[Biopython-dev] Deprecating Bio.mathfns, Bio.stringfns and their C code?

[Bruce Southey]

Peter wrote:
> Bruce wrote:
>   
>> Personally I would be inclined
>> to incorporate Bio.listfns into each of these and depreciate Bio.listfns. I
>> know this is duplication but hopefully this would be addressed if someone
>> updates the code.
>>     
>
> If it were only pure python, I might agree with you.  But given there
> is C code which might well be worthwhile for speed, I'm not in any
> hurry to remove Bio.listfns until Bio.MaxEntropy.py, Bio.NaiveBayes
> and Bio.MarkovModel don't need it.  Labelling Bio.listfns as likely to
> be deprecated and removed in the future should suffice for the time
> being.
>
> Peter
>
>   
Hi,
Bug 2627 should include a modified version of Bio.MarkovModel and the 
associated test that removes Bio.listfns as well as the usage of 
numpy.oldnumeric (I hate the use of 'from x import *'). I really don't 
know enough about MarkovModel to know how correct it really is. I only 
worked with getting the test code to work including the Baum-Welsh part 
that is commented out in the test.  I'll try to get to Bio.MaxEntropy.py 
and Bio.NaiveBayes over the next week as time permits.

Also, Bio.pairwise2 appears to require the functionality of Bio.listfns. 
Since speed is relative, I think you need some type of benchmarking on 
this, which in turn needs a good example.

Bruce