view libinterp/corefcn/lookup.cc @ 21200:fcac5dbbf9ed

maint: Indent #ifdef blocks in libinterp. * builtins.h, Cell.cc, __contourc__.cc, __dispatch__.cc, __dsearchn__.cc, __ichol__.cc, __ilu__.cc, __lin_interpn__.cc, __pchip_deriv__.cc, __qp__.cc, balance.cc, besselj.cc, betainc.cc, bitfcns.cc, bsxfun.cc, c-file-ptr-stream.cc, c-file-ptr-stream.h, cellfun.cc, colloc.cc, comment-list.cc, conv2.cc, daspk.cc, dasrt.cc, dassl.cc, data.cc, debug.cc, defaults.cc, defaults.in.h, defun-dld.h, defun.cc, defun.h, det.cc, dirfns.cc, display.cc, dlmread.cc, dot.cc, dynamic-ld.cc, eig.cc, ellipj.cc, error.cc, errwarn.cc, event-queue.cc, fft.cc, fft2.cc, fftn.cc, file-io.cc, filter.cc, find.cc, gammainc.cc, gcd.cc, getgrent.cc, getpwent.cc, getrusage.cc, givens.cc, gl-render.cc, gl2ps-print.cc, graphics.cc, graphics.in.h, gripes.cc, hash.cc, help.cc, hess.cc, hex2num.cc, input.cc, inv.cc, jit-ir.cc, jit-typeinfo.cc, jit-util.cc, jit-util.h, kron.cc, load-path.cc, load-save.cc, lookup.cc, ls-ascii-helper.cc, ls-hdf5.cc, ls-mat-ascii.cc, ls-mat4.cc, ls-mat5.cc, ls-oct-binary.cc, ls-oct-text.cc, ls-oct-text.h, ls-utils.cc, ls-utils.h, lsode.cc, lu.cc, luinc.cc, mappers.cc, matrix_type.cc, max.cc, mex.h, mexproto.h, mgorth.cc, nproc.cc, oct-errno.in.cc, oct-fstrm.cc, oct-hdf5-types.cc, oct-hdf5.h, oct-hist.cc, oct-iostrm.cc, oct-lvalue.cc, oct-map.cc, oct-prcstrm.cc, oct-procbuf.cc, oct-stream.cc, oct-strstrm.cc, octave-link.cc, ordschur.cc, pager.cc, pinv.cc, pr-output.cc, procstream.cc, profiler.cc, psi.cc, pt-jit.cc, quad.cc, quadcc.cc, qz.cc, rand.cc, rcond.cc, regexp.cc, schur.cc, sighandlers.cc, sparse-xdiv.cc, sparse-xpow.cc, sparse.cc, spparms.cc, sqrtm.cc, str2double.cc, strfind.cc, strfns.cc, sub2ind.cc, svd.cc, sylvester.cc, symtab.cc, syscalls.cc, sysdep.cc, sysdep.h, time.cc, toplev.cc, tril.cc, tsearch.cc, txt-eng-ft.cc, txt-eng.cc, typecast.cc, urlwrite.cc, utils.cc, variables.cc, xdiv.cc, xnorm.cc, xpow.cc, zfstream.cc, __delaunayn__.cc, __eigs__.cc, __fltk_uigetfile__.cc, __glpk__.cc, __init_fltk__.cc, __init_gnuplot__.cc, __magick_read__.cc, __osmesa_print__.cc, __voronoi__.cc, amd.cc, audiodevinfo.cc, audioread.cc, ccolamd.cc, chol.cc, colamd.cc, convhulln.cc, dmperm.cc, fftw.cc, oct-qhull.h, qr.cc, symbfact.cc, symrcm.cc, oct-conf.in.cc, ov-base-diag.cc, ov-base-int.cc, ov-base-mat.cc, ov-base-scalar.cc, ov-base-sparse.cc, ov-base.cc, ov-bool-mat.cc, ov-bool-sparse.cc, ov-bool.cc, ov-builtin.cc, ov-cell.cc, ov-ch-mat.cc, ov-class.cc, ov-classdef.cc, ov-colon.cc, ov-complex.cc, ov-cs-list.cc, ov-cx-diag.cc, ov-cx-mat.cc, ov-cx-sparse.cc, ov-dld-fcn.cc, ov-fcn-handle.cc, ov-fcn-inline.cc, ov-fcn.cc, ov-float.cc, ov-flt-complex.cc, ov-flt-cx-diag.cc, ov-flt-cx-mat.cc, ov-flt-re-diag.cc, ov-flt-re-mat.cc, ov-int16.cc, ov-int32.cc, ov-int64.cc, ov-int8.cc, ov-java.cc, ov-lazy-idx.cc, ov-mex-fcn.cc, ov-null-mat.cc, ov-oncleanup.cc, ov-perm.cc, ov-range.cc, ov-re-diag.cc, ov-re-mat.cc, ov-re-sparse.cc, ov-scalar.cc, ov-str-mat.cc, ov-struct.cc, ov-typeinfo.cc, ov-uint16.cc, ov-uint32.cc, ov-uint64.cc, ov-uint8.cc, ov-usr-fcn.cc, ov.cc, ovl.cc, octave.cc, op-b-b.cc, op-b-bm.cc, op-b-sbm.cc, op-bm-b.cc, op-bm-bm.cc, op-bm-sbm.cc, op-cdm-cdm.cc, op-cell.cc, op-chm.cc, op-class.cc, op-cm-cm.cc, op-cm-cs.cc, op-cm-m.cc, op-cm-s.cc, op-cm-scm.cc, op-cm-sm.cc, op-cs-cm.cc, op-cs-cs.cc, op-cs-m.cc, op-cs-s.cc, op-cs-scm.cc, op-cs-sm.cc, op-dm-dm.cc, op-dm-scm.cc, op-dm-sm.cc, op-dm-template.cc, op-dms-template.cc, op-double-conv.cc, op-fcdm-fcdm.cc, op-fcdm-fdm.cc, op-fcm-fcm.cc, op-fcm-fcs.cc, op-fcm-fm.cc, op-fcm-fs.cc, op-fcn.cc, op-fcs-fcm.cc, op-fcs-fcs.cc, op-fcs-fm.cc, op-fcs-fs.cc, op-fdm-fdm.cc, op-float-conv.cc, op-fm-fcm.cc, op-fm-fcs.cc, op-fm-fm.cc, op-fm-fs.cc, op-fs-fcm.cc, op-fs-fcs.cc, op-fs-fm.cc, op-fs-fs.cc, op-i16-i16.cc, op-i32-i32.cc, op-i64-i64.cc, op-i8-i8.cc, op-int-concat.cc, op-int-conv.cc, op-m-cm.cc, op-m-cs.cc, op-m-m.cc, op-m-s.cc, op-m-scm.cc, op-m-sm.cc, op-pm-pm.cc, op-pm-scm.cc, op-pm-sm.cc, op-pm-template.cc, op-range.cc, op-s-cm.cc, op-s-cs.cc, op-s-m.cc, op-s-s.cc, op-s-scm.cc, op-s-sm.cc, op-sbm-b.cc, op-sbm-bm.cc, op-sbm-sbm.cc, op-scm-cm.cc, op-scm-cs.cc, op-scm-m.cc, op-scm-s.cc, op-scm-scm.cc, op-scm-sm.cc, op-sm-cm.cc, op-sm-cs.cc, op-sm-m.cc, op-sm-s.cc, op-sm-scm.cc, op-sm-sm.cc, op-str-m.cc, op-str-s.cc, op-str-str.cc, op-struct.cc, op-ui16-ui16.cc, op-ui32-ui32.cc, op-ui64-ui64.cc, op-ui8-ui8.cc, pt-arg-list.cc, pt-array-list.cc, pt-assign.cc, pt-binop.cc, pt-bp.cc, pt-cbinop.cc, pt-cell.cc, pt-check.cc, pt-classdef.cc, pt-cmd.cc, pt-colon.cc, pt-colon.h, pt-const.cc, pt-decl.cc, pt-eval.cc, pt-except.cc, pt-exp.cc, pt-fcn-handle.cc, pt-funcall.cc, pt-id.cc, pt-idx.cc, pt-jump.cc, pt-loop.cc, pt-mat.cc, pt-misc.cc, pt-pr-code.cc, pt-select.cc, pt-stmt.cc, pt-unop.cc, pt.cc, token.cc, Array-jit.cc, Array-os.cc, Array-sym.cc, Array-tc.cc, version.cc: Indent #ifdef blocks in libinterp.
author Rik <rik@octave.org>
date Fri, 05 Feb 2016 16:29:08 -0800
parents 538b57866b90
children 40de9f8f23a6
line wrap: on
line source

/*

Copyright (C) 2008-2015 VZLU Prague a.s., Czech Republic

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/>.

*/

// Author: Jaroslav Hajek <highegg@gmail.com>

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

#include <cctype>
#include <functional>
#include <algorithm>

#include "dNDArray.h"
#include "CNDArray.h"

#include "Cell.h"
#include "defun.h"
#include "error.h"
#include "errwarn.h"
#include "ovl.h"
#include "ov.h"

static
bool
contains_char (const std::string& str, char c)
{
  return (str.find (c) != std::string::npos
          || str.find (std::toupper (c)) != std::string::npos);
}

// case-insensitive character comparison functors
struct icmp_char_lt : public std::binary_function<char, char, bool>
{
  bool operator () (char x, char y) const
  { return std::toupper (x) < std::toupper (y); }
};

struct icmp_char_gt : public std::binary_function<char, char, bool>
{
  bool operator () (char x, char y) const
  { return std::toupper (x) > std::toupper (y); }
};

// FIXME: maybe these should go elsewhere?
// FIXME: are they even needed now?
// case-insensitive ascending comparator
#if 0
static bool
stri_comp_lt (const std::string& a, const std::string& b)
{
  return std::lexicographical_compare (a.begin (), a.end (),
                                       b.begin (), b.end (),
                                       icmp_char_lt ());
}

// case-insensitive descending comparator
static bool
stri_comp_gt (const std::string& a, const std::string& b)
{
  return std::lexicographical_compare (a.begin (), a.end (),
                                       b.begin (), b.end (),
                                       icmp_char_gt ());
}
#endif

template <typename T>
inline sortmode
get_sort_mode (const Array<T>& array,
               typename octave_sort<T>::compare_fcn_type desc_comp
               = octave_sort<T>::descending_compare)
{
  octave_idx_type n = array.numel ();
  if (n > 1 && desc_comp (array (0), array (n-1)))
    return DESCENDING;
  else
    return ASCENDING;
}

// FIXME: perhaps there should be octave_value::lookup?
// The question is, how should it behave w.r.t. the second argument's type.
// We'd need a dispatch on two arguments. Hmmm...

#define INT_ARRAY_LOOKUP(TYPE) \
  (table.is_ ## TYPE ## _type () && y.is_ ## TYPE ## _type ()) \
    retval = do_numeric_lookup (table.TYPE ## _array_value (), \
                                y.TYPE ## _array_value (), \
                                left_inf, right_inf, \
                                match_idx, match_bool);
template <typename ArrayT>
static octave_value
do_numeric_lookup (const ArrayT& array, const ArrayT& values,
                   bool left_inf, bool right_inf,
                   bool match_idx, bool match_bool)
{
  octave_value retval;

  Array<octave_idx_type> idx = array.lookup (values);
  octave_idx_type n = array.numel ();
  octave_idx_type nval = values.numel ();

  // Post-process.
  if (match_bool)
    {
      boolNDArray match (idx.dims ());
      for (octave_idx_type i = 0; i < nval; i++)
        {
          octave_idx_type j = idx.xelem (i);
          match.xelem (i) = j != 0 && values(i) == array(j-1);
        }

      retval = match;
    }
  else if (match_idx || left_inf || right_inf)
    {
      if (match_idx)
        {
          NDArray ridx (idx.dims ());

          for (octave_idx_type i = 0; i < nval; i++)
            {
              octave_idx_type j = idx.xelem (i);
              ridx.xelem (i) = (j != 0 && values(i) == array(j-1)) ? j : 0;
            }

          retval = ridx;
        }
      else if (left_inf && right_inf)
        {
          // Results in valid indices. Optimize using lazy index.
          octave_idx_type zero = 0;
          for (octave_idx_type i = 0; i < nval; i++)
            {
              octave_idx_type j = idx.xelem (i) - 1;
              idx.xelem (i) = std::max (zero, std::min (j, n-2));
            }

          retval = idx_vector (idx);
        }
      else if (left_inf)
        {
          // Results in valid indices. Optimize using lazy index.
          octave_idx_type zero = 0;
          for (octave_idx_type i = 0; i < nval; i++)
            {
              octave_idx_type j = idx.xelem (i) - 1;
              idx.xelem (i) = std::max (zero, j);
            }

          retval = idx_vector (idx);
        }
      else if (right_inf)
        {
          NDArray ridx (idx.dims ());

          for (octave_idx_type i = 0; i < nval; i++)
            {
              octave_idx_type j = idx.xelem (i);
              ridx.xelem (i) = std::min (j, n-1);
            }

          retval = ridx;
        }
    }
  else
    retval = idx;

  return retval;
}

DEFUN (lookup, args, ,
       "-*- texinfo -*-\n\
@deftypefn  {} {@var{idx} =} lookup (@var{table}, @var{y})\n\
@deftypefnx {} {@var{idx} =} lookup (@var{table}, @var{y}, @var{opt})\n\
Lookup values in a sorted table.\n\
\n\
This function is usually used as a prelude to interpolation.\n\
\n\
If table is increasing and @code{idx = lookup (table, y)}, then\n\
@code{table(idx(i)) <= y(i) < table(idx(i+1))} for all @code{y(i)} within\n\
the table.  If @code{y(i) < table(1)} then @code{idx(i)} is 0.  If\n\
@code{y(i) >= table(end)} or @code{isnan (y(i))} then @code{idx(i)} is\n\
@code{n}.\n\
\n\
If the table is decreasing, then the tests are reversed.\n\
For non-strictly monotonic tables, empty intervals are always skipped.\n\
The result is undefined if @var{table} is not monotonic, or if\n\
@var{table} contains a NaN.\n\
\n\
The complexity of the lookup is O(M*log(N)) where N is the size of\n\
@var{table} and M is the size of @var{y}.  In the special case when @var{y}\n\
is also sorted, the complexity is O(min(M*log(N),M+N)).\n\
\n\
@var{table} and @var{y} can also be cell arrays of strings\n\
(or @var{y} can be a single string).  In this case, string lookup\n\
is performed using lexicographical comparison.\n\
\n\
If @var{opts} is specified, it must be a string with letters indicating\n\
additional options.\n\
\n\
@table @code\n\
@item m\n\
@code{table(idx(i)) == val(i)} if @code{val(i)}\n\
occurs in table; otherwise, @code{idx(i)} is zero.\n\
\n\
@item b\n\
@code{idx(i)} is a logical 1 or 0, indicating whether\n\
@code{val(i)} is contained in table or not.\n\
\n\
@item l\n\
For numeric lookups\n\
the leftmost subinterval shall be extended to infinity (i.e., all indices\n\
at least 1)\n\
\n\
@item r\n\
For numeric lookups\n\
the rightmost subinterval shall be extended to infinity (i.e., all indices\n\
at most n-1).\n\
@end table\n\
@end deftypefn")
{
  int nargin = args.length ();

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

  octave_value table = args(0);
  octave_value y = args(1);
  if (table.ndims () > 2 || (table.columns () > 1 && table.rows () > 1))
    warning ("lookup: table is not a vector");

  octave_value retval;

  bool num_case = ((table.is_numeric_type () && y.is_numeric_type ())
                   || (table.is_char_matrix () && y.is_char_matrix ()));
  bool str_case = table.is_cellstr () && (y.is_string () || y.is_cellstr ());
  bool left_inf = false;
  bool right_inf = false;
  bool match_idx = false;
  bool match_bool = false;

  if (nargin == 3)
    {
      std::string opt = args(2).xstring_value ("lookup: OPT must be a string");
      left_inf = contains_char (opt, 'l');
      right_inf = contains_char (opt, 'r');
      match_idx = contains_char (opt, 'm');
      match_bool = contains_char (opt, 'b');
      if (opt.find_first_not_of ("lrmb") != std::string::npos)
        error ("lookup: unrecognized option: %c",
               opt[opt.find_first_not_of ("lrmb")]);
    }

  if ((match_idx || match_bool) && (left_inf || right_inf))
    error ("lookup: m, b cannot be specified with l or r");
  else if (match_idx && match_bool)
    error ("lookup: only one of m or b can be specified");
  else if (str_case && (left_inf || right_inf))
    error ("lookup: l, r are not recognized for string lookups");

  if (num_case)
    {
      // In the case of a complex array, absolute values will be used for
      // compatibility (though it's not too meaningful).
      if (table.is_complex_type ())
        table = table.abs ();

      if (y.is_complex_type ())
        y = y.abs ();

      Array<octave_idx_type> idx;

      // PS: I learned this from data.cc
      if INT_ARRAY_LOOKUP (int8)
      else if INT_ARRAY_LOOKUP (int16)
      else if INT_ARRAY_LOOKUP (int32)
      else if INT_ARRAY_LOOKUP (int64)
      else if INT_ARRAY_LOOKUP (uint8)
      else if INT_ARRAY_LOOKUP (uint16)
      else if INT_ARRAY_LOOKUP (uint32)
      else if INT_ARRAY_LOOKUP (uint64)
      else if (table.is_char_matrix () && y.is_char_matrix ())
        retval = do_numeric_lookup (table.char_array_value (),
                                    y.char_array_value (),
                                    left_inf, right_inf,
                                    match_idx, match_bool);
      else if (table.is_single_type () || y.is_single_type ())
        retval = do_numeric_lookup (table.float_array_value (),
                                    y.float_array_value (),
                                    left_inf, right_inf,
                                    match_idx, match_bool);
      else
        retval = do_numeric_lookup (table.array_value (),
                                    y.array_value (),
                                    left_inf, right_inf,
                                    match_idx, match_bool);
    }
  else if (str_case)
    {
      Array<std::string> str_table = table.cellstr_value ();
      Array<std::string> str_y (dim_vector (1, 1));

      if (y.is_cellstr ())
        str_y = y.cellstr_value ();
      else
        str_y(0) = y.string_value ();

      Array<octave_idx_type> idx = str_table.lookup (str_y);
      octave_idx_type nval = str_y.numel ();

      // Post-process.
      if (match_bool)
        {
          boolNDArray match (idx.dims ());
          for (octave_idx_type i = 0; i < nval; i++)
            {
              octave_idx_type j = idx.xelem (i);
              match.xelem (i) = j != 0 && str_y(i) == str_table(j-1);
            }

          retval = match;
        }
      else if (match_idx)
        {
          NDArray ridx (idx.dims ());
          if (match_idx)
            {
              for (octave_idx_type i = 0; i < nval; i++)
                {
                  octave_idx_type j = idx.xelem (i);
                  ridx.xelem (i) = (j != 0 && str_y(i) == str_table(j-1)) ? j
                                                                          : 0;
                }
            }

          retval = ridx;
        }
      else
        retval = idx;
    }
  else
    print_usage ();

  return retval;
}

/*
%!assert (lookup (1:3, 0.5), 0)     # value before table
%!assert (lookup (1:3, 3.5), 3)     # value after table error
%!assert (lookup (1:3, 1.5), 1)     # value within table error
%!assert (lookup (1:3, [3,2,1]), [3,2,1])
%!assert (lookup ([1:4]', [1.2, 3.5]'), [1, 3]')
%!assert (lookup ([1:4], [1.2, 3.5]'), [1, 3]')
%!assert (lookup ([1:4]', [1.2, 3.5]), [1, 3])
%!assert (lookup ([1:4], [1.2, 3.5]), [1, 3])
%!assert (lookup (1:3, [3, 2, 1]), [3, 2, 1])
%!assert (lookup ([3:-1:1], [3.5, 3, 1.2, 2.5, 2.5]), [0, 1, 2, 1, 1])
%!assert (isempty (lookup ([1:3], [])))
%!assert (isempty (lookup ([1:3]', [])))
%!assert (lookup (1:3, [1, 2; 3, 0.5]), [1, 2; 3, 0])
%!assert (lookup (1:4, [1, 1.2; 3, 2.5], "m"), [1, 0; 3, 0])
%!assert (lookup (4:-1:1, [1, 1.2; 3, 2.5], "m"), [4, 0; 2, 0])
%!assert (lookup (1:4, [1, 1.2; 3, 2.5], "b"), logical ([1, 0; 3, 0]))
%!assert (lookup (4:-1:1, [1, 1.2; 3, 2.5], "b"), logical ([4, 0; 2, 0]))
%!
%!assert (lookup ({"apple","lemon","orange"}, {"banana","kiwi"; "ananas","mango"}), [1,1;0,2])
%!assert (lookup ({"apple","lemon","orange"}, "potato"), 3)
%!assert (lookup ({"orange","lemon","apple"}, "potato"), 0)
*/