What language - platform use to data acquisiton/numerical work?

[Alexander Schmolck]

mdtorre at freemail.it (Matteo) writes:

> Thank you all for the contributions.
> 
> I have understood that Python is good for GUIs, for serial port data
> acquisition and for numercal computation.
> 
> My only problem is that MatLab has some huge numerical libraries that
> Phyton does not have. I have to do experiments with pattern recognitin
> algoritmhs (pca, neural networks etc.) and these are already
> efficently coded in MatLab. I can't afford the effort to code a radial
> basis function network from scratch (I'm able to do it, but it's
> tiring and very time consuming).
> So I'd like to use pyhon for all except numerical computation. I've
> seen Pymat to conetct MatLab to Python, but it seem a little outdated
> and unsupported. Is there any other way to embed MatLab into Python?
> 

I've hacked up a module based on pymat, so that you now can use matlab pretty
much like it were just some python library.

Examples: 

>>> mlab.plot([1,2,3])
[plots a line]

or 

>>> mlab.sort([[3.,2.,1.]])
array([       [ 1.,  2.,  3.]])

You can even do things like

>>> help(mlab.sort)
[prints the docu of sort]

and pickle objects in matlab. Speaking of objects, unlike pymat you can not
only work with arrays and strings but also with other types of matlab objects.

I have been using it for my work for some time, but I haven't yet got around
to release it (as open source). If you are interested I could send you
the code, but you'll need python2.2.1 and it might still have rough edges.

> Thank you in advance,
> 
> Matteo Della Torre

alex