[Bioperl-l] new Search objects

[Jason Stajich]

I've checked in the migration from my original Bio::Search::HSP,
Bio::Search::Subject, Bio::Search::Report objects to
Bio::Search::HSP::GenericHSP, Bio::Search::Hit::GenericHit,
Bio::Search::Result::GenericResult.

The fairly vanilla blast text parser Bio::SearchIO::blast and the NCBI XML
blast parser Bio::SearchIO::blastxml have been updated to support these as
well as the fasta parser Bio::SearchIO::fasta.  Steve's parser
Bio::SearchIO::psiblast will either get merged into the blast.pm or into a
new object called anyblast/allblast/notxmlblast =) once we agree upon a
good name.  Currently there are some issues with Steve's parser which we
are working on to come up with the best solution.  I would also like to
see us port in the HMMer parsing code which I will try and address later
this week.

I do need to go back and address the e-value/P-value/significance issues
to make sure we are properly complying with the SeqFeature::Similarity
/ SeqFeature::SimilarityPair the way we should be (in HSPI & GenericHSP).
Hilmar if you have time to take a peek and let me know where I've strayed
that would be really helpful.

One side effect from Steve's recent commits was a Bio::Root::Interface
object.  We've talked about it and I think it is better to just merge its
functionality into Bio::Root::RootI.  Interface.pm only has the methods
throw_not_implemented and warn_not_implemented suitable for creating
abstract methods on an Interface definition in our roundabout Perl OO
implementation.  These methods can replace the _abstractDeath method that
has existed to throw when someone tries to execute an inherited method
that was not implemented by the subclass.  We can start to systematically
fix these and then I think take stock of what is left for 0.9.3 and
1.0preA (or whatever we want to call it).

With the currect state of the repository I see all tests passing - I've
taken out the SearchIO::psiblast tests from SearchIO.t (lots of chatter
from it ) so Steve and I need to sift through that and decide what stays
and what goes.

Just to clarify - Ewan - is 0.9.2 officially out the door?  We can update
the links on the webpage if that is so and I'll add a freshmeat entry.

-jason

-- 
Jason Stajich
Duke University
jason@cgt.mc.duke.edu