Simple Chinese word segmentation based on RMM

This program is a simple Chinese word segmentation based on the idea of ​​RMM Chinese word segmentation. There are still many loopholes in the program. I hope God can give me some guidance.... The garbled code problem has been optimized.

/**
 * Based on RMM Chinese word segmentation (reverse matching method)
 * @author tangpan
 * @date 2013-10-12
 * @version 1.0.0
 **/
class SplitWord {
 //public $Tag_dic = array(); //Storage dictionary word segmentation
 public $Rec_dic = array(); //Storage reorganization Word segmentation
 public $Split_char = ' '; //Separator
 public $Source_str = ''; //Storage source string
 public $Result_str = ''; //Storage word segmentation result string
 public $limit_lenght = 2;
 public $Dic_maxLen = 28; //The maximum length of words in the dictionary
 public $Dic_minLen = 2; //The minimum length of words in the dictionary

 public function SplitWord() { //Initialize the object and automatically execute member methods
 $ this->__construct();
 }
 public function __construct() {
 $dic_path = dirname(__FILE__).'/words.csv'; //Preload the dictionary to improve word segmentation speed
 $fp = fopen( $ dic_path, 'r' ); //Read the words in the vocabulary
 while( $line = fgets( $fp, 256 ) ) {
 $ws = explode(' ', $line); //Read the words in the vocabulary Split the words
 $ws[0] = trim(iconv('utf-8','GBK',$ws[0])); //Encoding conversion
 //$this->Tag_dic[$ws[ 0]] = true; //Use word as index, sequence number as value
 $this->Rec_dic[strlen($ws[0])][$ws[0]] = true; //Use word length and word They are the indexes of the two-dimensional array respectively, using n as the value to reorganize the vocabulary
 }
 fclose($fp); //Close the vocabulary
 }

 /**
 * Set the source string
 * @param The string to be segmented
*/
 public function SetSourceStr( $str ) {
 $str = iconv( 'utf-8', 'GBK', $str ); // Convert utf-8 encoded characters to GBK encoding
 $this->Source_str = $this->DealStr( $ str ); //Preliminary processing of string
 }

 /**
 * Check string
 * @param $str source string
 * @return bool
*/
 public function checkStr( $str ) {
 if ( trim($str) == '' ) return; //If string If it is empty, return directly
 if ( ord( $str[0] ) > 0x80 ) return true; //If it is a Chinese character, return true
 else return false; //If it is not a Chinese character, return false
 }


 / **
 * RMM word segmentation algorithm
 * @param $str String to be processed
*/
 public function SplitRMM( $str = '' ) {
 if ( trim( $str ) == '' ) return; //If the string is empty, return directly
 else $this- >SetSourceStr( $str ); //When the string is not empty, set the source string
 if ( $this->Source_str == ' ' ) return; //When the source string is empty, return directly
 $split_words = explode( ' ', $this->Source_str ); //Split the string with spaces
 $lenght = count( $split_words ); //Calculate the length of the array
 for ( $i = $lenght - 1 ; $i >= 0; $i-- ) {
 if ( trim( $split_words[$i] ) == ' ' ) continue; //If the character is empty, skip the following code and go directly to the next One loop
 if ( $this->checkStr( $split_words[$i] ) ) { //Check the string, if it is a Chinese character
 if ( strlen( $split_words[$i] ) >= $this-> ;limit_lenght ) { //When the string length is greater than the limit size
 //Reverse match the string
 $this->Result_str = $this->pregRmmSplit( $split_words[$i] ).$this-> Split_char.$this->Result_str;
 }
 } else {
 $this->Result_str = $split_words[$i].$this->Split_char.$this->Result_str;
 }
 }
 $ this->clear( $split_words ); //Release memory
 return iconv('GBK', 'utf-8', $this->Result_str);
 }

 /**
 * Decompose the Chinese string using reverse matching method
 * @param $str string
 * @return $retStr The string that has been segmented into words
*/
 public function pregRmmSplit( $str ) {
 if ( $str == ' ' ) return;
 $splen = strlen( $str );
 $Split_Result = array();
 for ( $j = $splen - 1; $j >= 0; $j--) { //Match characters in reverse
 if ( $splen <= $this->Dic_minLen ) { //When the character length is greater than the minimum character length in the dictionary
 if ( $j = = 1 ) { //When the length is 1
 $Split_Result[] = substr( $str, 0, 2 );
 }else {
 $w = trim( substr( $str, 0, $this->Dic_minLen + 1 ) ); //Truncate the first four characters
 if ( $this->IsWord( $w ) ) { // Determine whether the character exists in the dictionary
 $Split_Result[] = $w; //If it exists, write it to the array for storage
 } else {
 $Split_Result[] = substr( $str, 2, 2); //Reverse storage
 $Split_Result[] = substr( $str, 0, 2 );
 }
 }
 $j = -1; //Close the loop;
 break;
 }
 if ( $j >= $this->Dic_maxLen ) $max_len = $this->Dic_maxLen; //When the character length is greater than the length of the largest word in the dictionary, assign the maximum limit length
 else $max_len = $j;
 for ( $k = $max_len; $k >= 0; $k = $k - 2 ) { //One jump is one Chinese character
 $w = trim( substr( $str, $j - $k, $k + 1 ) );
 if ( $this-> ;IsWord( $w ) ) {
 $Split_Result[] = $w; //Save the word
 $j = $j - $k - 1; //Move the position to the position of the matched character
 break; // If the word segmentation is successful, it will jump out of the current loop and enter the next loop
 }
 }
 }
 $retStr = $this->resetWord( $Split_Result ); //Reorganize the string and return the processed string
 $this-> ;clear( $Split_Result ); //Release memory
 return $retStr;
 }

 /**
 * Re-identify and combine word segments
 * @param $Split_Result Restructure the target string
 * @return $ret_Str Restructure the string
*/
 public function resetWord( $Split_Result ) {
 if ( trim( $Split_Result[0] ) == '' ) return;
 $Len = count( $Split_Result ) - 1;
 $ret_Str = '';
 $spc = $this->Split_char;
 for ( $i = $Len; $i >= 0 ; $i-- ) {
 if ( trim( $Split_Result[$i] ) != '' ) {
 $Split_Result[$i] = iconv( 'GBK', 'utf-8', $Split_Result[$i ] );
 $ret_Str .= $spc.$Split_Result[$i].' ';
 }
 }
 //$ret_Str = preg_replace('/^'.$spc.'/',',',$ ret_Str);
 $ret_Str = iconv('utf-8','GBK',$ret_Str);
 return $ret_Str;
 }

 /**
 * Check if a word exists in the dictionary
 * @param $okWord The word to check
 * @return bool;
*/
 public function IsWord( $okWord ) {
 $len = strlen( $okWord );
 if ( $len > $this->Dic_maxLen + 1 ) return false;
 else { //According to the two-dimensional array index matching, whether the word exists
 return isset($this ->Rec_dic[$len][$okWord]);
 }

 }

 /**
 * Preliminary string processing (replacing special characters with spaces)
 * @param $str The source string to be processed
 * @return $okStr Returns the preprocessed string
*/
 public function DealStr( $str ) {
 $spc = $this->Split_char; / /Copy separator
 $slen = strlen( $str ); // Calculate the length of the character
 if ( $slen == 0 ) return; // If the character length is 0, return directly
 $okstr = ''; // Initialization variable
 $prechar = 0; //Character judgment variable (0-blank, 1-English, 2-Chinese, 3-symbol)
 for ( $i = 0; $i < $slen; $i++ ) {
 $str_ord = ord( $str[$i] );
 if ( $str_ord < 0x81 ) { //If it is an English character
 if ( $str_ord < 33 ) { // English blank symbol
 if ( $str [$i] != 'r' && $str[$i] != 'n' )
 $okstr .= $spc;
 $prechar = 0;
 continue;
 } else if ( ereg('[@. %#:^&_-]',$str[$i]) ) { //If the character of the keyword is a number or English or special character
 if ( $prechar == 0 ) { //When the character is a blank character
 $okstr .= $str[$i];
 $prechar = 3;
 } else {
 $okstr .= $spc.$str[$i]; //When the character is not a blank character, string in front of the character Top white space character
 $prechar = 3;
 }
 } else if ( ereg('[0-9a-zA-Z]', $str[$i]) ) { //Split English number combinations
 if ( (ereg ('[0-9]',$str[$i-1]) && ereg('[a-zA-Z]',$str[$i]))
 || (ereg('[a-zA -Z]',$str[$i-1]) && ereg('[0-9]',$str[$i])) ) {
 $okstr .= $spc.$str[$i];
 } else {
 $okstr .= $str[$i];
 }
 }
 }else { //If the second character of the keyword is a Chinese character
 if ( $prechar != 0 && $prechar != 2 ) // If the previous character is non-Chinese and non-space, add a space
 $okstr . = $spc;
 if ( isset( $str[$i+1] ) ) { //If it is a Chinese character
 $c = $str[$i].$str[$i+1]; //Convert the two characters strings together to form a Chinese character
 $n = hexdec( bin2hex( $c ) ); //Convert the ascii code to hexadecimal and then to decimal
 if ( $n > 0xA13F && $ n < 0xAA40 ) { //If it is a Chinese punctuation mark
 if ( $prechar != 0 ) $okstr .= $spc; //Replace the Chinese punctuation mark with an empty mark
 //else $okstr .= $spc; // If the previous character is empty, string it directly
 $prechar = 3;
 } else { //If it is not Chinese punctuation
 $okstr .= $c;
 $prechar = 2;
 }
 $i++; // $ i plus 1, even if it moves to one Chinese character at a time
 }
 }
 }
 return $okstr;
 }

 /**
 * Release memory
 * @param $data Temporary data
*/
 public function clear( $data ) {
 unset( $data ) ; //Delete temporary data
 }
}
?>