// Copyright (C) 1998-1999 Jean-Marc Valin

#include "covariance.h"
#include <math.h>
#include "misc.h"
#include <string>

using namespace std;

namespace FD {

DECLARE_TYPE(DiagonalCovariance)
//@implements GMM

void DiagonalCovariance::processMean(RCPtr<Mean> mean)
{
   if (mode == inverted) return;
   if (mode != accum) throw string("DiagonalCovariance::processMean");

   Mean &v=*mean;
   double accum_1 = 1.0/(.001+v.getAccum());
   for(int i = 0; i < dimension; i++ )
      data[i] -= sqr(v[i])*accum_1;
}

void DiagonalCovariance::invert()
{
   if (mode == inverted) return;
   if (mode != accum) throw string("DiagonalCovariance::invert");
   double accum_1 = 1.0/(.001+accum_count);
for(int i = 0; i < data.size(); i++ )
   {
      data[i] = 1.0 / (.001 + data[i] * accum_1);
   }
   mode = inverted;
}

double DiagonalCovariance::mahalanobisDistance(const float *x1, const double *x2) const
{
   if (mode != inverted) throw string ("DiagonalCovariance::mahalanobisDistance");
   double dist=0;
   for (int i=0;i<dimension;i++)
      dist += sqr(x1[i]-x2[i]) * data[i];
   return dist-getDeterminant();
}

void DiagonalCovariance::compute_determinant() const
{
   if (mode != inverted) throw string ("DiagonalCovariance::compute_determinant");
   determinant=0;

   for (unsigned int i=0;i<dimension;i++)
      determinant += .5*log(data[i]);

   determinant_is_valid = true;
}

void DiagonalCovariance::reset()
{
   determinant_is_valid=false;
   for (unsigned int i=0;i<dimension;i++)
      data[i]=0.0;
   mode = accum;
   accum_count=0;
}

void DiagonalCovariance::printOn(ostream &out) const
{
   out << "<DiagonalCovariance" << endl;
   out << "<dimension " << dimension << ">" << endl;
   out << "<mode " << mode << ">" << endl;
   if (mode == accum)
      out << "<accum_count " << accum_count << "> " << endl;
   out << "<data";
   for (int i=0;i<dimension;i++)
      out << " " << data[i];
   out << ">\n";
   out << ">\n";
   return;
}

void DiagonalCovariance::readFrom (istream &in)
{
   dimension=-1;
   string tag;
   while (1)
   {
      char ch;
      in >> ch;
      if (ch == '>') break;
      in >> tag;
      if (tag == "dimension")
      {
         in >> dimension;
         data.resize(dimension);
      } else if (tag == "mode")
         in >> mode;
            else if (tag == "accum_count")
         in >> accum_count;
      else if (tag == "data")
      {
         if (dimension==-1)
            throw new ParsingException("DiagonalCovariance::readFrom : dimension must be specified before data");
         for (int i=0;i<dimension;i++)
            in >> data[i];
      } else 
         throw new ParsingException ("DiagonalCovariance::readFrom : unknown argument: " + tag);
      if (!in) throw new ParsingException ("DiagonalCovariance::readFrom : Parse error trying to build " + tag);
      in >> tag;
      if (tag != ">") throw new ParsingException ("DiagonalCovariance::readFrom : Parse error: '>' expected ");
   }   
};

istream &operator >> (istream &in, DiagonalCovariance &cov)
{
   if (!isValidType(in, "DiagonalCovariance")) return in;

   cov.readFrom(in);   
   return in;
}
}//namespace FD