How to test if a function handle belongs to a particular package?
4 ビュー (過去 30 日間)
古いコメントを表示
I have a method that accepts a function handle and then needs to branch depending on whether that function is part of +package1 or +package2. While I can think of some hacks involving parsing the function name and/or package's help text and/or output of dir, is there a better or simpler way?
thanks, -n
0 件のコメント
採用された回答
Guillaume
2017 年 4 月 28 日
Probably the simplest would be to parse the output of
functions(yourfunctionhandle)
The file field of that output will contain the full path to the anonymous function, so will contain the package name, if any.
5 件のコメント
Guillaume
2017 年 4 月 29 日
編集済み: per isakson
2017 年 12 月 21 日
Re: documentation. Well, at least it is documented in all versions, so if a change occurs you'll know about it. Saying that, I'm not aware of any major change to the behavior of functions. If you keep with your current design, this is certainly what I'd use.
As Steven says, that design sounds a bit iffy. The way I would implement your requirements would be with functions objects (i.e. classes). I would have an abstract base class from which all functions derive. It is therefore trivial to check whether a function is allowed with isa (or validateattribute) based on its class. To differentiate between the two kind of functions, I would have two additional abstract classes (deriving from the main abstract base class). The whole thing would be something like:
classdef processingfunction
methods (Abstract)
function result = dosomething(this, arg); %the actual method equivalent to the function handles
end
end
classdef likelihoodfunction < processingfunction
%does not need anything extra above the base class. Just a new type
end
classdef lossesfunction < processingfunction
%does not need anything extra above the base class. Just a new type
end
classdef someactualfunction < likelihoodfunction
methods
function result = dosomething(this, arg)
%actual implementation
result = arg + 2;
end
end
end
function find_best(myset, func)
validateattributes(fun, {'processingfunction'}, {'scalar'}); %error if func is not derived from processing function
if isa(func, 'lossesfunction') %switch behaviour depending of subtype
result = func.dosomething(1);
else
result = func.dosomething(2);
end
end
There are other advantages to using function objects, such as caching of inputs, precalculating expensive parameters and possibly others. The downside is that OOP in matlab is not always very fast.
その他の回答 (1 件)
Image Analyst
2017 年 4 月 29 日
As an example, to find out the folder of the bwdist function in the Image Processing Toolbox:
functionInfo = which('bwdist')
[folder, baseFileNameNoExt, ext] = fileparts(functionInfo)
Here is the file/folder info that is returned:
functionInfo =
'C:\Program Files\MATLAB\R2017a\toolbox\images\images\bwdist.m'
folder =
'C:\Program Files\MATLAB\R2017a\toolbox\images\images'
baseFileNameNoExt =
'bwdist'
ext =
'.m'
Adapt as needed for your function names.
5 件のコメント
Image Analyst
2017 年 4 月 29 日
I got the impression that he "needs to branch depending on whether that function is part of +package1 or +package2" so I figured that if he could figure out the folder where that function lives, then he'd know which package it belonged to. That was my thinking.
参考
カテゴリ
Help Center および File Exchange で Construct and Work with Object Arrays についてさらに検索
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!