/*

Copyright (C) 1994-2017 John W. Eaton

This file is part of Octave.

Octave is free software; you can redistribute it and/or modify it
under the terms of the GNU General Public License as published by
the Free Software Foundation; either version 3 of the License, or
(at your option) any later version.

Octave 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 General Public License for more details.

You should have received a copy of the GNU General Public License
along with Octave; see the file COPYING.  If not, see
<http://www.gnu.org/licenses/>.

*/

#if defined (HAVE_CONFIG_H)
#  include "config.h"
#endif

#include "CMatrix.h"
#include "dMatrix.h"
#include "fCMatrix.h"
#include "fMatrix.h"
#include "lo-error.h"
#include "lo-lapack-proto.h"
#include "schur.h"

namespace octave
{
  namespace math
  {
    // For real types.

    static F77_INT
    select_ana (const double& a, const double&)
    {
      return (a < 0.0);
    }

    static F77_INT
    select_dig (const double& a, const double& b)
    {
      return (hypot (a, b) < 1.0);
    }

    static F77_INT
    select_ana (const float& a, const float&)
    {
      return (a < 0.0);
    }

    static F77_INT
    select_dig (const float& a, const float& b)
    {
      return (hypot (a, b) < 1.0);
    }

    // For complex types.

    static F77_INT
    select_ana (const F77_DBLE_CMPLX& a_arg)
    {
      const Complex a = reinterpret_cast<const Complex&> (a_arg);
      return a.real () < 0.0;
    }

    static F77_INT
    select_dig (const F77_DBLE_CMPLX& a_arg)
    {
      const Complex& a = reinterpret_cast<const Complex&> (a_arg);
      return (abs (a) < 1.0);
    }

    static F77_INT
    select_ana (const F77_CMPLX& a_arg)
    {
      const FloatComplex& a = reinterpret_cast<const FloatComplex&> (a_arg);
      return a.real () < 0.0;
    }

    static F77_INT
    select_dig (const F77_CMPLX& a_arg)
    {
      const FloatComplex& a = reinterpret_cast<const FloatComplex&> (a_arg);
      return (abs (a) < 1.0);
    }

    template <>
    F77_INT
    schur<Matrix>::init (const Matrix& a, const std::string& ord, bool calc_unitary)
    {
      F77_INT a_nr = octave::to_f77_int (a.rows ());
      F77_INT a_nc = octave::to_f77_int (a.cols ());

      if (a_nr != a_nc)
        (*current_liboctave_error_handler) ("schur: requires square matrix");

      if (a_nr == 0)
        {
          schur_mat.clear ();
          unitary_mat.clear ();
          return 0;
        }

      // Workspace requirements may need to be fixed if any of the
      // following change.

      char jobvs;
      char sense = 'N';
      char sort = 'N';

      if (calc_unitary)
        jobvs = 'V';
      else
        jobvs = 'N';

      char ord_char = (ord.empty () ? 'U' : ord[0]);

      if (ord_char == 'A' || ord_char == 'D' || ord_char == 'a' || ord_char == 'd')
        sort = 'S';

      volatile double_selector selector = 0;
      if (ord_char == 'A' || ord_char == 'a')
        selector = select_ana;
      else if (ord_char == 'D' || ord_char == 'd')
        selector = select_dig;

      F77_INT n = a_nc;
      F77_INT lwork = 8 * n;
      F77_INT liwork = 1;
      F77_INT info;
      F77_INT sdim;
      double rconde;
      double rcondv;

      schur_mat = a;

      if (calc_unitary)
        unitary_mat.clear (n, n);

      double *s = schur_mat.fortran_vec ();
      double *q = unitary_mat.fortran_vec ();

      Array<double> wr (dim_vector (n, 1));
      double *pwr = wr.fortran_vec ();

      Array<double> wi (dim_vector (n, 1));
      double *pwi = wi.fortran_vec ();

      Array<double> work (dim_vector (lwork, 1));
      double *pwork = work.fortran_vec ();

      // BWORK is not referenced for the non-ordered Schur routine.
      F77_INT ntmp = (ord_char == 'N' || ord_char == 'n') ? 0 : n;
      Array<F77_INT> bwork (dim_vector (ntmp, 1));
      F77_INT *pbwork = bwork.fortran_vec ();

      Array<F77_INT> iwork (dim_vector (liwork, 1));
      F77_INT *piwork = iwork.fortran_vec ();

      F77_XFCN (dgeesx, DGEESX, (F77_CONST_CHAR_ARG2 (&jobvs, 1),
                                 F77_CONST_CHAR_ARG2 (&sort, 1),
                                 selector,
                                 F77_CONST_CHAR_ARG2 (&sense, 1),
                                 n, s, n, sdim, pwr, pwi, q, n, rconde, rcondv,
                                 pwork, lwork, piwork, liwork, pbwork, info
                                 F77_CHAR_ARG_LEN (1)
                                 F77_CHAR_ARG_LEN (1)
                                 F77_CHAR_ARG_LEN (1)));

      return info;
    }

    template <>
    F77_INT
    schur<FloatMatrix>::init (const FloatMatrix& a, const std::string& ord,
                              bool calc_unitary)
    {
      F77_INT a_nr = octave::to_f77_int (a.rows ());
      F77_INT a_nc = octave::to_f77_int (a.cols ());

      if (a_nr != a_nc)
        (*current_liboctave_error_handler) ("SCHUR requires square matrix");

      if (a_nr == 0)
        {
          schur_mat.clear ();
          unitary_mat.clear ();
          return 0;
        }

      // Workspace requirements may need to be fixed if any of the
      // following change.

      char jobvs;
      char sense = 'N';
      char sort = 'N';

      if (calc_unitary)
        jobvs = 'V';
      else
        jobvs = 'N';

      char ord_char = (ord.empty () ? 'U' : ord[0]);

      if (ord_char == 'A' || ord_char == 'D' || ord_char == 'a' || ord_char == 'd')
        sort = 'S';

      volatile float_selector selector = 0;
      if (ord_char == 'A' || ord_char == 'a')
        selector = select_ana;
      else if (ord_char == 'D' || ord_char == 'd')
        selector = select_dig;

      F77_INT n = a_nc;
      F77_INT lwork = 8 * n;
      F77_INT liwork = 1;
      F77_INT info;
      F77_INT sdim;
      float rconde;
      float rcondv;

      schur_mat = a;

      if (calc_unitary)
        unitary_mat.clear (n, n);

      float *s = schur_mat.fortran_vec ();
      float *q = unitary_mat.fortran_vec ();

      Array<float> wr (dim_vector (n, 1));
      float *pwr = wr.fortran_vec ();

      Array<float> wi (dim_vector (n, 1));
      float *pwi = wi.fortran_vec ();

      Array<float> work (dim_vector (lwork, 1));
      float *pwork = work.fortran_vec ();

      // BWORK is not referenced for the non-ordered Schur routine.
      F77_INT ntmp = (ord_char == 'N' || ord_char == 'n') ? 0 : n;
      Array<F77_INT> bwork (dim_vector (ntmp, 1));
      F77_INT *pbwork = bwork.fortran_vec ();

      Array<F77_INT> iwork (dim_vector (liwork, 1));
      F77_INT *piwork = iwork.fortran_vec ();

      F77_XFCN (sgeesx, SGEESX, (F77_CONST_CHAR_ARG2 (&jobvs, 1),
                                 F77_CONST_CHAR_ARG2 (&sort, 1),
                                 selector,
                                 F77_CONST_CHAR_ARG2 (&sense, 1),
                                 n, s, n, sdim, pwr, pwi, q, n, rconde, rcondv,
                                 pwork, lwork, piwork, liwork, pbwork, info
                                 F77_CHAR_ARG_LEN (1)
                                 F77_CHAR_ARG_LEN (1)
                                 F77_CHAR_ARG_LEN (1)));

      return info;
    }

    template <>
    F77_INT
    schur<ComplexMatrix>::init (const ComplexMatrix& a, const std::string& ord,
                                bool calc_unitary)
    {
      F77_INT a_nr = octave::to_f77_int (a.rows ());
      F77_INT a_nc = octave::to_f77_int (a.cols ());

      if (a_nr != a_nc)
        (*current_liboctave_error_handler) ("SCHUR requires square matrix");

      if (a_nr == 0)
        {
          schur_mat.clear ();
          unitary_mat.clear ();
          return 0;
        }

      // Workspace requirements may need to be fixed if any of the
      // following change.

      char jobvs;
      char sense = 'N';
      char sort = 'N';

      if (calc_unitary)
        jobvs = 'V';
      else
        jobvs = 'N';

      char ord_char = (ord.empty () ? 'U' : ord[0]);

      if (ord_char == 'A' || ord_char == 'D' || ord_char == 'a' || ord_char == 'd')
        sort = 'S';

      volatile complex_selector selector = 0;
      if (ord_char == 'A' || ord_char == 'a')
        selector = select_ana;
      else if (ord_char == 'D' || ord_char == 'd')
        selector = select_dig;

      F77_INT n = a_nc;
      F77_INT lwork = 8 * n;
      F77_INT info;
      F77_INT sdim;
      double rconde;
      double rcondv;

      schur_mat = a;
      if (calc_unitary)
        unitary_mat.clear (n, n);

      Complex *s = schur_mat.fortran_vec ();
      Complex *q = unitary_mat.fortran_vec ();

      Array<double> rwork (dim_vector (n, 1));
      double *prwork = rwork.fortran_vec ();

      Array<Complex> w (dim_vector (n, 1));
      Complex *pw = w.fortran_vec ();

      Array<Complex> work (dim_vector (lwork, 1));
      Complex *pwork = work.fortran_vec ();

      // BWORK is not referenced for non-ordered Schur.
      F77_INT ntmp = (ord_char == 'N' || ord_char == 'n') ? 0 : n;
      Array<F77_INT> bwork (dim_vector (ntmp, 1));
      F77_INT *pbwork = bwork.fortran_vec ();

      F77_XFCN (zgeesx, ZGEESX, (F77_CONST_CHAR_ARG2 (&jobvs, 1),
                                 F77_CONST_CHAR_ARG2 (&sort, 1),
                                 selector,
                                 F77_CONST_CHAR_ARG2 (&sense, 1),
                                 n, F77_DBLE_CMPLX_ARG (s), n, sdim, F77_DBLE_CMPLX_ARG (pw),
                                 F77_DBLE_CMPLX_ARG (q), n, rconde, rcondv,
                                 F77_DBLE_CMPLX_ARG (pwork), lwork, prwork, pbwork, info
                                 F77_CHAR_ARG_LEN (1)
                                 F77_CHAR_ARG_LEN (1)
                                 F77_CHAR_ARG_LEN (1)));

      return info;
    }

    template <>
    schur<ComplexMatrix>
    rsf2csf<ComplexMatrix, Matrix> (const Matrix& s_arg, const Matrix& u_arg)
    {
      ComplexMatrix s (s_arg);
      ComplexMatrix u (u_arg);

      F77_INT n = octave::to_f77_int (s.rows ());

      if (s.columns () != n || u.rows () != n || u.columns () != n)
        (*current_liboctave_error_handler)
          ("rsf2csf: inconsistent matrix dimensions");

      if (n > 0)
        {
          OCTAVE_LOCAL_BUFFER (double, c, n-1);
          OCTAVE_LOCAL_BUFFER (double, sx, n-1);

          F77_XFCN (zrsf2csf, ZRSF2CSF, (n, F77_DBLE_CMPLX_ARG (s.fortran_vec ()),
                                         F77_DBLE_CMPLX_ARG (u.fortran_vec ()), c, sx));
        }

      return schur<ComplexMatrix> (s, u);
    }

    template <>
    F77_INT
    schur<FloatComplexMatrix>::init (const FloatComplexMatrix& a,
                                     const std::string& ord, bool calc_unitary)
    {
      F77_INT a_nr = octave::to_f77_int (a.rows ());
      F77_INT a_nc = octave::to_f77_int (a.cols ());

      if (a_nr != a_nc)
        (*current_liboctave_error_handler) ("SCHUR requires square matrix");

      if (a_nr == 0)
        {
          schur_mat.clear ();
          unitary_mat.clear ();
          return 0;
        }

      // Workspace requirements may need to be fixed if any of the
      // following change.

      char jobvs;
      char sense = 'N';
      char sort = 'N';

      if (calc_unitary)
        jobvs = 'V';
      else
        jobvs = 'N';

      char ord_char = (ord.empty () ? 'U' : ord[0]);

      if (ord_char == 'A' || ord_char == 'D' || ord_char == 'a' || ord_char == 'd')
        sort = 'S';

      volatile float_complex_selector selector = 0;
      if (ord_char == 'A' || ord_char == 'a')
        selector = select_ana;
      else if (ord_char == 'D' || ord_char == 'd')
        selector = select_dig;

      F77_INT n = a_nc;
      F77_INT lwork = 8 * n;
      F77_INT info;
      F77_INT sdim;
      float rconde;
      float rcondv;

      schur_mat = a;
      if (calc_unitary)
        unitary_mat.clear (n, n);

      FloatComplex *s = schur_mat.fortran_vec ();
      FloatComplex *q = unitary_mat.fortran_vec ();

      Array<float> rwork (dim_vector (n, 1));
      float *prwork = rwork.fortran_vec ();

      Array<FloatComplex> w (dim_vector (n, 1));
      FloatComplex *pw = w.fortran_vec ();

      Array<FloatComplex> work (dim_vector (lwork, 1));
      FloatComplex *pwork = work.fortran_vec ();

      // BWORK is not referenced for non-ordered Schur.
      F77_INT ntmp = (ord_char == 'N' || ord_char == 'n') ? 0 : n;
      Array<F77_INT> bwork (dim_vector (ntmp, 1));
      F77_INT *pbwork = bwork.fortran_vec ();

      F77_XFCN (cgeesx, CGEESX, (F77_CONST_CHAR_ARG2 (&jobvs, 1),
                                 F77_CONST_CHAR_ARG2 (&sort, 1),
                                 selector,
                                 F77_CONST_CHAR_ARG2 (&sense, 1),
                                 n, F77_CMPLX_ARG (s), n, sdim, F77_CMPLX_ARG (pw), F77_CMPLX_ARG (q), n, rconde,
                                 rcondv,
                                 F77_CMPLX_ARG (pwork), lwork, prwork, pbwork, info
                                 F77_CHAR_ARG_LEN (1)
                                 F77_CHAR_ARG_LEN (1)
                                 F77_CHAR_ARG_LEN (1)));

      return info;
    }

    template <>
    schur<FloatComplexMatrix>
    rsf2csf<FloatComplexMatrix, FloatMatrix> (const FloatMatrix& s_arg,
                                              const FloatMatrix& u_arg)
    {
      FloatComplexMatrix s (s_arg);
      FloatComplexMatrix u (u_arg);

      F77_INT n = octave::to_f77_int (s.rows ());

      if (s.columns () != n || u.rows () != n || u.columns () != n)
        (*current_liboctave_error_handler)
          ("rsf2csf: inconsistent matrix dimensions");

      if (n > 0)
        {
          OCTAVE_LOCAL_BUFFER (float, c, n-1);
          OCTAVE_LOCAL_BUFFER (float, sx, n-1);

          F77_XFCN (crsf2csf, CRSF2CSF, (n, F77_CMPLX_ARG (s.fortran_vec ()),
                                         F77_CMPLX_ARG (u.fortran_vec ()), c, sx));
        }

      return schur<FloatComplexMatrix> (s, u);
    }

    // Instantiations we need.

    template class schur<ComplexMatrix>;

    template class schur<FloatComplexMatrix>;

    template class schur<FloatMatrix>;

    template class schur<Matrix>;
  }
}