matrix.h

Go to the documentation of this file.

#ifndef MATRIX_H
#define MATRIX_H

#include <string>
#include <vector>
#include <iostream>
#include <fstream>
#include "bioseq.h"

using namespace std;

//int aachar2num(char a);
//declared in bioseq.h
/*
int max(int i1, int i2, int i3) {
   return max(max(i1,i2),i3);
}
int max(int i1, int i2) {
   if (i1>i2) return i1;
   return i2;
}
*/

void expandCombination(char** &buff, int &buffsize, int &ws, const char* alphabet, int alphabetsize);

class Matrix {
   public:
      Matrix() : path("/home/kzhou/proj/seqaln/matrix"),
            name("blosum50"), match(0), mismatch(0),
            words(0), wordSize(0), mat() { for (int i=0; i<26; i++) for (int j=0; j<26; j++) mat[i][j]=default_mat[i][j]; /*read();*/ }
      Matrix(const string& matrixName, bool nucMatrix=false) 
         : path("/home/kzhou/proj/seqaln/matrix"),
            name(matrixName), match(0), mismatch(0),
            words(0), wordSize(0) { read(nucMatrix); }
      Matrix(const Matrix& mt);

      Matrix(int m, int mis) : match(m), mismatch(mis) { }
      ~Matrix();
      int getMinScore() const { return mins; }
      int getMaxScore() const { return maxs; }
      Matrix& operator=(const Matrix& mt);
      void show();

      const char* getAlphabet() const { return aas; }
      int getNumberOfAlphabet() const { return numsymbol; }

      void getWords(vector<string> &words, int ws) const;
      void growWord(const vector<string> &in, vector<string> &ou) const;
      void expandWord(char** &in, int &insize, int &ws) const;
      char** allwords(int ws) const;
      int getNumberOfWords() const { return wordsArraySize; }
      int getCurrentWordSize() const { return wordSize; }
      void showWords() const; // debug function
      int similarWord_debug(ostream &ous, const char* w, int ws, int cutoff=10, float fractioncutoff=0.75) const;
      int similarWord(vector<char*> &neighbor, const char* w, int ws, float cutoff=11, float fractioncutoff=0.8) const;

      /* return true for success, false for failure.
       * if nucMatrix is set, then it will use the 
       * hashbase() function defined in codon.h to
       * convert Base Char to integer, other than the generic 
       * char X-'A' function
       * */
      bool read(bool nucMatrix=false);
      bool read(const string &p, const string &n) { setPath(p); name=n; read(); }

      void setMatrix(const string &n) { name=n; read(); }

      void setPath(const string &p) { path=p; }
      int getMatchScore(const char c1, const char c2) {
         if (c1==c2) return match; return mismatch; }
      int lookup(int row, int col) const { 
         if (row > 26 || col > 26) {
            cerr << "Matrix index out of bound " << row << ", " << col << endl;
            exit(1);
         }
         return mat[row][col]; }
      int lookup(char row, char col) const { return mat[aachar2num(row)][aachar2num(col)]; }

      int score(const int* x, const int* y, const int len) const;

      /* the gap score, this should have something to do 
       * with the matrix
       * g is the length of the gap.
       */
      //int gap(int g);
      //gap(g)=alpha + (g-1)*beta
      //this is used in the resursion
      int alpha, beta;

   private:
      string path;
      string name; // Blosum50, Blosum62, etc
      int match;
      int mismatch;
      int mat[32][32]; // use only 27 of the elements
      static int default_mat[32][32];
      int maxs, mins;
      char aas[32]; // amino acid symbols, at most 26 so we had enough
      int numsymbol; // the size of aas array
      mutable char **words;
      mutable int wordsArraySize;
      mutable int wordSize;  // recoreds the size of words
};

#endif