void MainWindow::textBagOfWords(std::set<std::string> &featureDic, std::map<std::string, int> &text,int &textSize)
{
if(!ICTCLAS_Init()) //初始化分词组件。
{
QMessageBox::warning(this,"Warnning","Init fails",QMessageBox::Yes);
return;
}
else
{
printf("Init ok\n");
}
ICTCLAS_SetPOSmap(2);
QString myStr = ui->text->toPlainText();
QByteArray ba = myStr.toLocal8Bit();
char* sText;
char* sSentence = (char*)malloc(ba.size()+10);
sText = ba.data();
int len = 0;
//这样读入为了将换行符去掉，ASCII码10、13
for(int i = 0; sText[i] != '\0'; i++)
{
if(sText[i] == '\n' || sText[i] == '\r')continue;
sSentence[len++] = sText[i];
}
sSentence[len] = '\0';
unsigned int nPaLen=strlen(sSentence); // 需要分词的长度
char* sRst=0; //用户自行分配空间，用于保存结果；
sRst=(char*)malloc(nPaLen*6); //建议长度为字符串长度的6倍。
int nRstLen=0; //分词结果的长度
nRstLen = ICTCLAS_ParagraphProcess(sSentence,nPaLen,sRst,CODE_TYPE_UNKNOWN,0); //字符串处理
//free(sText);
free(sSentence);
//收集单词，形成字典
std::string words;
std::istringstream istream(sRst);
std::set<std::string> txtWords;
while(istream>>words)
{
txtWords.insert(words);
if(featureDic.count(words))
{
text[words]++;
}
}
textSize = txtWords.size();
free(sRst);
txtWords.clear();
istream.clear();
ICTCLAS_Exit(); //释放资源退出
return;
}
int main(int argc, char **argv)
{
int c;
/* 默认参数设置 */
char *httpcws_settings_listen = "0.0.0.0";
int httpcws_settings_port = 1985;
char *httpcws_settings_datapath = NULL; /*中文词典数据库路径 */
bool httpcws_settings_daemon = false;
int httpcws_settings_timeout = 120; /* 单位：秒 */
/* process arguments */
while ((c = getopt(argc, argv, "l:p:x:t:dh")) != -1) {
switch (c) {
case 'l':
httpcws_settings_listen = strdup(optarg);
break;
case 'p':
httpcws_settings_port = atoi(optarg);
break;
case 'x':
httpcws_settings_datapath = strdup(optarg); /* 词库文件存储路径 */
break;
case 't':
httpcws_settings_timeout = atoi(optarg);
break;
case 'd':
httpcws_settings_daemon = true;
break;
case 'h':
default:
show_help();
return 1;
}
}
/* 判断是否加了必填参数 -x */
if (httpcws_settings_datapath == NULL) {
show_help();
fprintf(stderr, "Attention: Please use the indispensable argument: -x <path>\n\n");
exit(1);
}
/* 初始化分词组件 */
if(!ICTCLAS_Init(httpcws_settings_datapath))
{
printf("%s\n", httpcws_settings_datapath);
fprintf(stderr, "ERROR: Count not open the Chinese dictionary!\n\n");
exit(1);
}
ICTCLAS_SetPOSmap(ICT_POS_MAP_SECOND);
fprintf(stderr, "Loading Chinese dictionary 'httpcws_dict.txt' into memory, please waitting ......\n");
char *httpcws_settings_dataname = (char *)malloc(1024);
memset (httpcws_settings_dataname, '\0', 1024);
sprintf(httpcws_settings_dataname, "%s/httpcws_dict.txt", httpcws_settings_datapath);
int nCount = ICTCLAS_ImportUserDict(httpcws_settings_dataname);
ICTCLAS_SaveTheUsrDic();
free(httpcws_settings_dataname);
printf("OK! %d words has loaded into memory.\n\n", nCount);
printf("HTTPCWS Server running on %s:%d\n", httpcws_settings_listen, httpcws_settings_port);
/* 如果加了-d参数，以守护进程运行 */
if (httpcws_settings_daemon == true){
pid_t pid;
/* Fork off the parent process */
pid = fork();
if (pid < 0) {
exit(EXIT_FAILURE);
}
/* If we got a good PID, then
we can exit the parent process. */
if (pid > 0) {
exit(EXIT_SUCCESS);
}
}
/* 请求处理部分 */
struct evhttp *httpd;
event_init();
httpd = evhttp_start(httpcws_settings_listen, httpcws_settings_port);
evhttp_set_timeout(httpd, httpcws_settings_timeout);
/* Set a callback for all other requests. */
evhttp_set_gencb(httpd, httpcws_handler, NULL);
event_dispatch();
/* Not reached in this code as it is now. */
evhttp_free(httpd);
return 0;
}
void SplitGBK(const char *sInput)
{//分词演示
//初始化分词组件
if(!ICTCLAS_Init())//数据在当前路径下，默认为GBK编码的分词
{
printf("ICTCLAS INIT FAILED!\n");
return ;
}
ICTCLAS_SetPOSmap(ICT_POS_MAP_SECOND);
char sSentence[2000]="三枪拍案惊奇的主创人员包括孙红雷、小沈阳、闫妮等，导演为张艺谋";
const char * sResult;
int nCount;
ICTCLAS_ParagraphProcessA(sSentence,&nCount);
printf("nCount=%d\n",nCount);
ICTCLAS_AddUserWord("孙红雷 yym");//添加孙红雷，作为演员名称
sResult = ICTCLAS_ParagraphProcess(sSentence,1);
printf("%s\n", sResult);
ICTCLAS_AddUserWord("小沈阳 yym");//添加小沈阳，作为演员名称
sResult = ICTCLAS_ParagraphProcess(sSentence,1);
printf("%s\n", sResult);
ICTCLAS_AddUserWord("闫妮 yym");//添加闫妮，作为演员名称
sResult = ICTCLAS_ParagraphProcess(sSentence,1);
printf("%s\n", sResult);
ICTCLAS_AddUserWord("三枪拍案惊奇 dym");//添加三枪拍案惊奇，作为电影名称
sResult = ICTCLAS_ParagraphProcess(sSentence,1);
printf("%s\n", sResult);
while(_stricmp(sSentence,"q")!=0)
{
sResult = ICTCLAS_ParagraphProcess(sSentence,0);
printf("%s\nInput string now('q' to quit)!\n", sResult);
scanf("%s",sSentence);
}
//导入用户词典前
printf("未导入用户词典：\n");
sResult = ICTCLAS_ParagraphProcess(sInput, 0);
printf("%s\n", sResult);
//导入用户词典后
printf("\n导入用户词典后：\n");
nCount = ICTCLAS_ImportUserDict("userdic.txt");//userdic.txt覆盖以前的用户词典
//保存用户词典
ICTCLAS_SaveTheUsrDic();
printf("导入%d个用户词。\n", nCount);
sResult = ICTCLAS_ParagraphProcess(sInput, 1);
printf("%s\n", sResult);
//动态添加用户词
printf("\n动态添加用户词后：\n");
ICTCLAS_AddUserWord("计算机学院 xueyuan");
ICTCLAS_SaveTheUsrDic();
sResult = ICTCLAS_ParagraphProcess(sInput, 1);
printf("%s\n", sResult);
//对文件进行分词
ICTCLAS_FileProcess("testGBK.txt","testGBK_result.txt",1);
//释放分词组件资源
ICTCLAS_Exit();
}
