How to Simplify Passing Multiple Arguments from Structs in MATLAB Function Calls?

am working on a MATLAB project where I need to pass multiple arguments from two structs (n and d) to a function called mi_decode. Currently, I am passing each field individually, which makes the code quite verbose and hard to maintain. Here is a snippet of my current code:

function results = spectral_mi(labels, data, n, d)

d.pl_ u int8
d.prr_k cell
d.is_cnrm logical = true
d.is_chk logical = true
d.is_fbk logical = true
d.is_x logical = false
d.is_pl logical = false
d.is_prr logical = false

n.cl uint16 = numel(unique(labels))
n.t uint8 = size(data{1}, 2)
n.k uint8 = numel(data)
n.trials uint8 = numel(labels)
n.f uint8 = fix(size(data{1}, 2)/2)+1
n.pl uint16

% ... other code ...

pow_per_f = cell(1, n.f);
phase_per_f = cell(1, n.f);
for f = 1:n.f
    pow_per_f{f} = squeeze(power(abs(st_data(:, f, :)), 2));
    phase_per_f{f} = squeeze(angle(st_data(:, f, :)));
end
[mi_, zmi_, pmi_] = mi_decode( ...
    labels, pow_per_f{:}, ...
    "is_x", d.is_x, ...
    "t", n.t, ...
    "k", n.f, ...
    "pl", n.pl, ...
    "is_cnrm", d.is_cnrm, ...
    "is_chk", d.is_chk, ...
    "is_fbk", d.is_fbk, ...
    "is_pl", d.is_pl, ...
    "is_prr", d.is_prr);
[phase_mi_, phase_zmi_, phase_pmi_] = mi_decode( ...
    labels, phase_per_f{:}, ...
    "is_x", d.is_x, ...
    "t", n.t, ...
    "k", n.f, ...
    "pl", n.pl, ...
    "is_cnrm", d.is_cnrm, ...
    "is_chk", d.is_chk, ...
    "is_fbk", d.is_fbk, ...
    "is_pl", d.is_pl, ...
    "is_prr", d.is_prr);
end

Is there a more concise way to pass all the fields from the structs n and d to the mi_decode function without listing each field individually? Any suggestions or best practices for handling this in MATLAB would be greatly appreciated.

[Stephen23]

"Any suggestions or best practices for handling this in MATLAB"

Just pass the whole structure. Let the function decide what fields it needs to access.

[Ludwig]

How would this be done in practice? @Stephen23

If I'd try to pass the whole struct:

[mi_, zmi_, pmi_] = mi_decode( ...
    l_, pow_per_f{:}, "d", d, "n", n);

I get:

"Invalid argument at position 117. Value must be of type double or be convertible to double."

The whole argument list is:

arguments
l_ (:, 1) uint8
end
arguments (Repeating)
rr_k double
end
arguments
d.pl_ uint8
d.prr_k cell
d.is_cnrm logical = true
d.is_chk logical = true
d.is_fbk logical = true
d.is_x logical = false
d.is_pl logical = false
d.is_prr logical = false

n.cl uint16 = numel(unique(l_))
n.t uint8 = size(rr_k{1}, 2)
n.k uint8 = numel(rr_k)
n.trials uint8 = numel(l_)
n.f uint8 = fix(size(rr_k{1}, 2)/2)+1
n.pl uint16
end

[Stephen23]

"How would this be done in practice?"

You would have to rewrite the function to accept those scalar structures as inputs.

Note that INPUTPARSER can do this:

https://www.mathworks.com/help/matlab/ref/inputparser.html

or you could write the function itself to accept those structures and refer to their fields, exactly like spectral_mi is.

Answer 1

[Matt J]

You can use namedargs2cell and comma-separated list expansion to pass multiple name-value pairs,e.g.,

function main(struct_n,struct_d)

 cell_n=namedargs2cell(struct_n);
 cell_d=namedargs2cell(struct_d);

 someFunction(cell_n{:},cell_d{:}); 
 
end

function someFunction(n,d)

 arguments
    n.cl uint16 
    n.t uint8 
    n.k uint8
    n.trials uint8 
    n.f uint8 
    n.pl uint16
     
    d.pl_u int8
    d.prr_k cell
    d.is_cnrm logical = true
    d.is_chk logical = true
    d.is_fbk logical = true
    d.is_x logical = false
    d.is_pl logical = false
    d.is_prr logical = false
    
 end
 ...
end