return multiple variable from matlab to python

It is easy to return one variable from Matlab using

res = mlab.run_func('path/to/jk.m', {'arg1': 3, 'arg2': 5})

How can I return more than one variable from Matlab. For example, my .m function is as below
function [lol, sos] = jk(args)
arg1 = args.arg1;
arg2 = args.arg2;
lol = arg1 + arg2;
sos = arg1 * arg2;
end

[arokem]

It's probably doable. I think that it would require changing the way that the matlab feval function is called in here:

https://github.com/arokem/python-matlab-bridge/blob/master/pymatbridge/matlab/util/pymat_feval.m
https://github.com/arokem/python-matlab-bridge/blob/master/pymatbridge/matlab/util/run_dot_m.m

I am actually not sure: the problem is to make feval aware that you want a matrix of outputs, instead of just one. This is a very strange aspect of the Matlab design.

[blink1073]

We could use something like this:

[resp{1:1}] = run_dot_m(func_path, arguments);

Then the individual return values are available as resp{1}..resp{n}. This defaults to returning the full nargout of a function.

[blink1073]

Actually, I was mistaken, we can pass nargout as an argument and this allows us to select nargout:

[resp{1:nargout}] = run_dot_m(func_path, arguments);

[blink1073]

Pretty slick actually, it even works for nargout=0. See http://blogs.mathworks.com/loren/2009/04/14/convenient-nargout-behavior/ for where I got the idea.

[arokem]

This seems good for always capturing all of the outputs. What if the user
only wants the first output?

On Fri, Feb 13, 2015 at 1:51 PM, Steven Silvester notifications@github.com
wrote:

Actually, I was mistaken, we can pass nargout as an argument and this
allows us to select nargout:

[resp{1:nargout}] = run_dot_m(func_path, arguments);

—
Reply to this email directly or view it on GitHub
#133 (comment)
.

[blink1073]

Then they pass nargout=1.

[blink1073]

It works, I tried it.

[blink1073]

The current behaviour is nargout is always 1.

[arokem]

This: #135

Doesn't seem to quite cut it.

>>> res = mlab.run_func('/Users/arokem/tmp/jk/jk.m', {'arg1': 3, 'arg2':
5})

>>> res

 {'message': 'Successfully completed request', 'result': [8], 'success':
'true'}

We somehow need to set nargin in advance?

On Fri, Feb 13, 2015 at 4:55 PM, Steven Silvester notifications@github.com
wrote:

The current behaviour is nargout is always 1.

—
Reply to this email directly or view it on GitHub
#133 (comment)
.

[blink1073]

Right, it would be a kwarg in run_func. 

—
Sent from Mailbox

On Fri, Feb 13, 2015 at 7:09 PM, Ariel Rokem notifications@github.com
wrote:

This: #135
Doesn't seem to quite cut it.

>>> res = mlab.run_func('/Users/arokem/tmp/jk/jk.m', {'arg1': 3, 'arg2':
5})
>>> res
 {'message': 'Successfully completed request', 'result': [8], 'success':
'true'}

We somehow need to set nargin in advance?
On Fri, Feb 13, 2015 at 4:55 PM, Steven Silvester notifications@github.com
wrote:

The current behaviour is nargout is always 1.

—
Reply to this email directly or view it on GitHub
#133 (comment)
.

---

Reply to this email directly or view it on GitHub:
#133 (comment)

[blink1073]

I've got a mostly working version of this, I will send a PR against your PR later today.

[blink1073]

Closed by #135.