Database access for EMBOSS

[David Martin]

On Mon, 20 Aug 2001, Peter Rice wrote:

> Claude Bonnard wrote:
> > It is not very surprising that SRS is the  best mode  for a fast access
> > to the sequence databases from EMBOSS. As I understood, the URL mode
> > allows the access to a SINGLE sequence and would not support the
> > "USA" standard (wild card query) as SRS mode does.
>
> True. We could add an "SRSREMOTE" access mode to extend queries, easy to
> program but maybe limited practical use.
>
> > If it is the case, is there a solution when the SRS server is NOT
> > on the same machine, but on a machine which is dedicated to SRS?
> > I have in mind a rsh type of request and I would like to know if
> > someone experience this type of problem and could help me in
> > solving that.
>
> SRS access mode allows you to define the name of the getz program.
>
> How about an alternative name that is a script, and uses rsh to run a
> remote getz and returns the results?
>
> For example, if your script is called 'remotegetz' just add this to the
> database definition:
>
> app: remotegetz
>
> (you can use the full path if needed)
>
> Note: This was originally added because the Sanger Centre ran 2 versions of
> SRS (5.1 and 6.0) and I needed to switch between them, but it has other
> possible uses.

This would then allow one to add whichever script one wanted as long as it
could parse srs style arguements..

It doesn't have to be SRS, just look like it.. The potential is there for
wrapping in house rdbms with such a script.

I'll add some more comments to the admin guide if Peter can send me
details of how EMBOSS calls the wgetz program (not being much of an srs
hacker myself).



..d



----------------------------------
David Martin PhD
Bioinformatics Scientific Officer
Wellcome Trust Biocentre, Dundee
----------------------------------