third_codes/cppjieba
1
0
Fork 0
mirror of https://github.com/yanyiwu/cppjieba.git synced 2025-07-18 00:00:12 +08:00
Code Issues Packages Projects Releases Wiki Activity

update cppcommon for splitStr

Browse Source
This commit is contained in:
gwdwyy 2013-09-04 18:03:24 +08:00
parent 2319fd9307
commit a8374bd2b7
2 changed files with 283 additions and 337 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

512
src/cppcommon/str_functs.cpp
View File

@ -6,227 +6,231 @@
namespace CPPCOMMON namespace CPPCOMMON
{ {
//http://stackoverflow.com/questions/2342162/stdstring-formatting-like-sprintf //http://stackoverflow.com/questions/2342162/stdstring-formatting-like-sprintf
string string_format(const string fmt, ...) string string_format(const string fmt, ...)
{ {
int size = 100; int size = 100;
std::string str; std::string str;
va_list ap; va_list ap;
while (1) { while (1) {
str.resize(size); str.resize(size);
va_start(ap, fmt); va_start(ap, fmt);
int n = vsnprintf((char *)str.c_str(), size, fmt.c_str(), ap); int n = vsnprintf((char *)str.c_str(), size, fmt.c_str(), ap);
va_end(ap); va_end(ap);
if (n > -1 && n < size) { if (n > -1 && n < size) {
str.resize(n); str.resize(n);
return str; return str;
} }
if (n > -1) if (n > -1)
size = n + 1; size = n + 1;
else else
size *= 2; size *= 2;
} }
return str; return str;
} }
string joinStr(const vector<string>& src, const string& connectorStr) string joinStr(const vector<string>& src, const string& connectorStr)
{ {
string res = ""; string res = "";
int len = src.size(); int len = src.size();
if(0 == len) if(0 == len)
{ {
return ""; return "";
} }
for(int i = 0; i < len - 1; i++) for(int i = 0; i < len - 1; i++)
{ {
res += stripStr(src[i]); res += stripStr(src[i]);
res += connectorStr; res += connectorStr;
} }
res += stripStr(src[len-1]); res += stripStr(src[len-1]);
return res; return res;
} }
vector<string> splitStr(const string& source, const string& pattern)
{
vector<string> res;
splitStr(source, res, pattern);
return res;
}
void splitStr(const string& source, vector<string>& out_vec, const string& pattern)
{
if(source.empty())
{
return;
}
out_vec.clear();
string s = source + pattern;
string::size_type pos;
uint length = s.size();
for(uint i = 0; i < length; i++) vector<string> splitStr(const string& source, const string& pattern)
{ {
pos = s.find(pattern, i); vector<string> res;
if(pos < length) splitStr(source, res, pattern);
{ return res;
string tmp = stripStr(s.substr(i, pos - i)); }
if("" != tmp)
{
out_vec.push_back(tmp);
}
i = pos + pattern.size() - 1;
}
}
}
string stripStr(const string& str, const string& patternStr) bool splitStr(const string& source, vector<string>& res, const string& pattern)
{ {
if(str.empty()) if(source.empty())
{ {
return str; return false;
} }
string::size_type posL = str.find_first_not_of(patternStr); res.clear();
if(string::npos == posL)
{
return str;
}
string::size_type posR = str.find_last_not_of(patternStr);
return str.substr(posL, posR - posL + 1);
} size_t start = source.find_first_not_of(pattern);
size_t end;
if(string::npos == start)
//http://stackoverflow.com/questions/216823/whats-the-best-way-to-trim-stdstring
// trim from start
std::string &ltrim(std::string &s)
{
s.erase(s.begin(), std::find_if(s.begin(), s.end(), std::not1(std::ptr_fun<int, int>(std::isspace))));
return s;
}
// trim from end
std::string &rtrim(std::string &s)
{
s.erase(std::find_if(s.rbegin(), s.rend(), std::not1(std::ptr_fun<int, int>(std::isspace))).base(), s.end());
return s;
}
// trim from both ends
std::string &trim(std::string &s)
{
return ltrim(rtrim(s));
}
bool splitStrMultiPatterns(
const string& strSrc,
vector<string>& outVec,
const vector<string>& patterns
)
{
char transChar = '#';
uint transLenThreshold = 10;
string transStr;
transStr += transChar;
while(strSrc.find(transStr) != string::npos)
{
transStr += transChar;
if(transStr.size() > transLenThreshold)
return false;
}
string strSrcMutable = strSrc;
for(uint i = 0; i < patterns.size(); i++)
{
strSrcMutable = replaceStr(strSrcMutable, patterns[i], transStr);
}
splitStr(strSrcMutable, outVec, transStr);
return true;
}
string upperStr(const string& strIn)
{
string str = strIn;
transform(str.begin(), str.end(), str.begin(), (int (*)(int))toupper);
return str;
}
string lowerStr(const string& strIn)
{
string str = strIn;
transform(str.begin(), str.end(), str.begin(), (int (*)(int))tolower);
return str;
}
string replaceStr(const string& strSrc, const string& oldStr, const string& newStr, int count)
{
string strRet = strSrc;
size_t pos = 0;
int l_count = 0;
if(-1 == count)
count = strRet.size();
while((pos = strRet.find(oldStr, pos)) != string::npos)
{
strRet.replace(pos, oldStr.size(), newStr);
if(++l_count >= count)
break;
pos += newStr.size();
}
return strRet;
}
unsigned int countStrDistance(const string& A, const string& B)
{
unsigned int lenA = A.size();
unsigned int lenB = B.size();
unsigned int len = (lenA < lenB ? lenA : lenB);
unsigned int res = lenA + lenB - 2 * len;
for(size_t i = 0; i < len; i++)
{
if(A[i] != B[i])
res++;
}
return res;
}
unsigned int countStrSimilarity(const string& A, const string& B)
{
unsigned int lenA = A.size();
unsigned int lenB = B.size();
unsigned int len = (lenA < lenB ? lenA : lenB);
unsigned int res = 0;
for(size_t i = 0; i < len; i++)
{
if(A[i] == B[i])
res++;
}
return res;
}
//unicode str to vec
bool uniStrToVec(const string& str, Unicode& vec)
{
vec.clear();
if(str.empty() || str.size() % 2)
{ {
return false; return false;
} }
for(uint i = 0; i < str.size(); i+=2) while(string::npos != start)
{ {
vec.push_back(twocharToUint16(str[i], str[i + 1])); end = source.find_first_of(pattern, start);
if(string::npos == end)
{
res.push_back(source.substr(start));
return true;
}
res.push_back(source.substr(start, end - start));
start = source.find_first_not_of(pattern, end);
} }
return true;
}
return true; string stripStr(const string& str, const string& patternStr)
} {
if(str.empty())
{
return str;
}
string::size_type posL = str.find_first_not_of(patternStr);
if(string::npos == posL)
{
return str;
}
string::size_type posR = str.find_last_not_of(patternStr);
return str.substr(posL, posR - posL + 1);
//unicode vec to str }
string uniVecToStr(const Unicode& vec)
{
string res(""); //http://stackoverflow.com/questions/216823/whats-the-best-way-to-trim-stdstring
for(uint i = 0; i < vec.size(); i++) // trim from start
{ std::string &ltrim(std::string &s)
pair<char,char> pa = uint16ToChar2(vec[i]); {
res += pa.first; s.erase(s.begin(), std::find_if(s.begin(), s.end(), std::not1(std::ptr_fun<int, int>(std::isspace))));
res += pa.second; return s;
} }
return res;
} // trim from end
std::string &rtrim(std::string &s)
{
s.erase(std::find_if(s.rbegin(), s.rend(), std::not1(std::ptr_fun<int, int>(std::isspace))).base(), s.end());
return s;
}
// trim from both ends
std::string &trim(std::string &s)
{
return ltrim(rtrim(s));
}
bool splitStrMultiPatterns(
const string& strSrc,
vector<string>& outVec,
const vector<string>& patterns
)
{
char transChar = '#';
uint transLenThreshold = 10;
string transStr;
transStr += transChar;
while(strSrc.find(transStr) != string::npos)
{
transStr += transChar;
if(transStr.size() > transLenThreshold)
return false;
}
string strSrcMutable = strSrc;
for(uint i = 0; i < patterns.size(); i++)
{
strSrcMutable = replaceStr(strSrcMutable, patterns[i], transStr);
}
splitStr(strSrcMutable, outVec, transStr);
return true;
}
string upperStr(const string& strIn)
{
string str = strIn;
transform(str.begin(), str.end(), str.begin(), (int (*)(int))toupper);
return str;
}
string lowerStr(const string& strIn)
{
string str = strIn;
transform(str.begin(), str.end(), str.begin(), (int (*)(int))tolower);
return str;
}
string replaceStr(const string& strSrc, const string& oldStr, const string& newStr, int count)
{
string strRet = strSrc;
size_t pos = 0;
int l_count = 0;
if(-1 == count)
count = strRet.size();
while((pos = strRet.find(oldStr, pos)) != string::npos)
{
strRet.replace(pos, oldStr.size(), newStr);
if(++l_count >= count)
break;
pos += newStr.size();
}
return strRet;
}
unsigned int countStrDistance(const string& A, const string& B)
{
unsigned int lenA = A.size();
unsigned int lenB = B.size();
unsigned int len = (lenA < lenB ? lenA : lenB);
unsigned int res = lenA + lenB - 2 * len;
for(size_t i = 0; i < len; i++)
{
if(A[i] != B[i])
res++;
}
return res;
}
unsigned int countStrSimilarity(const string& A, const string& B)
{
unsigned int lenA = A.size();
unsigned int lenB = B.size();
unsigned int len = (lenA < lenB ? lenA : lenB);
unsigned int res = 0;
for(size_t i = 0; i < len; i++)
{
if(A[i] == B[i])
res++;
}
return res;
}
//unicode str to vec
bool uniStrToVec(const string& str, Unicode& vec)
{
vec.clear();
if(str.empty() || str.size() % 2)
{
return false;
}
for(uint i = 0; i < str.size(); i+=2)
{
vec.push_back(twocharToUint16(str[i], str[i + 1]));
}
return true;
}
//unicode vec to str
string uniVecToStr(const Unicode& vec)
{
string res("");
for(uint i = 0; i < vec.size(); i++)
{
pair<char,char> pa = uint16ToChar2(vec[i]);
res += pa.first;
res += pa.second;
}
return res;
}
} }
@ -236,78 +240,20 @@ using namespace CPPCOMMON;
using namespace std; using namespace std;
int main() int main()
{ {
//string s = " \t\n1 a h \n"; vector<string> vec;
//cout<<"["<<stripStr(s)<<"]"<<endl; splitStr("1 3 4", vec);
//cout<<countStrDistance("Aheheh","heheh1212")<<endl; for(uint i =0;i < vec.size(); i++)
//cout<<joinStr(splitStr(s), ",")<<endl; {
//vector<string> vec; cout<<vec[i]<<endl;
//splitStr("1 3 4", vec); }
//char * a[] = {"3","jaj","ads"}; cout<<strStartsWith("--help","--")<<endl;
//vector<string> pats(a,a+3); cout<<strStartsWith("--help","-")<<endl;
//vec.clear(); cout<<strStartsWith("--help","he")<<endl;
//splitStrMultiPattern("1 #3 jajs5 asdf3ads 4", vec, pats); cout<<strStartsWith("help","help")<<endl;
//for(int i=0;i<vec.size();i++) cout<<strStartsWith("","help")<<endl;
//{ cout<<strStartsWith("hel","")<<endl;
// cout<<vec[i]<<endl; cout<<strEndsWith("hel","")<<endl;
//} cout<<strEndsWith("hel","el")<<endl;
//string s = "1111aaafasfa,asdj.sadhashfhaha"; return 0;
//upperStr(s);
//cout<<s<<endl;
//
//s = "ab1ba2ab3";
//cout<<replaceStr(s,"ab","###")<<endl;
//ifstream ifile("testdata/dict.txt");
//string line;
//while(getline(ifile, line))
//{
// uint16_t strbuf[1024];
// size_t unilen = utf8ToUnicode(line.c_str(), line.size(), strbuf);
// for(int i = 0; i < unilen; i++)
// {
// // printf("%x\n", strbuf[i]);
// }
// char utf8str[512]={0};
// unicodeToUtf8(strbuf, unilen, utf8str);
// //cout<<strlen(utf8str);
// cout<<utf8str<<endl;
//}
//cout<<string_format("hehe%s11asd%dasf","[here]",2);
//ifstream ifile("testdata/dict.gbk");
//string line;
//Unicode unicode;
//while(getline(ifile, line))
//{
// cout<<line<<endl;
// utf8ToUnicode(line, unicode);
// printUnicode(unicode);
// cout<<unicodeToUtf8(unicode)<<endl;;
//}
//vector<string> tmp;
//tmp.push_back("1");
////tmp.push_back("2");
////tmp.clear();
//cout<<joinStr(tmp, ",")<<endl;
//ifstream ifile("testdata/dict.gbk");
//string line;
//while(getline(ifile, line))
//{
// cout<<line<<endl;
// string s = gbkToUtf8(line);
// s = utf8ToGbk(s);
// cout<<s<<endl;
//}
cout<<strStartsWith("--help","--")<<endl;
cout<<strStartsWith("--help","-")<<endl;
cout<<strStartsWith("--help","he")<<endl;
cout<<strStartsWith("help","help")<<endl;
cout<<strStartsWith("","help")<<endl;
cout<<strStartsWith("hel","")<<endl;
cout<<strEndsWith("hel","")<<endl;
cout<<strEndsWith("hel","el")<<endl;
//string s(" helloword heh\t");
//string b;
//cout<<trim(b)<<"11"<<endl;
return 0;
} }
#endif #endif

108
src/cppcommon/str_functs.h
View File

@ -21,66 +21,66 @@
#include <locale> #include <locale>
namespace CPPCOMMON namespace CPPCOMMON
{ {
using namespace std; using namespace std;
string string_format(const string fmt, ...) ; string string_format(const string fmt, ...) ;
string joinStr(const vector<string>& source, const string& connector); string joinStr(const vector<string>& source, const string& connector);
vector<string> splitStr(const string& source, const string& pattern = " \t\n"); vector<string> splitStr(const string& source, const string& pattern = " \t\n");
void splitStr(const string& source, vector<string>& out_vec, const string& pattern = " \t\n"); bool splitStr(const string& source, vector<string>& res, const string& pattern = " \t\n");
bool splitStrMultiPatterns( bool splitStrMultiPatterns(
const string& strSrc, const string& strSrc,
vector<string>& outVec, vector<string>& outVec,
const vector<string>& patterns const vector<string>& patterns
); );
string upperStr(const string& str); string upperStr(const string& str);
string lowerStr(const string& str); string lowerStr(const string& str);
string replaceStr(const string& strSrc, const string& oldStr, const string& newStr, int count = -1); string replaceStr(const string& strSrc, const string& oldStr, const string& newStr, int count = -1);
string stripStr(const string& str, const string& patternstr = " \n\t"); string stripStr(const string& str, const string& patternstr = " \n\t");
std::string &ltrim(std::string &s) ; std::string &ltrim(std::string &s) ;
std::string &rtrim(std::string &s) ; std::string &rtrim(std::string &s) ;
std::string &trim(std::string &s) ; std::string &trim(std::string &s) ;
unsigned int countStrDistance(const string& A, const string& B); unsigned int countStrDistance(const string& A, const string& B);
unsigned int countStrSimilarity(const string& A, const string& B); unsigned int countStrSimilarity(const string& A, const string& B);
bool uniStrToVec(const string& str, Unicode& vec); bool uniStrToVec(const string& str, Unicode& vec);
string uniVecToStr(const Unicode& vec); string uniVecToStr(const Unicode& vec);
inline uint16_t twocharToUint16(char high, char low) inline uint16_t twocharToUint16(char high, char low)
{ {
return (((uint16_t(high) & 0x00ff ) << 8) | (uint16_t(low) & 0x00ff)); return (((uint16_t(high) & 0x00ff ) << 8) | (uint16_t(low) & 0x00ff));
} }
inline pair<char, char> uint16ToChar2(uint16_t in) inline pair<char, char> uint16ToChar2(uint16_t in)
{ {
pair<char, char> res; pair<char, char> res;
res.first = (in>>8) & 0x00ff; //high res.first = (in>>8) & 0x00ff; //high
res.second = (in) & 0x00ff; //low res.second = (in) & 0x00ff; //low
return res; return res;
} }
inline void printUnicode(const Unicode& unicode) inline void printUnicode(const Unicode& unicode)
{ {
cout<<uniVecToStr(unicode)<<endl; cout<<uniVecToStr(unicode)<<endl;
} }
inline bool strStartsWith(const string& str, const string& prefix) inline bool strStartsWith(const string& str, const string& prefix)
{ {
//return str.substr(0, prefix.size()) == prefix; //return str.substr(0, prefix.size()) == prefix;
if(prefix.length() > str.length()) if(prefix.length() > str.length())
{ {
return false; return false;
} }
return 0 == str.compare(0, prefix.length(), prefix); return 0 == str.compare(0, prefix.length(), prefix);
} }
inline bool strEndsWith(const string& str, const string& suffix) inline bool strEndsWith(const string& str, const string& suffix)
{ {
if(suffix.length() > str.length()) if(suffix.length() > str.length())
{ {
return false; return false;
} }
return 0 == str.compare(str.length() - suffix.length(), suffix.length(), suffix); return 0 == str.compare(str.length() - suffix.length(), suffix.length(), suffix);
} }
} }
#endif #endif