released/pIRS.old/stat_illumina_reads/gcvgstater/self_util.cpp

   1 #include <string>
   2 #include <sstream>
   3 #include <iostream>
   4 #include "self_util.h"
   5
   6 using namespace std;
   7
   8 char alphabet[128] =
   9 {
  10     4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4,
  11     4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4,
  12     4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4,
  13     4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4,
  14     4, 0, 4, 1, 4, 4, 4, 2, 4, 4, 4, 4, 4, 4, 0, 4,
  15     4, 4, 4, 4, 3, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4,
  16     4, 0, 4, 1, 4, 4, 4, 2, 4, 4, 4, 4, 4, 4, 0, 4,
  17     4, 4, 4, 4, 3, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4
  18 };
  19
  20 char bases[5] =
  21 {
  22     'A', 'C', 'G', 'T', 'N'
  23 };
  24
  25 char c_bases[5] =
  26 {
  27     'T', 'G', 'C', 'A', 'N'
  28 };
  29
  30 uint64_t seq2bit(string &kseq)
  31 {
  32     uint64_t kbit=0;
  33
  34     for(int i=0; i<kseq.size(); i++)
  35     {
  36         kbit=(kbit<<2)|alphabet[kseq[i]];
  37     }
  38
  39     return kbit;
  40 }
  41
  42 string bit2seq(uint64_t kbit, int kmerSize)
  43 {
  44     string kseq;
  45
  46     for(int i=0; i<kmerSize; i++)
  47     {
  48         kseq.push_back(bases[(kbit>>(kmerSize-1-i)*2)&0x3]);
  49     }
  50
  51     return kseq;
  52 }
  53
  54 bool check_seq(string &seq)
  55 {
  56     bool is_good = true;
  57     for (int i = 0; i < seq.size(); i++)
  58     {
  59         if (alphabet[seq[i]] == 4)
  60         {
  61             is_good = false;
  62             break;
  63         }
  64     }
  65
  66     return is_good;
  67 }
  68
  69
  70 void reverse_complement (string &in_str, string &out_str)
  71 {
  72     for (int64_t i=in_str.size()-1; i>=0; i--)
  73     {
  74         out_str.push_back(c_bases[alphabet[in_str[i]]]);
  75     }
  76 }
  77
  78 void TrimRight(string& str)
  79 {
  80     int pos = str.length() - 1;
  81     if(pos >= 0)
  82     {
  83         char c = str[pos];
  84         while((c == '\t') || (c == '\n') || (c == ' '))
  85         {
  86             str = str.substr(0, pos);
  87             pos = str.length() - 1;
  88             if(pos < 0)
  89             {
  90                 break;
  91             }
  92             c = str[pos];
  93         }
  94     }
  95 }
  96
  97 void TrimLeft(string& str)
  98 {
  99     if(str.length() > 0)
 100     {
 101         char c = str[0];
 102         while((c == '\t') || (c == '\n') || (c == ' '))
 103         {
 104             str = str.substr(1);
 105             if(str.length() == 0)
 106             {
 107                 break;
 108             }
 109             c = str[0];
 110         }
 111     }
 112 }
 113
 114 vector<string> splitString(string str, string delims)
 115 {
 116     vector<string> vec;
 117     string temp;
 118     //const string delims(" \t\n");
 119     string::size_type begIdx, endIdx;
 120     begIdx = str.find_first_not_of(delims);
 121     while(begIdx != string::npos)
 122     {
 123         endIdx = str.find_first_of(delims, begIdx);
 124         if(endIdx == string::npos)
 125         {
 126             endIdx == str.length();
 127         }
 128
 129         vec.push_back(str.substr(begIdx, endIdx-begIdx));
 130         begIdx = str.find_first_not_of(delims, endIdx);
 131     }
 132     return vec;
 133 }
 134
 135 vector<int> splitStringToInt(string str, string delims)
 136 {
 137     vector<int> vec;
 138     string temp;
 139     string::size_type begIdx, endIdx;
 140     begIdx = str.find_first_not_of(delims);
 141     while(begIdx != string::npos)
 142     {
 143         endIdx = str.find_first_of(delims, begIdx);
 144         if(endIdx == string::npos)
 145         {
 146             endIdx == str.length();
 147         }
 148         vec.push_back(toInt(str.substr(begIdx, endIdx-begIdx)));
 149         begIdx = str.find_first_not_of(delims, endIdx);
 150     }
 151     return vec;
 152 }
 153
 154 int toInt(string str)
 155 {
 156     int i;
 157     stringstream ss;
 158     //int prec = numeric_limits<int>::digits10;
 159     //ss.precision(prec);
 160     ss << str;
 161     ss >> i;
 162     return i;
 163 }
 164
 165 uint64_t toUint64(string str)
 166 {
 167     int i;
 168     stringstream ss;
 169     //int prec = numeric_limits<int>::digits10;
 170     //ss.precision(prec);
 171     ss << str;
 172     ss >> i;
 173     return i;
 174 }
 175
 176 char toUpper(const char& ch)
 177 {
 178     if(islower(ch))
 179     {
 180         return ch & 0x5F;
 181     }
 182 }
 183
 184 char toLower(const char& ch)
 185 {
 186     if(isupper(ch))
 187     {
 188         return ch | 0x20;
 189     }
 190 }
 191 string ToUpper(string str)
 192 {
 193     string temp = "";
 194     for(int i=0; i<str.length(); ++i)
 195     {
 196         temp += toUpper(str[i]);
 197     }
 198     return temp;
 199 }
 200
 201 string ToLower(string str)
 202 {
 203     string temp = "";
 204     //transform (str.begin(),str.end(), temp.begin(), toLower);
 205     for(int i=0; i<str.length(); ++i)
 206     {
 207         temp += toUpper(str[i]);
 208     }
 209     return temp;
 210 }
 211