summaryrefslogtreecommitdiff
path: root/decoder/stringlib.h
blob: a7c6c3c42629878f8b27e30e2bc58cf182ef498e (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
#ifndef CDEC_STRINGLIB_H_
#define CDEC_STRINGLIB_H_

#include <map>
#include <vector>
#include <cctype>
#include <cstring>
#include <string>

// read line in the form of either:
//   source
//   source ||| target
// source will be returned as a string, target must be a sentence or
// a lattice (in PLF format) and will be returned as a Lattice object
void ParseTranslatorInput(const std::string& line, std::string* input, std::string* ref);
struct Lattice;
void ParseTranslatorInputLattice(const std::string& line, std::string* input, Lattice* ref);

inline std::string Trim(const std::string& str, const std::string& dropChars = " \t") {
  std::string res = str;
  res.erase(str.find_last_not_of(dropChars)+1);
  return res.erase(0, res.find_first_not_of(dropChars));
}

inline void Tokenize(const std::string& str, char delimiter, std::vector<std::string>* res) {
  std::string s = str;
  int last = 0;
  res->clear();
  for (int i=0; i < s.size(); ++i)
    if (s[i] == delimiter) {
      s[i]=0;
      if (last != i) {
        res->push_back(&s[last]);
      }
      last = i + 1;
    }
  if (last != s.size())
    res->push_back(&s[last]);
}

inline unsigned NTokens(const std::string& str, char delimiter)
{
  std::vector<std::string> r;
  Tokenize(str,delimiter,&r);
  return r.size();
}

inline std::string LowercaseString(const std::string& in) {
  std::string res(in.size(),' ');
  for (int i = 0; i < in.size(); ++i)
    res[i] = tolower(in[i]);
  return res;
}

inline int CountSubstrings(const std::string& str, const std::string& sub) {
  size_t p = 0;
  int res = 0;
  while (p < str.size()) {
    p = str.find(sub, p);
    if (p == std::string::npos) break;
    ++res;
    p += sub.size();
  }
  return res;
}

inline int SplitOnWhitespace(const std::string& in, std::vector<std::string>* out) {
  out->clear();
  int i = 0;
  int start = 0;
  std::string cur;
  while(i < in.size()) {
    if (in[i] == ' ' || in[i] == '\t') {
      if (i - start > 0)
        out->push_back(in.substr(start, i - start));
      start = i + 1;
    }
    ++i;
  }
  if (i > start)
    out->push_back(in.substr(start, i - start));
  return out->size();
}

inline std::vector<std::string> SplitOnWhitespace(std::string const& in)
{
  std::vector<std::string> r;
  SplitOnWhitespace(in,&r);
  return r;
}


struct mutable_c_str {
  // because making a copy of a string might not copy its storage, so modifying a c_str() could screw up original (nobody uses cow nowadays because it needs locking under threading)
  char *p;
  mutable_c_str(std::string const& s) : p((char *)::operator new(s.size()+1)) {
    std::memcpy(p,s.data(),s.size());
    p[s.size()]=0;
  }
  ~mutable_c_str() { ::operator delete(p); }
private:
  mutable_c_str(mutable_c_str const&);
};

// ' ' '\t' tokens hardcoded
//NOTE: you should have stripped endline chars out first.
inline bool IsWordSep(char c) {
  return c==' '||c=='\t';
}


template <class F>
// *end must be 0 (i.e. [p,end] is valid storage, which will be written to with 0 to separate c string tokens
void VisitTokens(char *p,char *const end,F f) {
  if (p==end) return;
  char *last; // 0 terminated already.  this is ok to mutilate because s is a copy of the string passed in.  well, barring copy on write i guess.
  while(IsWordSep(*p)) { ++p;if (p==end) return; } // skip init whitespace
  last=p; // first non-ws char
  for(;;) {
    ++p;
    // now last is a non-ws char, and p is one past it.
    for(;;) {// p to end of word
      if (p==end) { f(last); return; }
      if (!IsWordSep(*p)) break;
      ++p;
    }
    *p=0;
    f(last);
    for(;;) { // again skip extra whitespace
      ++p;
      if (p==end) return;
      if (!IsWordSep(*p)) break;
    }
    last=p;
  }
}

template <class F>
void VisitTokens(std::string const& s,F f) {
  std::vector<std::string> ss=SplitOnWhitespace(s);
  for (int i=0;i<ss.size();++i)
    f(ss[i]);
  return;
  //FIXME:
  if (s.empty()) return;
  mutable_c_str mp(s);
  VisitTokens(mp.p,mp.p+s.size(),f);
}

inline void SplitCommandAndParam(const std::string& in, std::string* cmd, std::string* param) {
  cmd->clear();
  param->clear();
  std::vector<std::string> x;
  SplitOnWhitespace(in, &x);
  if (x.size() == 0) return;
  *cmd = x[0];
  for (int i = 1; i < x.size(); ++i) {
    if (i > 1) { *param += " "; }
    *param += x[i];
  }
}

void ProcessAndStripSGML(std::string* line, std::map<std::string, std::string>* out);

// given the first character of a UTF8 block, find out how wide it is
// see http://en.wikipedia.org/wiki/UTF-8 for more info
inline unsigned int UTF8Len(unsigned char x) {
  if (x < 0x80) return 1;
  else if ((x >> 5) == 0x06) return 2;
  else if ((x >> 4) == 0x0e) return 3;
  else if ((x >> 3) == 0x1e) return 4;
  else return 0;
}

#endif