|
return ($type=='word' ? $data[$key] : $data);
|
|
}
|
|
|
|
/**
|
|
* 设置源字符串
|
|
* @param $source
|
|
* @param $source_charset
|
|
* @param $target_charset
|
|
*
|
|
* @return bool
|
|
*/
|
|
public function SetSource( $source, $source_charset='utf-8', $target_charset='utf-8' )
|
|
{
|
|
$this->sourceCharSet = strtolower($source_charset);
|
|
$this->targetCharSet = strtolower($target_charset);
|
|
$this->simpleResult = array();
|
|
$this->finallyResult = array();
|
|
$this->finallyIndex = array();
|
|
if( $source != '' )
|
|
{
|
|
$rs = true;
|
|
if( preg_match("/^utf/", $source_charset) ) {
|
|
$this->sourceString = iconv('utf-8', UCS2, $source);
|
|
}
|
|
else if( preg_match("/^gb/", $source_charset) ) {
|
|
$this->sourceString = iconv('utf-8', UCS2, iconv('gb18030', 'utf-8', $source));
|
|
}
|
|
else if( preg_match("/^big/", $source_charset) ) {
|
|
$this->sourceString = iconv('utf-8', UCS2, iconv('big5', 'utf-8', $source));
|
|
}
|
|
else {
|
|
$rs = false;
|
|
}
|
|
}
|
|
else
|
|
{
|
|
$rs = false;
|
|
}
|
|
return $rs;
|
|
}
|
|
|
|
/**
|
|
* 设置结果类型(只在获取finallyResult才有效)
|
|
* @param $rstype 1 为全部， 2去除特殊符号
|
|
*
|
|
* @return void
|
|
*/
|
|
public function SetResultType( $rstype )
|
|
{
|
|
$this->resultType = $rstype;
|
|
}
|
|
|
|
/**
|
|
* 载入词典
|
|
*
|
|
* @return void
|
|
*/
|
|
public function LoadDict( $maindic='' )
|
|
{
|
|
$startt = microtime(true);
|
|
//正常读取文件
|
|
$dicAddon = $this->addonDicFile;
|
|
if($maindic=='' || !file_exists($maindic) )
|
|
{
|
|
$dicWords = $this->mainDicFile ;
|
|
}
|
|
else
|
|
{
|
|
$dicWords = $maindic;
|
|
$this->mainDicFile = $maindic;
|
|
}
|
|
|
|
//加载主词典（只打开）
|
|
$this->mainDicHand = fopen($dicWords, 'r');
|
|
|
|
//载入副词典
|
|
$hw = '';
|
|
$ds = file($dicAddon);
|
|
foreach($ds as $d)
|
|
{
|
|
$d = trim($d);
|
|
if($d=='') continue;
|
|
$estr = substr($d, 1, 1);
|
|
if( $estr==':' ) {
|
|
$hw = substr($d, 0, 1);
|
|
}
|
|
else
|
|
{
|
|
$spstr = _SP_;
|
|
$spstr = iconv(UCS2, 'utf-8', $spstr);
|
|
$ws = explode(',', $d);
|
|
$wall = iconv('utf-8', UCS2, join($spstr, $ws));
|
|
$ws = explode(_SP_, $wall);
|
|
foreach($ws as $estr)
|
|
{
|
|
$this->addonDic[$hw][$estr] = strlen($estr);
|
|
}
|
|
}
|
|
}
|
|
$this->loadTime = microtime(true) - $startt;
|
|
$this->isLoadDic = true;
|
|
}
|
|
|
|
/**
|
|
* 检测某个词是否存在
|
|
*/
|
|
public function IsWord( $word )
|
|
{
|
|
$winfos = $this->GetWordInfos( $word );
|
|
return ($winfos !== false);
|
|
}
|
|
|
|
/**
|
|
* 获得某个词的词性及词频信息
|
|
* @parem $word unicode编码的词
|
|
* @return void
|
|
*/
|
|
public function GetWordProperty($word)
|
|
{
|
|
if( strlen($word)<4 )
|
|
{
|
|
return '/s';
|
|
}
|
|
$infos = $this->GetWordInfos($word);
|
|
return isset($infos[1]) ? "/{$infos[1]}{$infos[0]}" : "/s";
|
|
}
|
|
|
|
/**
|
|
* 指定某词的词性信息（通常是新词）
|
|
* @parem $word unicode编码的词
|
|
* @parem $infos array('c' => 词频, 'm' => 词性);
|
|
* @return void;
|
|
*/
|
|
public function SetWordInfos($word, $infos)
|
|
{
|
|
if( strlen($word)<4 )
|
|
{
|
|
return ;
|
|
}
|
|
if( isset($this->mainDicInfos[$word]) )
|
|
{
|
|
$this->newWords[$word]++;
|
|
$this->mainDicInfos[$word]['c']++;
|
|
}
|
|
else
|
|
{
|
|
$this->newWords[$word] = 1;
|
|
$this->mainDicInfos[$word] = $infos;
|
|
}
|
|
}
|
|
|
|
/**
|
|
* 开始执行分析
|
|
* @parem bool optimize 是否对结果进行优化
|
|
* @return bool
|
|
*/
|
|
public function StartAnalysis($optimize=true)
|
|
{
|
|
if( !$this->isLoadDic )
|
|
{
|
|
$this->LoadDict();
|
|
}
|
|
$this->simpleResult = $this->finallyResult = array();
|
|
$this->sourceString .= chr(0).chr(32);
|
|
$slen = strlen($this->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 = $this->sourceString[$i].$this->sourceString[++$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 != '') {
|
|
$this->simpleResult[$s]['w'] = $onstr;
|
|
$this->simpleResult[$s]['t'] = $lastc;
|
|
$this->_deep_analysis($onstr, $lastc, $s, $optimize);
|
|
$s++;
|
|
$onstr = '';
|
|
}
|
|
$lastc = 2;
|
|
$onstr .= chr(0).chr($cn);
|
|
}
|
|
else
|
|
{
|
|
if( $onstr != '' )
|
|
{
|
|
$this->simpleResult[$s]['w'] = $onstr;
|
|
if( $lastc==2 )
|
|
{
|
|
if( !preg_match('/'.$notNumberMatch.'/i', iconv(UCS2, 'utf-8', $onstr)) ) $lastc = 4;
|
|
}
|
|
$this->simpleResult[$s]['t'] = $lastc;
|
|
if( $lastc != 4 ) $this->_deep_analysis($onstr, $lastc, $s, $optimize);
|
|
$s++;
|
|
}
|
|
$onstr = '';
|
|
$lastc = 3;
|
|
if($cn < 31)
|
|
{
|
|
continue;
|
|
}
|
|
else
|
|
{
|
|
$this->simpleResult[$s]['w'] = chr(0).chr($cn);
|
|
$this->simpleResult[$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 != '')
|
|
{
|
|
$this->simpleResult[$s]['w'] = $onstr;
|
|
if( $lastc==2 )
|
|
{
|
|
if( !preg_match('/'.$notNumberMatch.'/i', iconv(UCS2, 'utf-8', $onstr)) ) $lastc = 4;
|
|
}
|
|
$this->simpleResult[$s]['t'] = $lastc;
|
|
if( $lastc != 4 ) $this->_deep_analysis($onstr, $lastc, $s, $optimize);
|
|
$s++;
|
|
$onstr = '';
|
|
}
|
|
$lastc = 1;
|
|
$onstr .= $c;
|
|
}
|
|
//特殊符号
|
|
else
|
|
{
|
|
if( $onstr != '' )
|
|
{
|
|
$this->simpleResult[$s]['w'] = $onstr;
|
|
if( $lastc==2 )
|
|
{
|
|
if( !preg_match('/'.$notNumberMatch.'/i', iconv(UCS2, 'utf-8', $onstr)) ) $lastc = 4;
|
|
}
|
|
$this->simpleResult[$s]['t'] = $lastc;
|
|
if( $lastc != 4 ) $this->_deep_analysis($onstr, $lastc, $s, $optimize);
|
|
$s++;
|
|
}
|
|
|
|
//检测书名
|
|
if( $cn == 0x300A )
|
|
{
|
|
$tmpw = '';
|
|
$n = 1;
|
|
$isok = false;
|
|
$ew = chr(0x30).chr(0x0B);
|
|
while(true)
|
|
{
|
|
if( !isset($this->sourceString[$i+$n+1]) ) break;
|
|
$w = $this->sourceString[$i+$n].$this->sourceString[$i+$n+1];
|
|
if( $w == $ew )
|
|
{
|
|
$this->simpleResult[$s]['w'] = $c;
|
|
$this->simpleResult[$s]['t'] = 5;
|
|
$s++;
|
|
|
|
$this->simpleResult[$s]['w'] = $tmpw;
|
|
$this->newWords[$tmpw] = 1;
|
|
if( !isset($this->newWords[$tmpw]) )
|
|
{
|
|
$this->foundWordStr .= $this->_out_string_encoding($tmpw).'/nb, ';
|
|
$this->SetWordInfos($tmpw, array('c'=>1, 'm'=>'nb'));
|
|
}
|
|
$this->simpleResult[$s]['t'] = 13;
|
|
|
|
$s++;
|
|
|
|
//最大切分模式对书名继续分词
|
|
if( $this->differMax )
|
|
{
|
|
$this->simpleResult[$s]['w'] = $tmpw;
|
|
$this->simpleResult[$s]['t'] = 21;
|
|
$this->_deep_analysis($tmpw, $lastc, $s, $optimize);
|
|
$s++;
|
|
}
|
|
|
|
$this->simpleResult[$s]['w'] = $ew;
|
|
$this->simpleResult[$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 )
|
|
{
|
|
$this->simpleResult[$s]['w'] = $c;
|
|
$this->simpleResult[$s]['t'] = 5;
|
|
$s++;
|
|
$onstr = '';
|
|
$lastc = 5;
|
|
}
|
|
continue;
|
|
}
|
|
|
|
$onstr = '';
|
|
$lastc = 5;
|
|
if( $cn==0x3000 )
|
|
{
|
|
continue;
|
|
}
|
|
else
|
|
{
|
|
$this->simpleResult[$s]['w'] = $c;
|
|
$this->simpleResult[$s]['t'] = 5;
|
|
$s++;
|
|
}
|
|
}//2byte symbol
|
|
|
|
}//end 2byte char
|
|
|
|
}//end for
|
|
|
|
//处理分词后的结果
|
|
$this->_sort_finally_result();
|
|
}
|
|
|
|
/**
|
|
* 深入分词
|
|
* @parem $str
|
|
* @parem $ctype (2 英文类， 3 中/韩/日文类)
|
|
* @parem $spos 当前粗分结果游标
|
|
* @return bool
|
|
*/
|
|
private function _deep_analysis( &$str, $ctype, $spos, $optimize=true )
|
|
{
|
|
|
|
//中文句子
|
|
if( $ctype==1 )
|
|
{
|
|
$slen = strlen($str);
|
|
//小于系统配置分词要求长度的句子
|
|
if( $slen < $this->notSplitLen )
|
|
{
|
|
$tmpstr = '';
|
|
$lastType = 0;
|
|
if( $spos > 0 ) $lastType = $this->simpleResult[$spos-1]['t'];
|
|
if($slen < 5)
|
|
{
|
|
//echo iconv(UCS2, 'utf-8', $str).'<br/>';
|
|
if( $lastType==4 && ( isset($this->addonDic['u'][$str]) || isset($this->addonDic['u'][substr($str, 0, 2)]) ) )
|
|
{
|
|
$str2 = '';
|
|
if( !isset($this->addonDic['u'][$str]) && isset($this->addonDic['s'][substr($str, 2, 2)]) )
|
|
{
|
|
$str2 = substr($str, 2, 2);
|
|
$str = substr($str, 0, 2);
|
|
}
|
|
$ww = $this->simpleResult[$spos - 1]['w'].$str;
|
|
$this->simpleResult[$spos - 1]['w'] = $ww;
|
|
$this->simpleResult[$spos - 1]['t'] = 4;
|
|
if( !isset($this->newWords[$this->simpleResult[$spos - 1]['w']]) )
|
|
{
|
|
$this->foundWordStr .= $this->_out_string_encoding( $ww ).'/mu, ';
|
|
$this->SetWordInfos($ww, array('c'=>1, 'm'=>'mu'));
|
|
}
|
|
$this->simpleResult[$spos]['w'] = '';
|
|
if( $str2 != '' )
|
|
{
|
|
$this->finallyResult[$spos-1][] = $ww;
|
|
$this->finallyResult[$spos-1][] = $str2;
|
|
}
|
|
}
|
|
else {
|
|
$this->finallyResult[$spos][] = $str;
|
|
}
|
|
}
|
|
else
|
|
{
|
|
$this->_deep_analysis_cn( $str, $ctype, $spos, $slen, $optimize );
|
|
}
|
|
}
|
|
//正常长度的句子，循环进行分词处理
|
|
else
|
|
{
|
|
$this->_deep_analysis_cn( $str, $ctype, $spos, $slen, $optimize );
|
|
}
|
|
}
|
|
//英文句子，转为小写
|
|
else
|
|
{
|
|
if( $this->toLower ) {
|
|
$this->finallyResult[$spos][] = strtolower($str);
|
|
}
|
|
else {
|
|
$this->finallyResult[$spos][] = $str;
|
|
}
|
|
}
|
|
}
|
|
|
|
/**
|
|
* 中文的深入分词
|
|
* @parem $str
|
|
* @return void
|
|
*/
|
|
private function _deep_analysis_cn( &$str, $lastec, $spos, $slen, $optimize=true )
|
|
{
|
|
$quote1 = chr(0x20).chr(0x1C);
|
|
$tmparr = array();
|
|
$hasw = 0;
|
|
//如果前一个词为 “ ， 并且字符串小于3个字符当成一个词处理。
|
|
if( $spos > 0 && $slen < 11 && $this->simpleResult[$spos-1]['w']==$quote1 )
|
|
{
|
|
$tmparr[] = $str;
|
|
if( !isset($this->newWords[$str]) )
|
|
{
|
|
$this->foundWordStr .= $this->_out_string_encoding($str).'/nq, ';
|
|
$this->SetWordInfos($str, array('c'=>1, 'm'=>'nq'));
|
|
}
|
|
if( !$this->differMax )
|
|
{
|
|
$this->finallyResult[$spos][] = $str;
|
|
return ;
|
|
}
|
|
}
|
|
//进行切分
|
|
for($i=$slen-1; $i > 0; $i -= 2)
|
|
{
|
|
//单个词
|
|
$nc = $str[$i-1].$str[$i];
|
|
//是否已经到最后两个字
|
|
if( $i <= 2 )
|
|
{
|
|
$tmparr[] = $nc;
|
|
$i = 0;
|
|
break;
|
|
}
|
|
$isok = false;
|
|
$i = $i + 1;
|
|
for($k=$this->dicWordMax; $k>1; $k=$k-2)
|
|
{
|
|
if($i < $k) continue;
|
|
$w = substr($str, $i-$k, $k);
|
|
if( strlen($w) <= 2 )
|
|
{
|
|
$i = $i - 1;
|
|
break;
|
|
}
|
|
if( $this->IsWord( $w ) )
|
|
{
|
|
$tmparr[] = $w;
|
|
$i = $i - $k + 1;
|
|
$isok = true;
|
|
break;
|
|
}
|
|
}
|
|
//echo '<hr />';
|
|
//没适合词
|
|
if(!$isok) $tmparr[] = $nc;
|
|
}
|
|
$wcount = count($tmparr);
|
|
if( $wcount==0 ) return ;
|
|
$this->finallyResult[$spos] = array_reverse($tmparr);
|
|
//优化结果(岐义处理、新词、数词、人名识别等)
|
|
if( $optimize )
|
|
{
|
|
$this->_optimize_result( $this->finallyResult[$spos], $spos );
|
|
}
|
|
}
|
|
|
|
/**
|
|
* 对最终分词结果进行优化（把simpleresult结果合并，并尝试新词识别、数词合并等）
|
|
* @parem $optimize 是否优化合并的结果
|
|
* @return bool
|
|
*/
|
|
//t = 1 中/韩/日文, 2 英文/数字/符号('.', '@', '#', '+'), 3 ANSI符号 4 纯数字 5 非ANSI符号或不支持字符
|
|
private function _optimize_result( &$smarr, $spos )
|
|
{
|
|
$newarr = array();
|
|
$prePos = $spos - 1;
|
|
$arlen = count($smarr);
|
|
$i = $j = 0;
|
|
//检测数量词
|
|
if( $prePos > -1 && !isset($this->finallyResult[$prePos]) )
|
|
{
|
|
$lastw = $this->simpleResult[$prePos]['w'];
|
|
$lastt = $this->simpleResult[$prePos]['t'];
|
|
if( ($lastt==4 || isset( $this->addonDic['c'][$lastw] )) && isset( $this->addonDic['u'][$smarr[0]] ) )
|
|
{
|
|
$this->simpleResult[$prePos]['w'] = $lastw.$smarr[0];
|
|
$this->simpleResult[$prePos]['t'] = 4;
|
|
if( !isset($this->newWords[ $this->simpleResult[$prePos]['w'] ]) )
|
|
{
|
|
$this->foundWordStr .= $this->_out_string_encoding( $this->simpleResult[$prePos]['w'] ).'/mu, ';
|
|
$this->SetWordInfos($this->simpleResult[$prePos]['w'], array('c'=>1, 'm'=>'mu'));
|
|
}
|
|
$smarr[0] = '';
|
|
$i++;
|
|
}
|
|
}
|
|
for(; $i < $arlen; $i++)
|
|
{
|
|
|
|
if( !isset( $smarr[$i+1] ) )
|
|
{
|
|
$newarr[$j] = $smarr[$i];
|
|
break;
|
|
}
|
|
$cw = $smarr[$i];
|
|
$nw = $smarr[$i+1];
|
|
$ischeck = false;
|
|
//检测数量词
|
|
if( isset( $this->addonDic['c'][$cw] ) && isset( $this->addonDic['u'][$nw] ) )
|
|
{
|
|
//最大切分时保留合并前的词
|
|
if($this->differMax)
|
|
{
|
|
$newarr[$j] = chr(0).chr(0x28);
|
|
$j++;
|
|
$newarr[$j] = $cw;
|
|
$j++;
|
|
$newarr[$j] = $nw;
|
|
$j++;
|
|
$newarr[$j] = chr(0).chr(0x29);
|
|
$j++;
|
|
}
|
|
$newarr[$j] = $cw.$nw;
|
|
if( !isset($this->newWords[$newarr[$j]]) )
|
|
{
|
|
$this->foundWordStr .= $this->_out_string_encoding( $newarr[$j] ).'/mu, ';
|
|
$this->SetWordInfos($newarr[$j], array('c'=>1, 'm'=>'mu'));
|
|
}
|
|
$j++; $i++; $ischeck = true;
|
|
}
|
|
//检测前导词(通常是姓)
|
|
else if( isset( $this->addonDic['n'][ $smarr[$i] ] ) )
|
|
{
|
|
$is_rs = false;
|
|
//词语是副词或介词或频率很高的词不作为人名
|
|
if( strlen($nw)==4 )
|
|
{
|
|
$winfos = $this->GetWordInfos($nw);
|
|
if(isset($winfos['m']) && ($winfos['m']=='r' || $winfos['m']=='c' || $winfos['c']>500) )
|
|
{
|
|
$is_rs = true;
|
|
}
|
|
}
|
|
if( !isset($this->addonDic['s'][$nw]) && strlen($nw)<5 && !$is_rs )
|
|
{
|
|
$newarr[$j] = $cw.$nw;
|
|
//echo iconv(UCS2, 'utf-8', $newarr[$j])."<br />";
|
|
//尝试检测第三个词
|
|
if( strlen($nw)==2 && isset($smarr[$i+2]) && strlen($smarr[$i+2])==2 && !isset( $this->addonDic['s'][$smarr[$i+2]] ) )
|
|
{
|
|
$newarr[$j] .= $smarr[$i+2];
|
|
$i++;
|
|
}
|
|
if( !isset($this->newWords[$newarr[$j]]) )
|
|
{
|
|
$this->SetWordInfos($newarr[$j], array('c'=>1, 'm'=>'nr'));
|
|
$this->foundWordStr .= $this->_out_string_encoding($newarr[$j]).'/nr, ';
|
|
}
|
|
//为了防止错误，保留合并前的姓名
|
|
if(strlen($nw)==4)
|
|
{
|
|
$j++;
|
|
$newarr[$j] = chr(0).chr(0x28);
|
|
$j++;
|
|
$newarr[$j] = $cw;
|
|
$j++;
|
|
$newarr[$j] = $nw;
|
|
$j++;
|
|
$newarr[$j] = chr(0).chr(0x29);
|
|
}
|
|
|
|
$j++; $i++; $ischeck = true;
|
|
}
|
|
}
|
|
//检测后缀词(地名等)
|
|
else if( isset($this->addonDic['a'][$nw]) )
|
|
{
|
|
$is_rs = false;
|
|
//词语是副词或介词不作为前缀
|
|
if( strlen($cw)>2 )
|
|
{
|
|
$winfos = $this->GetWordInfos($cw);
|
|
if(isset($winfos['m']) && ($winfos['m']=='a' || $winfos['m']=='r' || $winfos['m']=='c' || $winfos['c']>500) )
|
|
{
|
|
$is_rs = true;
|
|
}
|
|
}
|
|
if( !isset($this->addonDic['s'][$cw]) && !$is_rs )
|
|
{
|
|
$newarr[$j] = $cw.$nw;
|
|
if( !isset($this->newWords[$newarr[$j]]) )
|
|
{
|
|
$this->foundWordStr .= $this->_out_string_encoding($newarr[$j]).'/na, ';
|
|
$this->SetWordInfos($newarr[$j], array('c'=>1, 'm'=>'na'));
|
|
}
|
|
$i++; $j++; $ischeck = true;
|
|
}
|
|
}
|
|
//新词识别（暂无规则）
|
|
else if($this->unitWord)
|
|
{
|
|
if(strlen($cw)==2 && strlen($nw)==2
|
|
&& !isset($this->addonDic['s'][$cw]) && !isset($this->addonDic['t'][$cw]) && !isset($this->addonDic['a'][$cw])
|
|
&& !isset($this->addonDic['s'][$nw]) && !isset($this->addonDic['c'][$nw]))
|
|
{
|
|
$newarr[$j] = $cw.$nw;
|
|
//尝试检测第三个词
|
|
if( isset($smarr[$i+2]) && strlen($smarr[$i+2])==2 && (isset( $this->addonDic['a'][$smarr[$i+2]] ) || isset( $this->addonDic['u'][$smarr[$i+2]] )) )
|
|
{
|
|
$newarr[$j] .= $smarr[$i+2];
|
|
$i++;
|
|
}
|
|
if( !isset($this->newWords[$newarr[$j]]) )
|
|
{
|
|
$this->foundWordStr .= $this->_out_string_encoding($newarr[$j]).'/ms, ';
|
|
$this->SetWordInfos($newarr[$j], array('c'=>1, 'm'=>'ms'));
|
|
}
|
|
$i++; $j++; $ischeck = true;
|
|
}
|
|
}
|
|
|
|
//不符合规则
|
|
if( !$ischeck )
|
|
{
|
|
$newarr[$j] = $cw;
|
|
//二元消岐处理——最大切分模式
|
|
if( $this->differMax && !isset($this->addonDic['s'][$cw]) && strlen($cw) < 5 && strlen($nw) < 7)
|
|
{
|
|
$slen = strlen($nw);
|
|
$hasDiff = false;
|
|
for($y=2; $y <= $slen-2; $y=$y+2)
|
|
{
|
|
$nhead = substr($nw, $y-2, 2);
|
|
$nfont = $cw.substr($nw, 0, $y-2);
|
|
if( $this->IsWord( $nfont.$nhead ) )
|
|
{
|
|
if( strlen($cw) > 2 ) $j++;
|
|
$hasDiff = true;
|
|
$newarr[$j] = $nfont.$nhead;
|
|
}
|
|
}
|
|
}
|
|
$j++;
|
|
}
|
|
|
|
}//end for
|
|
$smarr = $newarr;
|
|
}
|
|
|
|
/**
|
|
* 转换最终分词结果到 finallyResult 数组
|
|
* @return void
|
|
*/
|
|
private function _sort_finally_result()
|
|
{
|
|
$newarr = array();
|
|
$i = 0;
|
|
foreach($this->simpleResult as $k=>$v)
|
|
{
|
|
if( empty($v['w']) ) continue;
|
|
if( isset($this->finallyResult[$k]) && count($this->finallyResult[$k]) > 0 )
|
|
{
|
|
foreach($this->finallyResult[$k] as $w)
|
|
{
|
|
if(!empty($w))
|
|
{
|
|
$newarr[$i]['w'] = $w;
|
|
$newarr[$i]['t'] = 20;
|
|
$i++;
|
|
}
|
|
}
|
|
}
|
|
else if($v['t'] != 21)
|
|
{
|
|
$newarr[$i]['w'] = $v['w'];
|
|
$newarr[$i]['t'] = $v['t'];
|
|
$i++;
|
|
}
|
|
}
|
|
$this->finallyResult = $newarr;
|
|
$newarr = '';
|
|
}
|
|
|
|
/**
|
|
* 把uncode字符串转换为输出字符串
|
|
* @parem str
|
|
* return string
|
|
*/
|
|
private function _out_string_encoding( &$str )
|
|
{
|
|
$rsc = $this->_source_result_charset();
|
|
if( $rsc==1 ) {
|
|
$rsstr = iconv(UCS2, 'utf-8', $str);
|
|
}
|
|
else if( $rsc==2 ) {
|
|
$rsstr = iconv('utf-8', 'gb18030', iconv(UCS2, 'utf-8', $str) );
|
|
}
|
|
else{
|
|
$rsstr = iconv('utf-8', 'big5', iconv(UCS2, 'utf-8', $str) );
|
|
}
|
|
return $rsstr;
|
|
}
|
|
|
|
/**
|
|
* 获取最终结果字符串（用空格分开后的分词结果）
|
|
* @return string
|
|
*/
|
|
public function GetFinallyResult($spword=' ', $word_meanings=false)
|
|
{
|
|
$rsstr = '';
|
|
foreach($this->finallyResult as $v)
|
|
{
|
|
if( $this->resultType==2 && ($v['t']==3 || $v['t']==5) )
|
|
{
|
|
continue;
|
|
}
|
|
$m = '';
|
|
if( $word_meanings )
|
|
{
|
|
$m = $this->GetWordProperty($v['w']);
|
|
}
|
|
$w = $this->_out_string_encoding($v['w']);
|
|
if( $w != ' ' )
|
|
{
|
|
if($word_meanings) {
|
|
$rsstr .= $spword.$w.$m;
|
|
}
|
|
else {
|
|
$rsstr .= $spword.$w;
|
|
}
|
|
}
|
|
}
|
|
return $rsstr;
|
|
}
|
|
|
|
/**
|
|
* 获取粗分结果，不包含粗分属性
|
|
* @return array()
|
|
*/
|
|
public function GetSimpleResult()
|
|
{
|
|
$rearr = array();
|
|
foreach($this->simpleResult as $k=>$v)
|
|
{
|
|
if( empty($v['w']) ) continue;
|
|
$w = $this->_out_string_encoding($v['w']);
|
|
if( $w != ' ' ) $rearr[] = $w;
|
|
}
|
|
return $rearr;
|
|
}
|
|
|
|
/**
|
|
* 获取粗分结果，包含粗分属性（1中文词句、2 ANSI词汇（包括全角），3 ANSI标点符号（包括全角），4数字（包括全角），5 中文标点或无法识别字符）
|
|
* @return array()
|
|
*/
|
|
public function GetSimpleResultAll()
|
|
{
|
|
$rearr = array();
|
|
foreach($this->simpleResult as $k=>$v)
|
|
{
|
|
$w = $this->_out_string_encoding($v['w']);
|
|
if( $w != ' ' )
|
|
{
|
|
$rearr[$k]['w'] = $w;
|
|
$rearr[$k]['t'] = $v['t'];
|
|
}
|
|
}
|
|
return $rearr;
|
|
}
|
|
|
|
/**
|
|
* 获取索引hash数组
|
|
* @return array('word'=>count,...)
|
|
*/
|
|
public function GetFinallyIndex()
|
|
{
|
|
$rearr = array();
|
|
foreach($this->finallyResult as $v)
|
|
{
|
|
if( $this->resultType==2 && ($v['t']==3 || $v['t']==5) )
|
|
{
|
|
continue;
|
|
}
|
|
$w = $this->_out_string_encoding($v['w']);
|
|
if( $w == ' ' )
|
|
{
|
|
continue;
|
|
}
|
|
if( isset($rearr[$w]) )
|
|
{
|
|
$rearr[$w]++;
|
|
}
|
|
else
|
|
{
|
|
$rearr[$w] = 1;
|
|
}
|
|
}
|
|
arsort( $rearr );
|
|
return $rearr;
|
|
}
|
|
|
|
/**
|
|
* 获取最终关键字(返回用 "," 间隔的关键字)
|
|
* @return string
|
|
*/
|
|
public function GetFinallyKeywords( $num = 10 )
|
|
{
|
|
$n = 0;
|
|
$arr = $this->GetFinallyIndex();
|
|
$okstr = '';
|
|
foreach( $arr as $k => $v )
|
|
{
|
|
//排除长度为1的词
|
|
if( strlen($k)==1 ) {
|
|
continue;
|
|
}
|
|
//排除长度为2的非英文词
|
|
elseif( strlen($k)==2 && preg_match('/[^0-9a-zA-Z]/', $k) ) {
|
|
continue;
|
|
|
|
}
|
|
//排除单个中文字
|
|
elseif( strlen($k) < 4 && !preg_match('/[a-zA-Z]/', $k)) {
|
|
continue;
|
|
}
|
|
$okstr .= ($okstr=='' ? $k : ','.$k);
|
|
$n++;
|
|
if( $n > $num ) break;
|
|
}
|
|
return $okstr;
|
|
}
|
|
|
|
/**
|
|
* 获得保存目标编码
|
|
* @return int
|
|
*/
|
|
private function _source_result_charset()
|
|
{
|
|
if( preg_match("/^utf/", $this->targetCharSet) ) {
|
|
$rs = 1;
|
|
}
|
|
else if( preg_match("/^gb/", $this->targetCharSet) ) {
|
|
$rs = 2;
|
|
}
|
|
else if( preg_match("/^big/", $this->targetCharSet) ) {
|
|
$rs = 3;
|
|
}
|
|
else {
|
|
$rs = 4;
|
|
}
|
|
return $rs;
|
|
}
|
|
|
|
/**
|
|
* 编译词典
|
|
* @parem $sourcefile utf-8编码的文本词典数据文件<参见范例dict/not-build/base_dic_full.txt>
|
|
* 注意, 需要PHP开放足够的内存才能完成操作
|
|
* @return void
|
|
*/
|
|
public function MakeDict( $source_file, $target_file='' )
|
|
{
|
|
$target_file = ($target_file=='' ? $this->mainDicFile : $target_file);
|
|
$allk = array();
|
|
$fp = fopen($source_file, 'r');
|
|
while( $line = fgets($fp, 512) )
|
|
{
|
|
if( $line[0]=='@' ) continue;
|
|
list($w, $r, $a) = explode(',', $line);
|
|
$a = trim( $a );
|
|
$w = iconv('utf-8', UCS2, $w);
|
|
$k = $this->_get_index( $w );
|
|
if( isset($allk[ $k ]) )
|
|
$allk[ $k ][ $w ] = array($r, $a);
|
|
else
|
|
$allk[ $k ][ $w ] = array($r, $a);
|
|
}
|
|
fclose( $fp );
|
|
$fp = fopen($target_file, 'w');
|
|
$heade_rarr = array();
|
|
$alldat = '';
|
|
$start_pos = $this->mask_value * 8;
|
|
foreach( $allk as $k => $v )
|
|
{
|
|
$dat = serialize( $v );
|
|
$dlen = strlen($dat);
|
|
$alldat .= $dat;
|
|
|
|
$heade_rarr[ $k ][0] = $start_pos;
|
|
$heade_rarr[ $k ][1] = $dlen;
|
|
$heade_rarr[ $k ][2] = count( $v );
|
|
|
|
$start_pos += $dlen;
|
|
}
|
|
unset( $allk );
|
|
for($i=0; $i < $this->mask_value; $i++)
|
|
{
|
|
if( !isset($heade_rarr[$i]) )
|
|
{
|
|
$heade_rarr[$i] = array(0, 0, 0);
|
|
}
|
|
fwrite($fp, pack("Inn", $heade_rarr[$i][0], $heade_rarr[$i][1], $heade_rarr[$i][2]));
|
|
}
|
|
fwrite( $fp, $alldat);
|
|
fclose( $fp );
|
|
}
|
|
|
|
/**
|
|
* 导出词典的词条
|
|
* @parem $targetfile 保存位置
|
|
* @return void
|
|
*/
|
|
public function ExportDict( $targetfile )
|
|
{
|
|
if( !$this->mainDicHand )
|
|
{
|
|
$this->mainDicHand = fopen($this->mainDicFile, 'r');
|
|
}
|
|
$fp = fopen($targetfile, 'w');
|
|
for($i=0; $i <= $this->mask_value; $i++)
|
|
{
|
|
$move_pos = $i * 8;
|
|
fseek($this->mainDicHand, $move_pos, SEEK_SET);
|
|
$dat = fread($this->mainDicHand, 8);
|
|
$arr = unpack('I1s/n1l/n1c', $dat);
|
|
if( $arr['l'] == 0 )
|
|
{
|
|
continue;
|
|
}
|
|
fseek($this->mainDicHand, $arr['s'], SEEK_SET);
|
|
$data = @unserialize(fread($this->mainDicHand, $arr['l']));
|
|
if( !is_array($data) ) continue;
|
|
foreach($data as $k => $v)
|
|
{
|
|
$w = iconv(UCS2, 'utf-8', $k);
|
|
fwrite($fp, "{$w},{$v[0]},{$v[1]}\n");
|
|
}
|
|
}
|
|
fclose( $fp );
|
|
return true;
|
|
}
|
|
}
|
|
|
|
?>
|