[Bioperl-l] SeqFetcherI (fwd)

[Val Curwen]

Ewan Birney wrote:
> 
Ewan wrote
> 
> Why is Bio::DB::RandomAccessI not ok with
> 
>   $db->get_Seq_by_acc();
> 
> and
> 
>   $db->get_Seq_by_id();
> 
> (you can implement only one if you wish).
> 

(Apologies to anyone who hasn't messed with Ensembl pipeline code, this
will probably be gobbledygook)

This came about because we're trying to make the sequence fetching in
the ensembl pipeline easier for people to use who are setting up their
own installations outside - at the moment it's very nasty and hidden
down inside Runnables making it very hard to use. We have our
EnsEMBL::Pipeline::Seqfetcher that I have split down into little modules
to run pfetch, efetch, getz, whatever. We thought it would make sense to
make it compliant with a Bioperl SeqFetcher if such a thing were to be
written ... we don't currently use bioperl-db for our seq storing here,
but it is conceivable that people outside will (and I think Elia is
talking to some people who are planning just that) and we didn't want
them to have to rewrite Runnable/Runnable DB code if we could just have
those modules do a 

my $seq = $seqfetcher->fetch_Seq_by_db_accession

Does that sound reasonable? I am not wedded to any particular scheme, I
just want the pipeline to run cleanly and easily for everyone :-) 

Val