[Biojava-l] SQL-backed persistent Biojava sequence/feature objects

David Huen smh1008@cus.cam.ac.uk
Mon, 30 Jul 2001 15:34:08 +0100 (BST)


On Mon, 30 Jul 2001, Thomas Down wrote:

> On Mon, Jul 30, 2001 at 01:48:57PM +0100, Ewan Birney wrote:
> >
> 
> David, have you looked at the BPDB schema?  How welll do you
> think it meets your needs?
> 
I haven't looked at the BPDB schema yet, I'll try to do that this evening.
To save me navigating thru a framework for a package in a language I don't
understand, do you have a URL to the relevant SQL schema please?

Will Ewan and you be down at the Brocklehurst beer bash on Wed?  Perhaps
it might be easier to trash out with a pad of paper how far we can achieve
commonality without extreme complexity.

What design I've got so far is a pretty direct mapping of Biojava
sequence/feature objects onto a database and relies on using sequence and
feature ids to get a hierarchy (features have parent fields so select on
that field gets me the next level) and as you mentioned, a BLOB to chuck
the bundle into (I was tempted to use a BLOB for an entire sequence but
that's taking the piss and not very editable).  I'm trying to make it such
that I can either choose some form of lazy instantiation for Biojava
objects (and save memory at the expense of more database transactions) or
have the whole sequence object instantiated at a go.  I'm trying to keep
it simple as I'm sure I'll find design errors necessitating a rewrite and
I'd rather that happened earlier rather than later when a lot has been
invested into a design.  The primary objective for me is that other than
the original call to get such an object, all subsequent manipulations
should as if you were working on a standard Biojava object.

Thanks,
David Huen