Mercurial > hg > octave-nkf
view scripts/general/validateattributes.m @ 18754:0ede4dbb37f1
Overhaul interp1, interp2, interp3 functions.
* NEWS: Announce change in 'cubic' interpolation method for interp2
to match Matlab.
* bicubic.m: Use interp2 (..., "spline") in %!tests.
* interp1.m: Improve docstring. Use switch statement instead of if/elseif tree
for simpler code. Use more informative error message than 'table too short'.
Add titles to demo plots. Add new demo block showing difference between 'pchip'
and 'spline' methods.
* interp2.m: Rewrite docstring. Use variable 'extrap' instead of 'extrapval' to
match documentation. Use clearer messages in error() calls. Make 'cubic' use
the same algorithm as 'pchip' for Matlab compatibility. Use Octave coding
conventions regarding spaces between variable and parenthesis. Added input
validation tests.
* interp3.m: Rewrite docstring. Use clearer messages in error() calls. Make
'cubic' use the same algorithm as 'pchip' for Matlab compatibility. Simplify
input processing. Rewrite some %!tests for clarity. Added input validation
tests.
author | Rik <rik@octave.org> |
---|---|
date | Sun, 30 Mar 2014 14:18:43 -0700 |
parents | 2b01c11197d6 |
children | efd5cf93013b |
line wrap: on
line source
## Copyright (C) 2013 Carnë Draug ## ## 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/>. ## -*- texinfo -*- ## @deftypefn {Function File} {} validateattributes (@var{A}, @var{classes}, @var{attributes}) ## @deftypefnx {Function File} {} validateattributes (@var{A}, @var{classes}, @var{attributes}, @var{arg_idx}) ## @deftypefnx {Function File} {} validateattributes (@var{A}, @var{classes}, @var{attributes}, @var{func_name}) ## @deftypefnx {Function File} {} validateattributes (@var{A}, @var{classes}, @var{attributes}, @var{func_name}, @var{arg_name}) ## @deftypefnx {Function File} {} validateattributes (@var{A}, @var{classes}, @var{attributes}, @var{func_name}, @var{arg_name}, @var{arg_idx}) ## Check validity of input argument. ## ## Confirms that the argument @var{A} is valid by belonging to one of ## @var{classes}, and holding all of the @var{attributes}. If it does not, ## an error is thrown, with a message formatted accordingly. The error ## message can be made further complete by the function name @var{fun_name}, ## the argument name @var{arg_name}, and its position in the input ## @var{arg_idx}. ## ## @var{classes} must be a cell array of strings (an empty cell array is ## allowed) with the name of classes (remember that a class name is case ## sensitive). In addition to the class name, the following categories ## names are also valid: ## ## @table @asis ## @item @qcode{"float"} ## Floating point value comprising classes @qcode{"double"} and ## @qcode{"single"}. ## ## @item @qcode{"integer"} ## Integer value comprising classes (u)int8, (u)int16, (u)int32, (u)int64. ## ## @item @qcode{"numeric"} ## Numeric value comprising either a floating point or integer value. ## ## @end table ## ## @var{attributes} must be a cell array with names of checks for @var{A}. ## Some of them require an additional value to be supplied right after the ## name (see details for each below). ## ## @table @asis ## @item @qcode{"<="} ## All values are less than or equal to the following value in @var{attributes}. ## ## @item @qcode{"<"} ## All values are less than the following value in @var{attributes}. ## ## @item @qcode{">="} ## All values are greater than or equal to the following value in ## @var{attributes}. ## ## @item @qcode{">"} ## All values are greater than the following value in @var{attributes}. ## ## @item @qcode{"2d"} ## A 2-dimensional matrix. Note that vectors and empty matrices have ## 2 dimensions, one of them being of length 1, or both length 0. ## ## @item @qcode{"3d"} ## Has no more than 3 dimensions. A 2-dimensional matrix is a 3-D matrix ## whose 3rd dimension is of length 1. ## ## @item @qcode{"binary"} ## All values are either 1 or 0. ## ## @item @qcode{"column"} ## Values are arranged in a single column. ## ## @item @qcode{"decreasing"} ## No value is @var{NaN}, and each is less than the preceding one. ## ## @item @qcode{"even"} ## All values are even numbers. ## ## @item @qcode{"finite"} ## All values are finite. ## ## @item @qcode{"increasing"} ## No value is @var{NaN}, and each is greater than the preceding one. ## ## @item @qcode{"integer"} ## All values are integer. This is different than using @code{isinteger} ## which only checks its an integer type. This checks that each value in ## @var{A} is an integer value, i.e., it has no decimal part. ## ## @item @qcode{"ncols"} ## Has exactly as many columns as the next value in @var{attributes}. ## ## @item @qcode{"ndims"} ## Has exactly as many dimensions as the next value in @var{attributes}. ## ## @item @qcode{"nondecreasing"} ## No value is @var{NaN}, and each is greater than or equal to the preceding ## one. ## ## @item @qcode{"nonempty"} ## It is not empty. ## ## @item @qcode{"nonincreasing"} ## No value is @var{NaN}, and each is less than or equal to the preceding one. ## ## @item @qcode{"nonnan"} ## No value is a @code{NaN}. ## ## @item @qcode{"non-negative"} ## All values are non negative. ## ## @item @qcode{"nonsparse"} ## It is not a sparse matrix. ## ## @item @qcode{"nonzero"} ## No value is zero. ## ## @item @qcode{"nrows"} ## Has exactly as many rows as the next value in @var{attributes}. ## ## @item @qcode{"numel"} ## Has exactly as many elements as the next value in @var{attributes}. ## ## @item @qcode{"odd"} ## All values are odd numbers. ## ## @item @qcode{"positive"} ## All values are positive. ## ## @item @qcode{"real"} ## It is a non-complex matrix. ## ## @item @qcode{"row"} ## Values are arranged in a single row. ## ## @item @qcode{"scalar"} ## It is a scalar. ## ## @item @qcode{"size"} ## Its size has length equal to the values of the next in @var{attributes}. ## The next value must is an array with the length for each dimension. To ## ignore the check for a certain dimension, the value of @code{NaN} can be ## used. ## ## @item @qcode{"square"} ## Is a square matrix. ## ## @item @qcode{"vector"} ## Values are arranged in a single vector (column or vector). ## ## @end table ## ## @seealso{nargin, varargin, isa, validatestring} ## @end deftypefn function validateattributes (A, cls, attr, varargin) if (nargin < 3 || nargin > 6) print_usage (); elseif (! iscellstr (cls)) error ("validateattributes: CLASSES must be a cell array of strings"); elseif (! iscell (attr)) error ("validateattributes: ATTRIBUTES must be a cell array"); endif ## Built start of error message from the extra optional arguments func_name = ""; var_name = "input"; if (nargin > 3) fourth = varargin{1}; if (ischar (fourth)) func_name = [fourth ": "]; elseif (nargin == 4 && valid_arg_idx (fourth)) var_name = sprintf ("input %d", fourth); else error ("validateattributes: 4th input argument must be ARG_IDX or FUNC_NAME"); endif if (nargin > 4) var_name = varargin{2}; if (! ischar (var_name)) error ("validateattributes: VAR_NAME must be a string"); endif if (nargin > 5) arg_idx = varargin{3}; if (! valid_arg_idx (arg_idx)) error ("validateattributes: ARG_IDX must be a positive integer"); endif var_name = sprintf ("%s (argument #%i)", var_name, arg_idx); endif endif endif err_ini = [func_name var_name]; check_cl = isa (A, cls); if (! isempty (check_cl) && ! any (check_cl)) ## Allowing for an empty list of classes is Matlab incompatible but ## that should count as a just a Matlab bug, not an incompatibility. ## Replace the category names with the classes that belong to it. integer = { "int8" "int16" "int32" "int64" ... "uint8" "uint16" "uint32" "uint64"}; float = {"single" "double"}; numeric = {integer{:} float{:}}; cls = replace_cl_group (cls, "integer", integer); cls = replace_cl_group (cls, "float", float ); cls = replace_cl_group (cls, "numeric", numeric); cls = unique (cls); classes = sprintf (" %s", cls{:}); error ("%s must be of class:\n\n %s\n\nbut was of class %s", err_ini, classes, class (A)); endif ## We use a while loop because some attributes require the following value ## in the cell array. Also, we can't just get the boolean value for the ## test and check at the end the error message since some of the tests ## require some more complex error message. ## It may look like that we don't perform enough input check in this ## function (e.g., we don't check if there's a value after the size ## attribute). The reasoning is that this will be a function mostly used ## by developers with fairly static input so any problem would be caught ## immediately during that functino development, it's no dependent on the ## final user input. In addition, it can be called so many times at the ## start of every function, we want it to run specially fast. idx = 1; problem = false; # becomes true when one of the tests fails while (idx <= numel (attr)) ## TODO: once we use this in Octave core, it might be worthy to find ## which attributes are checked more often, and place them in that ## order inside the switch block. switch (tolower (attr{idx++})) case "2d", problem = ndims (A) != 2; case "3d", problem = ndims (A) > 3; case "column", problem = ! iscolumn (A); case "row", problem = ! isrow (A); case "scalar", problem = ! isscalar (A); case "vector", problem = ! isvector (A); case "square", problem = ! issquare (A); case "nonempty", problem = isempty (A); case "nonsparse", problem = issparse (A); case "binary", problem = ! islogical (A) && ... any ((A(:) != 1) & (A(:) != 0)); case "even", problem = any (rem (A(:), 2) != 0); case "odd", problem = any (mod (A(:), 2) != 1); case "integer", problem = ! isinteger (A) && ... any (ceil (A(:)) != A(:)); case "real", problem = ! isreal (A); case "finite", problem = ! isinteger (A) && ... ! all (isfinite (A(:))); case "nonnan", problem = ! isinteger (A) && ... any (isnan (A(:))); case "nonnegative", problem = any (A(:) < 0); case "nonzero", problem = any (A(:) == 0); case "positive", problem = any (A(:) <= 0); case "decreasing", problem = (any (isnan (A(:))) || any (diff (A(:)) >= 0)); case "increasing", problem = (any (isnan (A(:))) || any (diff (A(:)) <= 0)); case "nondecreasing", problem = (any (isnan (A(:))) || any (diff (A(:)) < 0)); case "nonincreasing", problem = (any (isnan (A(:))) || any (diff (A(:)) > 0)); case "size", A_size = size (A); w_size = attr{idx++}; A_size(isnan (w_size)) = NaN; if (! isequaln (A_size, w_size)) A_size_str = sprintf ("%dx", size (A))(1:end-1); w_size_str = sprintf ("%ix", w_size)(1:end-1); w_size_str = strrep (w_size_str, "NaN", "N"); error ("%s must be of size %s but was %s", err_ini, w_size_str, A_size_str); endif case "numel", if (numel (A) != attr{idx++}) error ("%s must have %d elements", err_ini, attr{idx-1}); endif case "ncols", if (columns (A) != attr{idx++}) error ("%s must have %d columns", err_ini, attr{idx-1}); endif case "nrows", if (rows (A) != attr{idx++}) error ("%s must have %d rows", err_ini, attr{idx-1}); endif case "ndims", ## Note that a [4 5 1] matrix is not considered to have ndims == 3 ## but is ok for "3d". This is not a bug. if (ndims (A) != attr{idx++}) error ("%s must have %d dimensions", err_ini, attr{idx-1}); endif case ">" if (! all (A(:) > attr{idx++})) error ("%s must be greater than %f", err_ini, attr{idx-1}); endif case ">=" if (! all (A(:) >= attr{idx++})) error ("%s must be greater than or equal to %f", err_ini, attr{idx-1}); endif case "<" if (! all (A(:) < attr{idx++})) error ("%s must be less than %f", err_ini, attr{idx-1}); endif case "<=" if (! all (A(:) <= attr{idx++})) error ("%s must be less than or equal to %f", err_ini, attr{idx-1}); endif otherwise error ("validateattributes: unknown ATTRIBUTE %s", attr{idx-1}); endswitch if (problem) error ("%s must be %s", err_ini, attr{idx-1}); endif endwhile endfunction function retval = valid_arg_idx (arg) retval = isnumeric (arg) && isscalar (arg) && arg > 0 && ceil (arg) == arg; endfunction function cls = replace_cl_group (cls, name, group) num_pos = strcmpi (cls, name); if (any (num_pos)) cls(num_pos) = []; cls(end+1:end+numel(group)) = group; endif endfunction %!error <double> validateattributes (rand (5), {"uint8"}, {}) %!error <single> validateattributes (uint8 (rand (5)), {"float"}, {}) %!error <2d> validateattributes (rand (5, 5, 5), {}, {"2d"}) %!error <3d> validateattributes (rand (5, 5, 5, 7), {}, {"3d"}) %!error <column> validateattributes (rand (5, 5), {}, {"column"}) %!error <column> validateattributes (rand (1, 5), {}, {"column"}) %!error <row> validateattributes (rand (5, 5), {}, {"row"}) %!error <row> validateattributes (rand (5, 1), {}, {"row"}) %!error <scalar> validateattributes (rand (1, 5), {}, {"scalar"}) %!error <vector> validateattributes (rand (5), {}, {"vector"}) %!error <square> validateattributes (rand (5, 6), {}, {"square"}) %!error <nonempty> validateattributes ([], {}, {"nonempty"}) %!error <nonsparse> validateattributes (sparse(rand(5)), {}, {"nonsparse"}) %!error <binary> validateattributes ("text", {}, {"binary"}) %!error <binary> validateattributes ([0 1 0 3 0], {}, {"binary"}) %!error <even> validateattributes ([2 3 6 8], {}, {"even"}) %!error <even> validateattributes ([2 NaN], {}, {"even"}) %!error <odd> validateattributes ([3 4 7 5], {}, {"odd"}) %!error <odd> validateattributes ([5 NaN], {}, {"odd"}) %!error <integer> validateattributes ([5 5.2 5.7], {}, {"integer"}) %!error <real> validateattributes ([5i 8 9], {}, {"real"}) %!error <finite> validateattributes ([5i Inf 8], {}, {"finite"}) %!error <nonnan> validateattributes ([NaN Inf 8], {}, {"nonnan"}) %!error <nonnegative> validateattributes ([7 8 -9], {}, {"nonnegative"}) %!error <nonzero> validateattributes ([7 8 0], {}, {"nonzero"}) %!error <positive> validateattributes ([7 0 8], {}, {"positive"}) %!error <decreasing> validateattributes ([7 8 4 3 -5], {}, {"decreasing"}) %!error <decreasing> validateattributes ([7 NaN 4 3 -5], {}, {"decreasing"}) %!error <increasing> validateattributes ([7 8 4 9 20], {}, {"increasing"}) %!error <increasing> validateattributes ([7 8 NaN 9 20], {}, {"increasing"}) %!error <nonincreasing> validateattributes ([7 8 4 9 20], {}, {"nonincreasing"}) %!error <nonincreasing> validateattributes ([7 8 NaN 9 20], {}, {"nonincreasing"}) %!error <nondecreasing> validateattributes ([7 8 4 3 -5], {}, {"nondecreasing"}) %!error <nondecreasing> validateattributes ([7 NaN 4 3 -5], {}, {"nondecreasing"}) %!error <size> validateattributes (ones (5, 3, 6), {}, {"size", [5 4 7]}) %!error <size> validateattributes (ones (5, 3, 6), {}, {"size", [5 NaN 7]}) %!error <size> validateattributes (ones (5, 3, 6), {}, {"size", [5 3 6 2]}) %!error <elements> validateattributes (ones (6, 3), {}, {"numel", 12}) %!error <columns> validateattributes (ones (6, 2), {}, {"ncols", 3}) %!error <rows> validateattributes (ones (6, 2), {}, {"nrows", 3}) %!error <dimensions> validateattributes (ones (6, 2, 6, 3), {}, {"ndims", 3}) %!error <greater than> validateattributes ([6 7 8 5], {}, {">", 5}) %!error <greater than> validateattributes ([6 7 8 5], {}, {">=", 6}) %!error <less than> validateattributes ([6 7 8 5], {}, {"<", 8}) %!error <less than> validateattributes ([6 7 8 5], {}, {"<=", 7}) %!test %! validateattributes (rand (5), {"numeric"}, {}) %! validateattributes (rand (5), {"float"}, {}) %! validateattributes (rand (5), {"double"}, {}) %! validateattributes ("text", {"char"}, {}) %! validateattributes (rand (5), {}, {"2d"}) %! validateattributes (rand (5), {}, {"3d"}) %! validateattributes (rand (5, 5, 5), {}, {"3d"}) %! validateattributes (rand (5, 1), {}, {"column"}) %! validateattributes (rand (1, 5), {}, {"row"}) %! validateattributes ("a", {}, {"scalar"}) %! validateattributes (5, {}, {"scalar"}) %! validateattributes (rand (1, 5), {}, {"vector"}) %! validateattributes (rand (5, 1), {}, {"vector"}) %! validateattributes (rand (5), {}, {"square"}) %! validateattributes (rand (5), {}, {"nonempty"}) %! validateattributes (rand (5), {}, {"nonsparse"}) %! validateattributes ([0 1 0 1 0], {}, {"binary"}) %! validateattributes (rand (5) > 0.5, {}, {"binary"}) %! validateattributes ([8 4 0 6], {}, {"even"}) %! validateattributes ([-1 3 5], {}, {"odd"}) %! validateattributes ([8 4 0 6], {}, {"real"}) %! validateattributes ([8 4i 0 6], {}, {"finite"}) %! validateattributes (uint8 ([8 4]), {}, {"finite"}) %! validateattributes ([8 Inf], {}, {"nonnan"}) %! validateattributes ([0 7 4], {}, {"nonnegative"}) %! validateattributes ([-8 7 4], {}, {"nonzero"}) %! validateattributes ([8 7 4], {}, {"positive"}) %! validateattributes ([8 7 4 -5], {}, {"decreasing"}) %! validateattributes ([-8 -7 4 5], {}, {"increasing"}) %! validateattributes ([8 4 4 -5], {}, {"nonincreasing"}) %! validateattributes ([-8 -8 4 5], {}, {"nondecreasing"}) %! validateattributes (rand (4, 6, 7, 2), {}, {"size", [4 6 7 2]}) %! validateattributes (rand (4, 6, 7, 2), {}, {"size", [4 NaN 7 2]}) %! validateattributes (rand (4, 6, 7, 2), {}, {"size", [4 6 NaN 2 NaN]}) %! validateattributes (rand (6, 2), {}, {"numel", 12}) %! validateattributes (rand (6, 2), {}, {"ncols", 2}) %! validateattributes (rand (6, 2), {}, {"nrows", 6}) %! validateattributes (rand (6, 2, 4, 5), {}, {"ndims", 4}) %! validateattributes ([4 5 6 7], {}, {">", 3}) %! validateattributes ([4 5 6 7], {}, {">=", 4}) %! validateattributes ([4 5 6 7], {}, {"<", 8}) %! validateattributes ([4 5 6 7], {}, {"<=", 7}) %!test %! validateattributes ([0 1 0 1], {"double", "uint8"}, {"binary", "size", [NaN 4], "nonnan"})