STK_lapack_Svd.h

Go to the documentation of this file.

/*--------------------------------------------------------------------*/
/*     Copyright (C) 2004-2016  Serge Iovleff, Université Lille 1, Inria
 
    This program is free software; you can redistribute it and/or modify
    it under the terms of the GNU Lesser General Public License as
    published by the Free Software Foundation; either version 2 of the
    License, or (at your option) any later version.
 
    This program is distributed in the hope that it will be useful,
    but WITHOUT ANY WARRANTY; without even the implied warranty of
    MERCHANTABILITY or FITNESS FOR a PARTICULAR PURPOSE.  See the
    GNU Lesser General Public License for more details.
 
    You should have received a copy of the GNU Lesser General Public
    License along with this program; if not, write to the
    Free Software Foundation, Inc.,
    59 Temple Place,
    Suite 330,
    Boston, MA 02111-1307
    USA
 
    Contact : S..._Dot_I..._At_stkpp_Dot_org (see copyright for ...)
*/
 
/*
 * Project:  stkpp::Algebra
 * created on: 20 nov. 2013
 * Author:   iovleff, S..._Dot_I..._At_stkpp_Dot_org (see copyright for ...)
 **/
 
#ifndef STK_LAPACK_SVD_H
#define STK_LAPACK_SVD_H
 
#include <Arrays/include/STK_CArray.h>
#include <Arrays/include/STK_CArrayVector.h>
 
#include "STK_ISvd.h"
#include "STK_lapack_Util.h"
#include "STK_transpose.h"
 
namespace STK
{
 
namespace lapack
{
class Svd;
}
 
namespace hidden
{
template<>
struct AlgebraTraits< lapack::Svd >
{
  typedef CArrayXX ArrayU;
  typedef CVectorX ArrayD;
  typedef CArrayXX ArrayV;
};
 
} // namespace hidden
 
 
namespace lapack
{
class Svd: public ISvd<Svd>
{
  public:
    typedef ISvd< Svd > Base;
    typedef typename hidden::Traits<CArrayXX>::Col ColVector;
    typedef typename hidden::Traits<CArrayXX>::Row RowVector;
 
    using Base::norm_;
    using Base::rank_;
    using Base::U_;
    using Base::D_;
    using Base::V_;
    using Base::withU_;
    using Base::withV_;
    using Base::nrowU;
    using Base::ncolU;
    using Base::ncolV;
 
    inline Svd( CArrayXX const&  A, bool ref = false, bool withU= true, bool withV= true)
              : Base(A, ref, withU, withV), jobz_( (withU|withV) ? 'O':'N') {}
    template<class OtherArray>
    inline Svd( ArrayBase<OtherArray> const& A, bool withU= true, bool withV= true)
              : Base(A, withU, withV), jobz_( (withU|withV) ? 'O':'N') {}
    inline Svd( Svd const& decomp): Base(decomp), jobz_(decomp.jobz_) {}
    inline virtual ~Svd() {}
    inline virtual Svd* clone() const { return new Svd(*this);}
    inline Svd& operator=(Svd const& decomp)
    {
      Base::operator=(decomp);
      jobz_ = decomp.jobz_;
      return *this;
    }
    // getters
    char jobz() const { return jobz_;}
    // setters
    void setJobz(char jobz) { jobz_ = jobz;}
 
    bool runImpl();
 
  private:
     char jobz_;
     bool computeSvd(CArrayXX& a, CArrayXX& u, CVectorX& s, CArrayXX& v);
};
 
/* @brief Run svd decomposition */
inline bool Svd::runImpl()
{
  if (jobz_ == 'A' || jobz_ == 'S')
  {
    msg_error_ = _T("In lapack::Svd::runImpl, the options 'A' and 'S' are not available");
    return false;
  }
  int beginRow = U_.beginRows(), beginCol = U_.beginCols();
  bool result = true;
  // compute results
  if ( (jobz_ == 'O' && U_.sizeRows() >= U_.sizeCols()) || jobz_ == 'N')
  {
    if (!computeSvd(U_, U_, D_, V_)) { result = false;}
  }
  else
  { // jobz_ == 'O' and m<n, V_ will contain u and U_ will contain vt
    if (!computeSvd(U_, V_, D_, V_)) { result = false;};
    U_.exchange(V_); // !now U_ is (m,m) and V_ is (m,n)
  }
  U_.shift(beginRow, beginCol);
  D_.shift(beginCol);
  transpose(V_.shift(beginCol));
  return result;
}
 
 
/* Computing the Svd decomposition of the matrix U_. */
inline bool Svd::computeSvd(CArrayXX& a, CArrayXX& u, CVectorX& s, CArrayXX& vt)
{
  int m = a.sizeRows(), n = a.sizeCols(), nbSv = std::min(m,n);
  a.shift(0,0);
  // Workspace and status variables:
  double workSize;
  double *p_work = &workSize;
  int* p_iwork = new int[8*nbSv];
  int lwork = -1;
  int info;
  //
  // Call dgesdd_ with lwork = -1 to query optimal workspace size:
  info = gesdd( jobz_, m, n
               , a.p_data(), m, s.p_data(), u.p_data(), m, vt.p_data(), n
               , p_work, lwork, p_iwork);
  // check any error
  if (info!=0)
  {
    if (info>0)
    { this->msg_error_ = STKERROR_NO_ARG(lapack::Svd::computeSvd,internal error);
      return false;
    }
    this->msg_error_= STKERROR_1ARG(lapack::Svd::computeSvd get,-info,error parameter);
    return false;
  }
  // optimal storage dimension is returned in work[0]
  lwork = workSize;
  p_work = new Real[lwork];
  // resize u
  if ( !((jobz_ == 'O' && m >= n) || (jobz_ == 'N')) )
  {
    int ucol = (jobz_ == 'A' || jobz_ == 'O') ? m : nbSv;
    u.resize(m, ucol);
  }
  if ( !((jobz_ == 'O' && m < n)  || (jobz_ == 'N')) )
  {
    int ldvt = (jobz_ == 'A' || jobz_ == 'O') ? n : nbSv;
    vt.resize(ldvt,n);
  }
  s.resize(nbSv).shift(0);
  u.shift(0,0);
  vt.shift(0,0);
 
  // Call dgesdd_ to do the actual computation:
  info = gesdd( jobz_, m, n
              , a.p_data(), a.sizeRows(), s.p_data()
              , u.p_data(), u.sizeRows()
              , vt.p_data(), vt.sizeRows()
              , p_work, lwork, p_iwork);
  // clean
  delete[] p_work;
  delete[] p_iwork;
  // check any error
  if (info!=0)
  {
    if (info>0)
    { this->msg_error_ = STKERROR_NO_ARG(lapack::Svd::computeSvd,internal error);
      return false;
    }
    this->msg_error_= STKERROR_1ARG(lapack::Svd::computeSvd get,-info,error parameter);
    return false;
  }
  return true;
}
 
} // namespace lapack
 
} // namespace STK
 
#endif /* STK_LAPACK_SVD_H */