[Bioperl-l] genewise runnable

[Jason Stajich]

As far as I know the WrapperBase object is a good starting point - if you
find it lacks things that you are copy+pasting into separate runnables,
then it should be made a function in the wrapper base and I leave that up
to whoever is implementing runnables.  There should only be one function -
executable() -which returns the appropriate path to the wrapper
executable, no more exists_executable() calls.  (not sure what to do when
we have wrappers that use >1 executable...).

Various existing wrapper need to be rewritten - EMBOSS needs to be moved
completely to bioperl-run instead of its hooks in Bio::Factory::EMBOSS
(perhaps keep old module as a ptr), and StandAloneBlast will need a
thorough going through.  Someone should also write a blastcl3 wrapper if
the blastcl3 parser is written.

At some point Martin will propose his analysisfactory + analysis objects &
namespaces which will hopefully be the best way to go forward that will
map to local and remote analyses in a uniform fashion.  The runnables will
be accessible in the AnalysisFactory.

As an aside/announcement: I'm not going to be able to do much
of the this though so someone in your group or Martin is going to have to
drive this so that it is a robust and easy to configure system.  There are
also lots of small bite-sized projects there for someone who wants to try
their hand at Bioperl who has been sitting on the sidelines.


-jason

On Mon, 26 Aug 2002 shawnh@worf.fugu-sg.org wrote:

> Guilty ;) we will start moving them out..
> btw how is the wrapper factory idea coming along?
>
>
> shawn
>
> On Mon, 26 Aug 2002, Jason Stajich wrote:
>
> > Is it possible to abstract your parsing code from the genewise runnables
> > (and any other -- repeatmasker?) and make them into proper Bio::Tools::XX
> > parsers s.t. we can use these modules in the general bioperl.  I sort of
> > thought we were working on the model that the runnables would only contain
> > code for setting up and running and exe and that they would delegate to
> > parsers for the processing of the output files.  This way different
> > parsers can be used and people can parse genewise output, etc w/o running
> > the apps.
> >
> > -j
> >
> >
>
>

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