RMM Segmentation Algorithm Class

RMM Segmentation Algorithm Class RMM segmentation algorithm Class splitword{ var $TagDic = Array (); var $RankDic = Array (); var $SourceStr = '; var $ResultStr = '; var $SplitChar = '; Separator var $SplitLen = 4; Reserved word length var $MaxLen = 7; Dictionary maximum Chinese text, where the value is the largest index of a byte array var $MinLen = 3; Minimum Chinese text, where the value is the maximum index of the byte array function Splitword () { $this->__construct (); } function __construct () { Advanced participle, pre-loaded dictionary to mention the high speed of participle $dicfile = DirName (__file__). " /ppldic.csv "; $fp = fopen ($dicfile, ' r ');//Read words in a thesaurus while ($line = fgets ($FP, 256)) { $ws = Explode (', $line);//split the words in the thesaurus $this->tagdic[$ws [0]] = $ws [1]; $this->rankdic[strlen ($ws [0]) [$WS [0]] = $WS [2]; } Fclose ($FP);//Close the thesaurus file } Analysis and release resources function Clear () { @fclose ($this->quickdic); } Set source string function SetSource ($STR) { $this->sourcestr = $this->updatestr ($STR); $this->resultstr = ""; } Check if the string does not exist in Chinese function NOTGBK ($STR) { if ($str = = "") return ""; if (Ord ($str [0]) >0x80) return false; else return true; } RMM segmentation algorithm function splitrmm ($str = "") { if ($str! = "") $this->setsource ($STR); if ($this->sourcestr== "") return ""; $this->sourcestr = $this->updatestr ($this->sourcestr); $spwords = Explode ("", $this->sourcestr); $spLen = count ($spwords); $SPC = $this->splitchar; for ($i = ($spLen-1); $i >=0; $i-) { if ($spwords [$i]== "") continue; if ($this->notgbk ($spwords [$i]) { if (Preg_match ("/[^0-9\.\+\-]/", $spwords [$i])) {$this->resultstr = $spwords [$i]. $spc. $this->resultstr;} Else { $nextword = ""; @ $nextword = substr ($this->resultstr,0,strpos ($this->resultstr, "")); } } Else { $c = $spwords [$i][0]. $spwords [$i][1]; $n = Hexdec (Bin2Hex ($c)); if (strlen ($spwords [$i]) <= $this->splitlen) { } Else { $this->resultstr = $this->runrmm ($spwords [$i]). $spc. $this->resultstr; } } } return $this->resultstr; } Decomposition of all Chinese strings by inverse matching method function RUNRMM ($STR) { $SPC = $this->splitchar; $spLen = strlen ($STR); $RSSTR = ""; $okWord = ""; $tmpWord = ""; $WordArray = Array (); Inverse dictionary Matching for ($i = ($spLen-1); $i >=0;) { When I reach the smallest possible word if ($i <= $this->minlen) { if ($i ==1) { $WordArray [] = substr ($str, 0,2); }else { $w = substr ($str, 0, $this->minlen+1); if ($this->isword ($w)) { $WordArray [] = $w; }else{ $WordArray [] = substr ($str, 2,2); $WordArray [] = substr ($str, 0,2); } } $i =-1; Break } Analyze the situation when the minimum word is above if ($i >= $this->maxlen) $maxPos = $this->maxlen; else $maxPos = $i; $isMatch = false; for ($j = $maxPos; $j >=0; $j = $j-2) { $w = substr ($str, $i-$j, $j + 1); if ($this->isword ($w)) { $WordArray [] = $w; $i = $i-$j-1; $isMatch = true; Break } } } $RSSTR = $this->otherword ($WordArray); return $rsStr; } function Otherword ($WordArray) { $wlen = count ($WordArray) -1;//counts the number of elements in the array $RSSTR = "";//Initialize variable $SPC = $this->splitchar; for ($i = $wlen; $i >=0; $i--) { $rsStr. = $spc. $WordArray [$i]. ",";//splitting the array into comma } Returns the result of this paragraph participle $RSSTR = Preg_replace ("/^". $spc. " /",", ", $RSSTR); return $rsStr; } Determine if there is a word in the dictionary function Isword ($okWord) { $slen = strlen ($okWord); if ($slen > $this->maxlen) return false; else return isset ($this->rankdic[$slen [$okWord]); } Collation of strings (punctuation, Chinese and English mixed, etc. preliminary processing) function Updatestr ($STR) { $SPC = $this->splitchar; $slen = strlen ($STR); if ($slen ==0) return '; $okstr = "; $prechar = 0; 0-Blank 1-English-Chinese 3-symbol for ($i =0; $i < $slen; $i + +) { if (Ord ($str [$i]) < 0X81) { Blank symbols in English if (Ord ($str [$i]) < 33) { if ($prechar!=0&& $str [$i]!= "\ r" && $str [$i]!= "\ n") $okstr. = $SPC; $prechar = 0; Continue }else if (Preg_match ("/[^0-9a-za-z@\.%#:\\&_-]/", $str [$i])) { if ($prechar ==0) {$okstr. = $str [$i]; $prechar = 3;} else{$okstr. = $spc. $str [$i]; $prechar = 3;} }else{ if ($prechar ==2| | $prechar ==3) {$okstr. = $spc. $str [$i]; $prechar = 1;} Else { if (Preg_match ("/@#%:/", $str [$i])) {$okstr. = $str [$i]; $prechar = 3; } else {$okstr. = $str [$i]; $prechar = 1;} } } } else{ If the previous character is non-Chinese and non-whitespace, add a space if ($prechar!=0 && $prechar!=2) $okstr. = $SPC; If Chinese characters if (Isset ($str [$i +1])) { $c = $str [$i]. $str [$i +1]; $n = Hexdec (Bin2Hex ($c)); if ($n <0xa13f && $n > 0xaa40) { if ($prechar!=0) $okstr. = $spc. $c; else $okstr. = $c; $prechar = 3; } else{ $okstr. = $c; $prechar = 2; } $i + +; } } } return $okstr; } } Call $split =new Splitword (); echo $split->splitrmm ("PHP Search Technology"); Note The format of the Ppldic.csv dictionary is word + space + number +n Copy Code

