////////////////////////////////////////////////////////////////////////
//
// Copyright (C) 1998-2023 The Octave Project Developers
//
// See the file COPYRIGHT.md in the top-level directory of this
// distribution or <https://octave.org/copyright/>.
//
// 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
// <https://www.gnu.org/licenses/>.
//
////////////////////////////////////////////////////////////////////////

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

#include <cstdlib>
#include <string>

#include "variables.h"
#include "utils.h"
#include "pager.h"
#include "defun.h"
#include "errwarn.h"
#include "quit.h"
#include "unwind-prot.h"

#include "ov-re-sparse.h"
#include "ov-cx-sparse.h"
#include "ov-bool-sparse.h"

OCTAVE_BEGIN_NAMESPACE(octave)

DEFUN (issparse, args, ,
       doc: /* -*- texinfo -*-
@deftypefn {} {@var{tf} =} issparse (@var{x})
Return true if @var{x} is a sparse matrix.
@seealso{ismatrix}
@end deftypefn */)
{
  if (args.length () != 1)
    print_usage ();

  return ovl (args(0).issparse ());
}

/*
%!assert (issparse (sparse (1)), true)
%!assert (issparse (1), false)
%!assert (issparse (sparse (false)), true)
%!assert (issparse (true), false)
%!test
%! warning ("off", "Octave:sparse:double-conversion", "local");
%! assert (issparse (sparse (single ([1 2]))), true);
%!assert (issparse (single ([1, 2])), false)
%!assert (issparse (sparse ([1+i, 2]')), true)
%!assert (issparse ([1+i, 2]'), false)

%!assert (issparse ([]), false)
%!assert (issparse (sparse([])), true)
%!assert (issparse ("test"), false)
%!assert (issparse (struct ("one", {1})), false)
%!assert (issparse (cell (1)), false)

## Test input validation
%!error issparse ()
%!error issparse (1,2)
*/

DEFUN (sparse, args, ,
       doc: /* -*- texinfo -*-
@deftypefn  {} {@var{S} =} sparse (@var{A})
@deftypefnx {} {@var{S} =} sparse (@var{m}, @var{n})
@deftypefnx {} {@var{S} =} sparse (@var{i}, @var{j}, @var{sv})
@deftypefnx {} {@var{S} =} sparse (@var{i}, @var{j}, @var{sv}, @var{m}, @var{n})
@deftypefnx {} {@var{S} =} sparse (@var{i}, @var{j}, @var{sv}, @var{m}, @var{n}, "unique")
@deftypefnx {} {@var{S} =} sparse (@var{i}, @var{j}, @var{sv}, @var{m}, @var{n}, @var{nzmax})
Create a sparse matrix from a full matrix @var{A} or row, column, value
triplets.

If @var{A} is a full matrix, convert it to a sparse matrix representation,
removing all zero values in the process.  The matrix @var{A} should be of type
logical or double.

If two inputs @var{m} (rows) and @var{n} (columns) are specified then create
an empty sparse matrix with the specified dimensions.

Given the integer index vectors @var{i} and @var{j}, and a 1-by-@code{nnz}
vector of real or complex values @var{sv}, construct the sparse matrix
@code{S(@var{i}(@var{k}),@var{j}(@var{k})) = @var{sv}(@var{k})} with overall
dimensions @var{m} and @var{n}.  If any of @var{i}, @var{j}, or @var{sv} are
scalars, they are expanded to have a common size.

If @var{m} or @var{n} are not specified then their values are derived from the
maximum index in the vectors @var{i} and @var{j} as given by
@w{@code{@var{m} = max (@var{i})}}, @w{@code{@var{n} = max (@var{j})}}.

@strong{Note}: If multiple values are specified with the same @var{i},
@var{j} indices, the corresponding value in @var{S} will be the sum of the
values at the repeated location.  @xref{XREFaccumarray,,@code{accumarray}}, for
an example of how to produce different behavior such as taking the minimum
instead.

If the option @qcode{"unique"} is given, and more than one value is specified
at the same @var{i}, @var{j} indices, then only the last specified value will
be used.  For completeness, the option @qcode{"sum"} can be given and will
be ignored as the default behavior is to sum values at repeated locations.

@code{sparse (@var{m}, @var{n})} will create an empty @var{m}x@var{n} sparse
matrix and is equivalent to @code{sparse ([], [], [], @var{m}, @var{n})}

The optional final argument reserves space for @var{nzmax} values in the sparse
array and is useful if the eventual number of nonzero values will be greater
than the number of values in @var{sv} used during the initial construction of
the array.  @xref{XREFspalloc,,@code{spalloc}}, for more information and usage
instructions.

Example 1 (convert full matrix to sparse to save memory):

@example
@group
x = full (diag (1:1000));
sizeof (x)
@result{}  8000000
s = sparse (x);
sizeof (xs)
@result{}  24008
@end group
@end example

Example 2 (sum at repeated indices):

@example
@group
@var{i} = [1 1 2]; @var{j} = [1 1 2]; @var{sv} = [3 4 5];
sparse (@var{i}, @var{j}, @var{sv}, 3, 4)
@result{}
   Compressed Column Sparse (rows = 3, cols = 4, nnz = 2 [17%])

     (1, 1) ->  7
     (2, 2) ->  5
@end group
@end example

Example 3 ("unique" option):

@example
@group
@var{i} = [1 1 2]; @var{j} = [1 1 2]; @var{sv} = [3 4 5];
sparse (@var{i}, @var{j}, @var{sv}, 3, 4, "unique")
@result{}
   Compressed Column Sparse (rows = 3, cols = 4, nnz = 2 [17%])

     (1, 1) ->  4
     (2, 2) ->  5
@end group
@end example
@seealso{full, accumarray, spalloc, spdiags, speye, spones, sprand, sprandn,
sprandsym, spconvert, spfun}
@end deftypefn */)
{
  int nargin = args.length ();

  if (nargin == 0 || nargin > 6)
    print_usage ();

  octave_value retval;

  if (nargin == 1)
    {
      octave_value arg = args(0);
      if (arg.isfloat ())
        {
          if (arg.is_single_type ())
            warning_with_id ("Octave:sparse:double-conversion",
                             "sparse: input array cast to double");
          if (arg.iscomplex ())
            retval = arg.sparse_complex_matrix_value ();
          else
            retval = arg.sparse_matrix_value ();
        }
      else if (arg.islogical ())
        retval = arg.sparse_bool_matrix_value ();
      else
        err_wrong_type_arg ("sparse", arg);
    }
  else if (nargin == 2)
    {
      octave_idx_type m = 0;
      octave_idx_type n = 0;

      get_dimensions (args(0), args(1), "sparse", m, n);

      // FIXME: this code is never active because get_dimensions()
      //        replaces negative dimensions with 0.
      if (m < 0 || n < 0)
        error ("sparse: dimensions must be non-negative");

      retval = SparseMatrix (m, n);
    }
  else if (nargin >= 3)
    {
      bool summation = true;
      if (nargin > 3 && args(nargin-1).is_string ())
        {
          std::string opt = args(nargin-1).string_value ();
          if (opt == "unique")
            summation = false;
          else if (opt == "sum" || opt == "summation")
            summation = true;
          else
            error ("sparse: invalid option: %s", opt.c_str ());

          nargin -= 1;
        }

      octave_idx_type m, n, nzmax;
      m = n = nzmax = -1;
      if (nargin == 6)
        {
          nzmax = args(5).idx_type_value ();
          nargin--;
        }

      if (nargin == 5)
        {
          get_dimensions (args(3), args(4), "sparse", m, n);

          // FIXME: this code is never active because get_dimensions()
          //        replaces negative dimensions with 0.
          if (m < 0 || n < 0)
            error ("sparse: dimensions must be non-negative");
        }

      int argidx = 0;    // index we're checking when index_vector throws
      try
        {
          idx_vector i = args(0).index_vector ();
          argidx = 1;
          idx_vector j = args(1).index_vector ();

          octave_value arg = args(2);  // temp var for code readability
          if (arg.isfloat ())
            {
              if (arg.is_single_type ())
                warning_with_id ("Octave:sparse:double-conversion",
                                 "sparse: input array cast to double");
              if (arg.iscomplex ())
                retval = SparseComplexMatrix (arg.complex_array_value (),
                                              i, j, m, n, summation, nzmax);
              else
                retval = SparseMatrix (arg.array_value (),
                                       i, j, m, n, summation, nzmax);
            }
          else if (arg.islogical ())
            retval = SparseBoolMatrix (arg.bool_array_value (),
                                       i, j, m, n, summation, nzmax);
          else
            err_wrong_type_arg ("sparse", arg);
        }
      catch (index_exception& ie)
        {
          // Rethrow to allow more info to be reported later.
          ie.set_pos_if_unset (2, argidx+1);
          throw;
        }
    }

  return retval;
}

/*
## Tests for sparse constructor are in test/sparse.tst
%!assert (1)
*/

DEFUN (spalloc, args, ,
       doc: /* -*- texinfo -*-
@deftypefn {} {@var{s} =} spalloc (@var{m}, @var{n}, @var{nz})
Create an @var{m}-by-@var{n} sparse matrix with pre-allocated space for at
most @var{nz} nonzero elements.

This is useful for building a matrix incrementally by a sequence of indexed
assignments.  Subsequent indexed assignments after @code{spalloc} will reuse
the pre-allocated memory, provided they are of one of the simple forms

@itemize
@item @code{@var{s}(I:J) = @var{x}}

@item @code{@var{s}(:,I:J) = @var{x}}

@item @code{@var{s}(K:L,I:J) = @var{x}}
@end itemize

@b{and} that the following conditions are met:

@itemize
@item the assignment does not decrease @code{nnz (@var{S})}.

@item after the assignment, @code{nnz (@var{S})} does not exceed @var{nz}.

@item no index is out of bounds.
@end itemize

Partial movement of data may still occur, but in general the assignment will
be more memory and time efficient under these circumstances.  In particular,
it is possible to efficiently build a pre-allocated sparse matrix from a
contiguous block of columns.

The amount of pre-allocated memory for a given matrix may be queried using
the function @code{nzmax}.

Programming Note: Octave always reserves memory for at least one value,
even if @var{nz} is 0.
@seealso{nzmax, sparse}
@end deftypefn */)
{
  int nargin = args.length ();

  if (nargin < 2 || nargin > 3)
    print_usage ();

  octave_idx_type m = args(0).idx_type_value ();
  octave_idx_type n = args(1).idx_type_value ();

  octave_idx_type nz = 0;
  if (nargin == 3)
    nz = args(2).idx_type_value ();

  if (m < 0 || n < 0 || nz < 0)
    error ("spalloc: M, N, and NZ must be non-negative");

  return ovl (SparseMatrix (dim_vector (m, n), nz));
}

/*
%!assert (issparse (spalloc (1,1)))
%!assert (spalloc (1,1), sparse (1,1))
%!test
%! s = spalloc (1,1,5);
%! assert (s, sparse (1,1));
%! assert (nzmax (s), 5);
%!assert (spalloc (1,2), sparse (1,2))
%!assert (spalloc (1,2,2), sparse (1,2))
%!assert (spalloc (2,1), sparse (2,1))
%!assert (spalloc (2,1,2), sparse (2,1))

%!error spalloc ()
%!error spalloc (1)
%!error spalloc (1,2,3,4)
%!error <M, N, and NZ must be non-negative> spalloc (-1, 1, 1)
%!error <M, N, and NZ must be non-negative> spalloc (1, -1, 1)
%!error <M, N, and NZ must be non-negative> spalloc (1, 1, -1)
*/

OCTAVE_END_NAMESPACE(octave)