[SciPy-Dev] Procedure for new code submission (pyHHT)

Stuart Mumford stuart@mumford.me...
Fri Jun 8 09:09:59 CDT 2012


Hello,

> It is fine to create your project with whatever structure you like.

Fair enough!!

> The Hilbert-Huang Transform is encumbered by a patent in the US.
Oh great!

> I recommend keeping it contained to a single package and not integrate
it into another open source library.

Seems sensible, then people can take it or leave it as required!

Thanks for your help.


Moving away from HHT, what are the documentation requirements and testing
requirements for submissions in general?

The test classes as written like
https://github.com/numpy/numpy/blob/master/doc/TESTS.rst.txt<https://github.com/numpy/numpy/blob/master/doc/TESTS.rst.txt%20>test
the execution of the code and the results? So in writing a test class
do you need to develop a test set of data and check that the result of the
test is what is expected? Then more over how do you justifiy what is
expected is correct. For a certian function (like HHT or Wavelet) do you
have to check the transform against a simple generated signal that makes it
clear the output is correct or a test 'real life' dataset?

Also for dicumentation am I right in thinking that API reference is created
from docstrings and a tutorial style documentation should be made seperatly?

Thanks
Stuart
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mail.scipy.org/pipermail/scipy-dev/attachments/20120608/6bd44e2a/attachment.html 


More information about the SciPy-Dev mailing list