开通博客赚积分
发布资源赚积分
分类

源码开发语言/平台
当前位置：首页 > 源码/资料 > 多国语言处理 > 查看源码
SegGraph.cpp
资源名称：ICTCLASV1.2.rar　[点击查看]
上传用户：sunyong76
上传日期：2021-10-03
资源大小：2236k
文件大小：17k
源码类别：

多国语言处理

开发平台：

Java

SegGraph.cpp：源码内容
  1. //////////////////////////////////////////////////////////////////////
  2. //ICTCLAS简介：计算所汉语词法分析系统ICTCLAS(Institute of Computing Technology, Chinese Lexical Analysis System)，
  3. //             功能有：中文分词；词性标注；未登录词识别。
  4. //             分词正确率高达97.58%(973专家评测结果)，
  5. //             未登录词识别召回率均高于90%，其中中国人名的识别召回率接近98%;
  6. //             处理速度为31.5Kbytes/s。
  7. //著作权：  Copyright?2002-2005中科院计算所 职务著作权人：张华平 刘群
  8. //遵循协议：自然语言处理开放资源许可证1.0
  9. //Email: zhanghp@software.ict.ac.cn
  10. //Homepage:www.nlp.org.cn;mtgroup.ict.ac.cn
  11. /****************************************************************************
  12.  *
  13.  * Copyright (c) 2000, 2001 
  14.  *     Machine Group
  15.  *     Software Research Lab.
  16.  *     Institute of Computing Tech.
  17.  *     Chinese Academy of Sciences
  18.  *     All rights reserved.
  19.  *
  20.  * This file is the confidential and proprietary property of 
  21.  * Institute of Computing Tech. and the posession or use of this file requires 
  22.  * a written license from the author.
  23.  * Filename: SegGraph.cpp
  24.  * Abstract:
  25.  *            implement for the Word Segmentation Directed Graph.
  26.  *
  27.  * Author:   Kevin Zhang 
  28.  *          (zhanghp@software.ict.ac.cn)
  29.  * Date:     2002-1-8
  30.  *
  31.  * Notes:
  32.  *                
  33.  * 
  34.  ****************************************************************************/
  35. // SegGraph.cpp: implementation of the CSegGraph class.
  36. //
  37. //////////////////////////////////////////////////////////////////////
  39. #include "stdafx.h"
  40. #include "SegGraph.h"
  41. #include "..\Utility\Utility.h"
  42. #include <string.h>
  43. #include <math.h>
  45. //////////////////////////////////////////////////////////////////////
  46. // Construction/Destruction
  47. //////////////////////////////////////////////////////////////////////
  49. CSegGraph::CSegGraph()
  50. {
  51. m_segGraph.SetRowFirst();
  52. //segGraph: The segmentation word graph
  53. //Row first array
  54. }
  55. CSegGraph::~CSegGraph()
  56. {
  57. }
  58. bool CSegGraph::GenerateWordNet(char *sSentence,CDictionary &dictCore,bool bOriginalFreq)
  59. {
  60. //Gernerate the word net from the sLine, that's list all the possible word
  61. unsigned int i=0,j,nLen=strlen(sSentence);
  63. /* 
  64. * ----- commented by huangjin@ict.ac.cn 2006-6-8 ------ 
  65. *
  66. *  char sWord[WORD_MAXLENGTH]="",sTempWord[WORD_MAXLENGTH]="",sWordMatch[WORD_MAXLENGTH];
  67. *
  68. */
  69. /*----Added By huangjin@ict.ac.cn 2006-6-8----*/
  70. char sWord[WORD_MAXLENGTH]="",sTempWord[WORD_MAXLENGTH]="",sWordMatch[WORD_MAXLENGTH]="";
  71. /*---------------------------------------------*/
  73. int nWordIndex=0,nHandleTemp,k,nPOS;
  74. int nMatchFreq[20],nMatchHandle[20],nTotalFreq,nMatchCount;
  75. double dValue=0;
  76. m_nAtomCount=0;
  77. m_segGraph.SetEmpty();//Set segmentation graph empty
  79. AtomSegment(sSentence);
  80. //Atomic Segmentation
  82.     for(i=0;i<m_nAtomCount;i++)//Init the cost array
  83.     {
  84. if(m_nAtomPOS[i]==CT_CHINESE)//The atom is a Chinese Char
  85. {
  86. if(!bOriginalFreq)//Not original frequency
  87. /* 
  88. * ----- commented by huangjin@ict.ac.cn 2006-5-30 ------ 
  89. * 
  90. *  m_segGraph.SetElement(i,i+1,log(MAX_FREQUENCE),0);//init the link with the maximum value
  91. *
  92. */
  93. /*----Added By huangjin@ict.ac.cn 2006-5-30----*/
  94. m_segGraph.SetElement(i,i+1,log(double(MAX_FREQUENCE)),0);//init the link with the maximum value
  95. /*--------------------------------------------*/
  97. else
  98. m_segGraph.SetElement(i,i+1,0,0,m_sAtom[i]);//init the link with the maximum value
  99. }
  100. else//Other atom
  101. {
  102. strcpy(sWord,m_sAtom[i]);//init the word 
  103. dValue=MAX_FREQUENCE;
  104. switch(m_nAtomPOS[i])
  105. {
  106. case CT_INDEX:
  107. case CT_NUM:
  108. /*----Added By huangjin@ict.ac.cn 2006-7-11----*/
  109. case CT_SINGLE_NUM:
  110. /*---------------------------------------------*/
  111. nPOS=-27904;//'m'*256
  112. strcpy(sWord,"未##数");
  113. dValue=0;
  114. break;
  115. case CT_DELIMITER:
  116. nPOS=30464;//'w'*256;
  117. break;
  118. case CT_LETTER:
  119. nPOS=-'n'*256-'x';//
  120. dValue=0;
  121. strcpy(sWord,"未##串");
  122. break;
  123. /*----Added By huangjin@ict.ac.cn 2006-7-11----*/
  124. case CT_SINGLE_DELIMITER://12021-2129-3121
  125. if(GetCharCount("+-1234567890",m_sAtom[i])==(int)strlen(m_sAtom[i]))
  126. {
  127. nPOS=-27904;//'m'*256
  128. dValue=0;
  129. strcpy(sWord,"未##数");
  130. }
  131. else
  132. {
  133. nPOS=30464; //'w'*256
  134. }
  135. break;
  136. /*---------------------------------------------*/
  137. case CT_SINGLE://12021-2129-3121
  138. if(GetCharCount("+-1234567890",m_sAtom[i])==(int)strlen(m_sAtom[i]))
  139. {
  140. nPOS=-27904;//'m'*256
  141. strcpy(sWord,"未##数");
  142. }
  143. else
  144. {
  145. nPOS=-'n'*256-'x';//
  146. strcpy(sWord,"未##串");
  147. }
  148. dValue=0;
  149. break;
  150. default:
  151. nPOS=m_nAtomPOS[i];//'?'*256;
  152. break;
  153. }
  154. if(!bOriginalFreq)//Not original frequency
  155. m_segGraph.SetElement(i,i+1,0,nPOS);//init the link with minimum
  156. else
  157. m_segGraph.SetElement(i,i+1,dValue,nPOS,sWord);//init the link with minimum
  158. }
  159.     }
  160. i=0;
  162. while(i<m_nAtomCount)//All the word
  163. {
  164.   strcpy(sWord,m_sAtom[i]);//Get the current atom
  165.   j=i+1;
  166.   /* 
  167.   * ----- commented by huangjin@ict.ac.cn 2006-5-31 ------ 
  168.   *
  169.   *  if(strcmp(sWord,"月")==0&&strcmp(m_sAtom[i+1],"份")==0)//Don't split 月份
  170.   *
  171.   */
  172.   /*----Added By huangjin@ict.ac.cn 2006-5-31----*/
  173.   //Add "i<m_nAtomCount-1" so that the i will not slop over when compare m_sAtom[i+1]
  174.   if(strcmp(sWord,"月")==0&&i<m_nAtomCount-1&&strcmp(m_sAtom[i+1],"份")==0)//Don't split 月份
  175.   /*---------------------------------------------*/
  176.   j+=1;
  177.   /*----Added By huangjin@ict.ac.cn 2006-7-17----*/
  178.   if((strcmp(sWord,"点")==0||strcmp(sWord,"刻")==0)&&
  179.   i<m_nAtomCount-1&&strcmp(m_sAtom[i+1],"钟")==0)//Don't split 点钟｜刻钟
  180.   j+=1;
  181.   /*---------------------------------------------*/
  182.   //while(j<=m_nAtomCount&&dictCore.GetMaxMatch(sWord,sWordMatch,&nHandleTemp))
  184. //=============================modify by Jiang Wenbin===================================
  185.   while(j<=m_nAtomCount)
  186.   {
  187.   
  188.   bool find=dictCore.GetMaxMatch(sWord,sWordMatch,&nHandleTemp);
  189.   if(find)
  190.   {
  191.   
  192. if(strcmp(sWordMatch,sWord)==0)//find the current word
  193. {
  194.   
  195. nTotalFreq=0;
  196. dictCore.GetHandle(sWord,&nMatchCount,nMatchHandle,nMatchFreq);
  197. for(k=0;k<nMatchCount;k++)//Add the frequency
  198. {
  199. nTotalFreq+=nMatchFreq[k];
  200. }
  201. //Adding a rule to exclude some words to be formed.
  202. if(strlen(sWord)==4&&i>=1&&(IsAllNum((unsigned char *)m_sAtom[i-1])||IsAllChineseNum(m_sAtom[i-1]))&&(strncmp(sWord,"年",2)==0||strncmp(sWord,"月",2)==0))
  203. {//1年内、1999年末
  204. if(CC_Find("末内中底前间初",sWord+2))
  205. break;
  206.  }
  207. if(nMatchCount==1)//The possible word has only one POS, store it
  208. {
  209. if(!bOriginalFreq)//Not original frequency
  210. /* 
  211. * ----- commented by huangjin@ict.ac.cn 2006-5-30 ------ 
  212. * 
  213. *  m_segGraph.SetElement(i,j,-log(nTotalFreq+1)+log(MAX_FREQUENCE),nMatchHandle[0]);
  214. *
  215. */
  216. /*----Added By huangjin@ict.ac.cn 2006-5-30----*/
  217. m_segGraph.SetElement(i,j,-log(double(nTotalFreq+1))+log(double(MAX_FREQUENCE)),nMatchHandle[0]);
  218. /*---------------------------------------------*/
  219. else
  220. m_segGraph.SetElement(i,j,nTotalFreq,nMatchHandle[0],sWord);
  221. }
  222. else 
  223. {
  224. if(!bOriginalFreq)//Not original frequency
  225. /* 
  226. * ----- commented by huangjin@ict.ac.cn 2006-5-30 ------ 
  227. * 
  228. *  m_segGraph.SetElement(i,j,-log(nTotalFreq+1)+log(MAX_FREQUENCE),0);
  229. *
  230. */
  231. /*----Added By huangjin@ict.ac.cn 2006-5-30----*/
  232. m_segGraph.SetElement(i,j,-log(double(nTotalFreq+1))+log(double(MAX_FREQUENCE)),0);
  233. /*---------------------------------------------*/
  234. else
  235. m_segGraph.SetElement(i,j,nTotalFreq,0,sWord);
  236. }
  237. }
  238.   }
  239.   else if(j-i>8)
  240.   {
  241. break;
  242.   }
  244.   //Add a condition to control the end of string
  245.    //retrieve the dictionary with the word
  246.   //cout<<"Word: "<<sWord<<endl;
  247.   
  248.   strcat(sWord,m_sAtom[j++]);
  249.   }
  250.   i+=1;//Start from i++;
  251. }
  253. return true;
  254. }
  256. /* 
  257. * ----- commented by huangjin@ict.ac.cn 2006-7-11 ------ 
  258. *
  259. * 
  260. //DELbool CSegGraph::AtomSegment(char *sSentence)
  261. //DEL{
  262. //DEL unsigned int i=0,j=0,nCurType,nNextType;
  263. //DEL //i is the pointer of sentence string
  264. //DEL //j is the pointer of pAtoms
  265. //DEL char sChar[3];
  266. //DEL sChar[2]=0;//Set the char ending
  267. //DEL m_sAtom[j][0]=0;//Set the first word as null
  268. //DEL m_nAtomLength[j]=0;
  269. //DEL if(strncmp(sSentence,SENTENCE_BEGIN,strlen(SENTENCE_BEGIN))==0)
  270. //DEL {
  271. //DEL strcpy(m_sAtom[j],SENTENCE_BEGIN);//Set the first word as sentence begining
  272. //DEL m_nAtomLength[j]=strlen(SENTENCE_BEGIN);
  273. //DEL m_nAtomPOS[j]=CT_SENTENCE_BEGIN;//init
  274. //DEL i+=m_nAtomLength[j];
  275. //DEL j+=1;
  276. //DEL m_sAtom[j][0]=0;//Set the first word as null
  277. //DEL m_nAtomLength[j]=0;
  278. //DEL }
  279. //DEL while(i<strlen(sSentence))
  280. //DEL {
  281. //DEL if(strncmp(sSentence+i,SENTENCE_END,strlen(SENTENCE_END))==0)
  282. //DEL {
  283. //DEL strcpy(m_sAtom[j],SENTENCE_END);//Set the first word as null
  284. //DEL m_nAtomLength[j]=strlen(SENTENCE_END);
  285. //DEL m_nAtomPOS[j]=CT_SENTENCE_END;//init
  286. //DEL i+=m_nAtomLength[j];
  287. //DEL j+=1;
  288. //DEL m_sAtom[j][0]=0;//Set the first word as null
  289. //DEL m_nAtomLength[j]=0;
  290. //DEL continue;
  291. //DEL }
  292. //DEL /*----Added By huangjin@ict.ac.cn 2006-6-8----*/
  293. //DEL //连续的三个点到六个点认为是省略号，不要切开
  294. //DEL char c=*(sSentence+i);
  295. //DEL if(c=='.')
  296. //DEL {
  297. //DEL char ellipsis[]= "......";
  298. //DEL bool bEllipsis = false;
  299. //DEL for( int tmpindex = 5; tmpindex>=2; tmpindex-- )
  300. //DEL {
  301. //DEL if( strncmp(sSentence+i,ellipsis,strlen(ellipsis))==0)
  302. //DEL {
  303. //DEL strcpy(m_sAtom[j],ellipsis);
  304. //DEL m_nAtomLength[j]=strlen(ellipsis);
  305. //DEL m_nAtomPOS[j]=CT_DELIMITER;//init
  306. //DEL i+=m_nAtomLength[j];
  307. //DEL j+=1;
  308. //DEL m_sAtom[j][0]=0;//Set the first word as null
  309. //DEL m_nAtomLength[j]=0;
  310. //DEL bEllipsis=true;
  311. //DEL break;
  312. //DEL }
  313. //DEL else
  314. //DEL {
  315. //DEL ellipsis[tmpindex]='';
  316. //DEL }
  317. //DEL }
  318. //DEL if( bEllipsis )
  319. //DEL {
  320. //DEL continue;
  321. //DEL }
  322. //DEL }
  323. //DEL else if(c=='-'&&i+1<strlen(sSentence))
  324. //DEL {
  325. //DEL c=*(sSentence+i+1);
  326. //DEL if(c=='-')
  327. //DEL {
  328. //DEL strcpy(m_sAtom[j],"--");
  329. //DEL m_nAtomLength[j]=strlen("--");
  330. //DEL m_nAtomPOS[j]=CT_DELIMITER;//init
  331. //DEL i+=m_nAtomLength[j];
  332. //DEL j+=1;
  333. //DEL m_sAtom[j][0]=0;//Set the first word as null
  334. //DEL m_nAtomLength[j]=0;
  335. //DEL continue;
  336. //DEL }
  337. //DEL }
  338. //DEL /*---------------------------------------------*/
  339. //DEL
  340. //DEL sChar[0]=*(sSentence+i);//Get the char with first byte
  341. //DEL sChar[1]=0;//
  342. //DEL i+=1;
  343. //DEL if(sChar[0]<0)//Two byte char
  344. //DEL {
  345. //DEL sChar[1]=*(sSentence+i);//Get the char with second byte
  346. //DEL i+=1;//i increased by 1
  347. //DEL }
  348. //DEL strcat(m_sAtom[j],sChar);
  349. //DEL nCurType=charType((unsigned char *)sChar);
  350. //DEL if(sChar[0]=='.'&&(charType((unsigned char *)sSentence+i)==CT_NUM||(*(sSentence+i)>='0'&&*(sSentence+i)<='9')))
  351. //DEL nCurType=CT_NUM;//Digit after . indicate . as a point in the numeric
  352. //DEL
  353. //DEL m_nAtomPOS[j]=nCurType;
  354. //DEL //Record its property, just convience for continuous processing
  355. //DEL
  356. //DEL if(nCurType==CT_CHINESE||nCurType==CT_INDEX||nCurType==CT_DELIMITER||nCurType==CT_OTHER)
  357. //DEL {//Chinese char, index number,delimiter and other is treated as atom
  358. //DEL m_nAtomLength[j]=strlen(m_sAtom[j]);//Save its length
  359. //DEL j+=1;//Skip to next atom
  360. //DEL m_sAtom[j][0]=0;//init
  361. //DEL /*----Added By huangjin@ict.ac.cn 2006-6-8----*/
  362. //DEL m_nAtomLength[j]=0;
  363. //DEL /*---------------------------------------------*/
  364. //DEL }
  365. //DEL else
  366. //DEL {//Number,single char, letter
  367. //DEL nNextType=255;
  368. //DEL if(i<strlen(sSentence))
  369. //DEL nNextType=charType((unsigned char *)(sSentence+i));
  370. //DEL if(nNextType!=nCurType||i==strlen(sSentence))
  371. //DEL //Reaching end or next char type is different from current char
  372. //DEL {
  373. //DEL m_nAtomLength[j]=strlen(m_sAtom[j]);//Save its length
  374. //DEL j+=1;
  375. //DEL m_sAtom[j][0]=0;//init
  376. //DEL /*----Added By huangjin@ict.ac.cn 2006-6-8----*/
  377. //DEL   m_nAtomLength[j]=0;
  378. //DEL   /*---------------------------------------------*/
  379. //DEL   }
  380. //DEL  }
  381. //DEL }
  382. //DEL m_nAtomCount=j;//The count of segmentation atoms
  383. //DEL return true;
  384. //DEL}
  385. //DEL*
  386. //DEL*/
  388. /*----Added By huangjin@ict.ac.cn 2006-9-12----*/
  389. bool CSegGraph::AtomSegment(char *sSentence)
  390. {
  391. unsigned int i=0, j=0, nCurType, nNextType;
  392. //i is the pointer of sSentence string
  393. //j is the pointer of m_sAtom, m_nAtomLength an m_nAtomPOS
  394. char sChar[3]="";
  395. const unsigned int nLen = strlen(sSentence); //store the length of sSentence
  397. //Set the first word as null
  398. UpdateAtoms(j,"",-1,false);
  400. if(!strncmp(sSentence,SENTENCE_BEGIN,strlen(SENTENCE_BEGIN))) //sentence begin
  401. {
  402. i+=UpdateAtoms(j,SENTENCE_BEGIN,CT_SENTENCE_BEGIN,false);//Set the word as sentence beginning
  403. j++;
  404. UpdateAtoms(j,"",-1,false);
  405. }
  407. while(i<nLen)
  408. {
  409. if(strncmp(sSentence+i,SENTENCE_END,strlen(SENTENCE_END))==0)
  410. {//Set the word as sentence ending
  411. i+=UpdateAtoms(j,SENTENCE_END,CT_SENTENCE_END,false);
  412. j++;
  413. UpdateAtoms(j,"",-1,false);
  414. continue;
  415. }
  417. i+=GetChar(sSentence+i, sChar); //Get current word
  418. nCurType=charType((unsigned char *)sChar);
  419. UpdateAtoms(j,sChar,nCurType);
  421. if ( nCurType==CT_CHINESE )
  422. {//一万八千六百五十二
  423. unsigned int k=i;
  424. char sNumCandidate[WORD_MAXLENGTH]="";
  425. if(k<nLen)
  426. {
  427. bool bNum=false;
  428. do
  429. {
  430. bNum=IsChineseNumCadidate(sChar);
  431. if(bNum)
  432. {
  433. strcat(sNumCandidate,sChar);
  434. k+=GetChar(sSentence+k,sChar);
  435. }
  436. else if(!strncmp(sSentence+k-2,"分之",4))
  437. {
  438. strcat(sNumCandidate,"分之");
  439. bNum=true;
  440. k+=GetChar(sSentence+k+2,sChar)+2;
  441. }
  442. }while(k<nLen&&bNum);
  443. }
  444. unsigned int h=ChineseNumRecognize(sNumCandidate, sSentence+k-strlen(sChar));
  445. if(h!=(unsigned int)-1)
  446. {
  447. sNumCandidate[h]=0;//截断
  448. UpdateAtoms(j,sNumCandidate,CT_NUM,false);
  449. i+=strlen(sNumCandidate)-2;
  450. }
  451. }
  452. else
  453. {
  454. nNextType=255;
  455. if(i<nLen)
  456. nNextType=charType((unsigned char*)(sSentence+i));
  457. //Numeric
  458. if(nCurType==CT_NUM||nCurType==CT_SINGLE_NUM||
  459. strchr("+-.",sChar[0])&&nNextType==CT_SINGLE_NUM||
  460. CC_Find("±－＋",sChar)&&nNextType==CT_NUM)
  461. {//Numeric and Numeric Prefix
  462. unsigned int k=i; //tmp index save i
  463. bool bSBC = false;
  464. if(strchr("+-.",sChar[0])||nCurType==CT_SINGLE_NUM)
  465. bSBC = true;
  467. bool bPuncEnd=true;
  468. if(nCurType==CT_NUM||nCurType==CT_SINGLE_NUM)
  469. bPuncEnd=false;
  471. while(k<nLen)
  472. {
  473. k+=GetChar(sSentence+k,sChar);
  474. nNextType=charType((unsigned char*)sChar);
  475. if(!bPuncEnd&&(strchr(".+-:",sChar[0])&&sChar[1]==0)||CC_Find("．·：∶／",sChar))
  476. bPuncEnd=true;
  477. else if(bSBC&&nNextType==CT_SINGLE_NUM||!bSBC&&nNextType==CT_NUM)
  478. bPuncEnd=false;
  479. else
  480. break;
  481. strcat(m_sAtom[j],sChar);//Get the atom
  482. i=k;
  483. }
  484. if(bPuncEnd)//末尾是符号或者标点，要分开
  485. {
  486. if(CC_Find("．·：∶／",m_sAtom[j]+strlen(m_sAtom[j])-2))
  487. {
  488. m_sAtom[j][strlen(m_sAtom[j])-2]=0;
  489. i-=2;
  490. }
  491. else //.+-:
  492. {
  493. m_sAtom[j][strlen(m_sAtom[j])-1]=0;
  494. i-=1;
  495. }
  496. }
  497. UpdateAtoms(j,"",CT_NUM);
  498. }//end of Numeric
  499. else if(i<nLen-1&&strchr(".-",sChar[0])&&!sChar[1]&&*(sSentence+i)==sChar[0])
  500. {//for ...... and --
  501. if(sChar[0]=='.')
  502. {
  503. char ellipsis[]= ".....";
  504. bool bEllipsis = false;
  505. for( int tmpindex = 4; tmpindex>=1; tmpindex-- )
  506. {
  507. if( strncmp(sSentence+i,ellipsis,strlen(ellipsis))==0)
  508. {
  509. bEllipsis=true;
  510. break;
  511. }
  512. else
  513. {
  514. ellipsis[tmpindex]='';
  515. }
  516. }
  517. if( bEllipsis )
  518. {
  519. i+=UpdateAtoms(j,ellipsis,CT_DELIMITER);
  520. }
  521. }
  522. else if(sChar[0]=='-')//the next character must be '-'
  523. {
  524. i+=UpdateAtoms(j,"-",CT_DELIMITER);
  525. }
  526. }
  527. else if(nCurType==CT_LETTER||
  528. nCurType==CT_SINGLE&&sChar[0]!='+'&&sChar[0]!='-'||
  529. (strchr("+-",sChar[0])||CC_Find("±－＋",sChar))&&nNextType!=CT_CHINESE)
  530. {//letters
  531. unsigned int k=i;
  532. bool bSBC = true;
  533. if(nCurType==CT_SINGLE||strchr("+-",sChar[0]))
  534. bSBC=false;
  536. while(k<nLen)
  537. {
  538. i=k;
  539. k+=GetChar(sSentence+k,sChar);
  540. nNextType=charType((unsigned char *)sChar);//Get the type
  542. if( bSBC&&(nNextType==CT_LETTER||nNextType==CT_NUM) ||
  543. !bSBC&&(nNextType==CT_SINGLE||nNextType==CT_SINGLE_NUM))
  544. {
  545. strcat(m_sAtom[j],sChar);
  546. }
  547. else
  548. {
  549. break;
  550. }
  551. }
  552. m_nAtomPOS[j]=CT_LETTER;
  553. m_nAtomLength[j]=strlen(m_sAtom[j]);
  554. }
  555. /*---------------------------------------------*/
  556. else if(!strcmp(sChar,"－") && !strncmp(sSentence+i,"－",2) )
  557. {
  558. i+=UpdateAtoms(j,"－",CT_DELIMITER);
  559. }
  560. }
  561. j++;
  562. UpdateAtoms(j,"",-1,false);
  563. }
  564. m_nAtomCount=j;
  565. return true;
  566. }
  567. /*---------------------------------------------*/
  569. /*----Added By huangjin@ict.ac.cn 2006-9-12----*/
  570. int CSegGraph::UpdateAtoms( int j, const char* str, int nPOS, bool bApp )
  571. {
  572. if(!bApp)
  573. {
  574. m_sAtom[j][0]=0;
  575. }
  576. strcat(m_sAtom[j],str);
  577. m_nAtomLength[j]=strlen(m_sAtom[j]);
  578. m_nAtomPOS[j]=nPOS;
  579. return strlen(str);
  580. }
  581. /*---------------------------------------------*/