/*

Copyright (C) 1994-2015 John W. Eaton
Copyright (C) 2008-2009 Jaroslav Hajek

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/>.

*/

#ifdef HAVE_CONFIG_H
#  include <config.h>
#endif

#include <vector>

#include "fRowVector.h"
#include "floatCHOL.h"
#include "f77-fcn.h"
#include "lo-error.h"
#include "oct-locbuf.h"
#include "oct-norm.h"
#ifndef HAVE_QRUPDATE
#  include "dbleQR.h"
#endif

extern "C"
{
  F77_RET_T
  F77_FUNC (spotrf, SPOTRF) (F77_CONST_CHAR_ARG_DECL,
                             const octave_idx_type&, float*,
                             const octave_idx_type&, octave_idx_type&
                             F77_CHAR_ARG_LEN_DECL);

  F77_RET_T
  F77_FUNC (spotri, SPOTRI) (F77_CONST_CHAR_ARG_DECL,
                             const octave_idx_type&, float*,
                             const octave_idx_type&, octave_idx_type&
                             F77_CHAR_ARG_LEN_DECL);

  F77_RET_T
  F77_FUNC (spocon, SPOCON) (F77_CONST_CHAR_ARG_DECL,
                             const octave_idx_type&, float*,
                             const octave_idx_type&, const float&,
                             float&, float*, octave_idx_type*,
                             octave_idx_type&
                             F77_CHAR_ARG_LEN_DECL);
#ifdef HAVE_QRUPDATE

  F77_RET_T
  F77_FUNC (sch1up, SCH1UP) (const octave_idx_type&, float*,
                             const octave_idx_type&, float*, float*);

  F77_RET_T
  F77_FUNC (sch1dn, SCH1DN) (const octave_idx_type&, float*,
                             const octave_idx_type&, float*, float*,
                             octave_idx_type&);

  F77_RET_T
  F77_FUNC (schinx, SCHINX) (const octave_idx_type&, float*,
                             const octave_idx_type&, const octave_idx_type&,
                             float*, float*, octave_idx_type&);

  F77_RET_T
  F77_FUNC (schdex, SCHDEX) (const octave_idx_type&, float*,
                             const octave_idx_type&, const octave_idx_type&,
                             float*);

  F77_RET_T
  F77_FUNC (schshx, SCHSHX) (const octave_idx_type&, float*,
                             const octave_idx_type&, const octave_idx_type&,
                             const octave_idx_type&, float*);
#endif
}

octave_idx_type
FloatCHOL::init (const FloatMatrix& a, bool upper, bool calc_cond)
{
  octave_idx_type a_nr = a.rows ();
  octave_idx_type a_nc = a.cols ();

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

  octave_idx_type n = a_nc;
  octave_idx_type info;

  is_upper = upper;

  chol_mat.clear (n, n);
  if (is_upper)
    for (octave_idx_type j = 0; j < n; j++)
      {
        for (octave_idx_type i = 0; i <= j; i++)
          chol_mat.xelem (i, j) = a(i, j);
        for (octave_idx_type i = j+1; i < n; i++)
          chol_mat.xelem (i, j) = 0.0f;
      }
  else
    for (octave_idx_type j = 0; j < n; j++)
      {
        for (octave_idx_type i = 0; i < j; i++)
          chol_mat.xelem (i, j) = 0.0f;
        for (octave_idx_type i = j; i < n; i++)
          chol_mat.xelem (i, j) = a(i, j);
      }
  float *h = chol_mat.fortran_vec ();

  // Calculate the norm of the matrix, for later use.
  float anorm = 0;
  if (calc_cond)
    anorm = xnorm (a, 1);

  if (is_upper)
    F77_XFCN (spotrf, SPOTRF, (F77_CONST_CHAR_ARG2 ("U", 1), n, h, n, info
                               F77_CHAR_ARG_LEN (1)));
  else
    F77_XFCN (spotrf, SPOTRF, (F77_CONST_CHAR_ARG2 ("L", 1), n, h, n, info
                               F77_CHAR_ARG_LEN (1)));

  xrcond = 0.0;
  if (info > 0)
    chol_mat.resize (info - 1, info - 1);
  else if (calc_cond)
    {
      octave_idx_type spocon_info = 0;

      // Now calculate the condition number for non-singular matrix.
      Array<float> z (dim_vector (3*n, 1));
      float *pz = z.fortran_vec ();
      Array<octave_idx_type> iz (dim_vector (n, 1));
      octave_idx_type *piz = iz.fortran_vec ();
      if (is_upper)
        F77_XFCN (spocon, SPOCON, (F77_CONST_CHAR_ARG2 ("U", 1), n, h,
                                   n, anorm, xrcond, pz, piz, spocon_info
                                   F77_CHAR_ARG_LEN (1)));
      else
        F77_XFCN (spocon, SPOCON, (F77_CONST_CHAR_ARG2 ("L", 1), n, h,
                                   n, anorm, xrcond, pz, piz, spocon_info
                                   F77_CHAR_ARG_LEN (1)));

      if (spocon_info != 0)
        info = -1;
    }

  return info;
}

static FloatMatrix
chol2inv_internal (const FloatMatrix& r, bool is_upper = true)
{
  FloatMatrix retval;

  octave_idx_type r_nr = r.rows ();
  octave_idx_type r_nc = r.cols ();

  if (r_nr != r_nc)
    (*current_liboctave_error_handler) ("chol2inv requires square matrix");

  octave_idx_type n = r_nc;
  octave_idx_type info = 0;

  FloatMatrix tmp = r;
  float *v = tmp.fortran_vec ();

  if (info == 0)
    {
      if (is_upper)
        F77_XFCN (spotri, SPOTRI, (F77_CONST_CHAR_ARG2 ("U", 1), n,
                                   v, n, info
                                   F77_CHAR_ARG_LEN (1)));
      else
        F77_XFCN (spotri, SPOTRI, (F77_CONST_CHAR_ARG2 ("L", 1), n,
                                   v, n, info
                                   F77_CHAR_ARG_LEN (1)));

      // If someone thinks of a more graceful way of doing this (or
      // faster for that matter :-)), please let me know!

      if (n > 1)
        {
          if (is_upper)
            for (octave_idx_type j = 0; j < r_nc; j++)
              for (octave_idx_type i = j+1; i < r_nr; i++)
                tmp.xelem (i, j) = tmp.xelem (j, i);
          else
            for (octave_idx_type j = 0; j < r_nc; j++)
              for (octave_idx_type i = j+1; i < r_nr; i++)
                tmp.xelem (j, i) = tmp.xelem (i, j);
        }

      retval = tmp;
    }

  return retval;
}

// Compute the inverse of a matrix using the Cholesky factorization.
FloatMatrix
FloatCHOL::inverse (void) const
{
  return chol2inv_internal (chol_mat, is_upper);
}

void
FloatCHOL::set (const FloatMatrix& R)
{
  if (! R.is_square ())
    (*current_liboctave_error_handler) ("FloatCHOL requires square matrix");

  chol_mat = R;
}

#ifdef HAVE_QRUPDATE

void
FloatCHOL::update (const FloatColumnVector& u)
{
  octave_idx_type n = chol_mat.rows ();

  if (u.numel () != n)
    (*current_liboctave_error_handler) ("cholupdate: dimension mismatch");

  FloatColumnVector utmp = u;

  OCTAVE_LOCAL_BUFFER (float, w, n);

  F77_XFCN (sch1up, SCH1UP, (n, chol_mat.fortran_vec (), chol_mat.rows (),
                             utmp.fortran_vec (), w));
}

octave_idx_type
FloatCHOL::downdate (const FloatColumnVector& u)
{
  octave_idx_type info = -1;

  octave_idx_type n = chol_mat.rows ();

  if (u.numel () != n)
    (*current_liboctave_error_handler) ("cholupdate: dimension mismatch");

  FloatColumnVector utmp = u;

  OCTAVE_LOCAL_BUFFER (float, w, n);

  F77_XFCN (sch1dn, SCH1DN, (n, chol_mat.fortran_vec (), chol_mat.rows (),
                             utmp.fortran_vec (), w, info));

  return info;
}

octave_idx_type
FloatCHOL::insert_sym (const FloatColumnVector& u, octave_idx_type j)
{
  octave_idx_type info = -1;

  octave_idx_type n = chol_mat.rows ();

  if (u.numel () != n + 1)
    (*current_liboctave_error_handler) ("cholinsert: dimension mismatch");
  if (j < 0 || j > n)
    (*current_liboctave_error_handler) ("cholinsert: index out of range");

  FloatColumnVector utmp = u;

  OCTAVE_LOCAL_BUFFER (float, w, n);

  chol_mat.resize (n+1, n+1);

  F77_XFCN (schinx, SCHINX, (n, chol_mat.fortran_vec (), chol_mat.rows (),
                             j + 1, utmp.fortran_vec (), w, info));

  return info;
}

void
FloatCHOL::delete_sym (octave_idx_type j)
{
  octave_idx_type n = chol_mat.rows ();

  if (j < 0 || j > n-1)
    (*current_liboctave_error_handler) ("choldelete: index out of range");

  OCTAVE_LOCAL_BUFFER (float, w, n);

  F77_XFCN (schdex, SCHDEX, (n, chol_mat.fortran_vec (), chol_mat.rows (),
                             j + 1, w));

  chol_mat.resize (n-1, n-1);
}

void
FloatCHOL::shift_sym (octave_idx_type i, octave_idx_type j)
{
  octave_idx_type n = chol_mat.rows ();

  if (i < 0 || i > n-1 || j < 0 || j > n-1)
    (*current_liboctave_error_handler) ("cholshift: index out of range");

  OCTAVE_LOCAL_BUFFER (float, w, 2*n);

  F77_XFCN (schshx, SCHSHX, (n, chol_mat.fortran_vec (), chol_mat.rows (),
                             i + 1, j + 1, w));
}

#else

void
FloatCHOL::update (const FloatColumnVector& u)
{
  warn_qrupdate_once ();

  octave_idx_type n = chol_mat.rows ();

  if (u.numel () != n)
    (*current_liboctave_error_handler) ("cholupdate: dimension mismatch");

  init (chol_mat.transpose () * chol_mat
        + FloatMatrix (u) * FloatMatrix (u).transpose (), true, false);
}

static bool
singular (const FloatMatrix& a)
{
  for (octave_idx_type i = 0; i < a.rows (); i++)
    if (a(i,i) == 0.0f) return true;
  return false;
}

octave_idx_type
FloatCHOL::downdate (const FloatColumnVector& u)
{
  warn_qrupdate_once ();

  octave_idx_type info = -1;

  octave_idx_type n = chol_mat.rows ();

  if (u.numel () != n)
    (*current_liboctave_error_handler) ("cholupdate: dimension mismatch");

  if (singular (chol_mat))
    info = 2;
  else
    {
      info = init (chol_mat.transpose () * chol_mat
                   - FloatMatrix (u) * FloatMatrix (u).transpose (), true,
                   false);
      if (info) info = 1;
    }

  return info;
}

octave_idx_type
FloatCHOL::insert_sym (const FloatColumnVector& u, octave_idx_type j)
{
  warn_qrupdate_once ();

  octave_idx_type info = -1;

  octave_idx_type n = chol_mat.rows ();

  if (u.numel () != n + 1)
    (*current_liboctave_error_handler) ("cholinsert: dimension mismatch");
  if (j < 0 || j > n)
    (*current_liboctave_error_handler) ("cholinsert: index out of range");

  if (singular (chol_mat))
    info = 2;
  else
    {
      FloatMatrix a = chol_mat.transpose () * chol_mat;
      FloatMatrix a1 (n+1, n+1);
      for (octave_idx_type k = 0; k < n+1; k++)
        for (octave_idx_type l = 0; l < n+1; l++)
          {
            if (l == j)
              a1(k, l) = u(k);
            else if (k == j)
              a1(k, l) = u(l);
            else
              a1(k, l) = a(k < j ? k : k-1, l < j ? l : l-1);
          }
      info = init (a1, true, false);
      if (info) info = 1;
    }

  return info;
}

void
FloatCHOL::delete_sym (octave_idx_type j)
{
  warn_qrupdate_once ();

  octave_idx_type n = chol_mat.rows ();

  if (j < 0 || j > n-1)
    (*current_liboctave_error_handler) ("choldelete: index out of range");

  FloatMatrix a = chol_mat.transpose () * chol_mat;
  a.delete_elements (1, idx_vector (j));
  a.delete_elements (0, idx_vector (j));
  init (a, true, false);
}

void
FloatCHOL::shift_sym (octave_idx_type i, octave_idx_type j)
{
  warn_qrupdate_once ();

  octave_idx_type n = chol_mat.rows ();

  if (i < 0 || i > n-1 || j < 0 || j > n-1)
    (*current_liboctave_error_handler) ("cholshift: index out of range");

  FloatMatrix a = chol_mat.transpose () * chol_mat;
  Array<octave_idx_type> p (dim_vector (n, 1));
  for (octave_idx_type k = 0; k < n; k++) p(k) = k;
  if (i < j)
    {
      for (octave_idx_type k = i; k < j; k++) p(k) = k+1;
      p(j) = i;
    }
  else if (j < i)
    {
      p(j) = i;
      for (octave_idx_type k = j+1; k < i+1; k++) p(k) = k-1;
    }

  init (a.index (idx_vector (p), idx_vector (p)), true, false);
}

#endif

FloatMatrix
chol2inv (const FloatMatrix& r)
{
  return chol2inv_internal (r);
}