[Biojava-dev] First draft of a remote blast service class

[Sylvain Foisy]

Hi Richard,

On 11/06/09 10:17, "[NAME]" <[ADDRESS]> wrote:

> Good stuff! My 2p's worth:

Thanks. It's my first java code in the past year or so... Managing projects
really kills programming habits :-(
 
> setSequence() should be overloaded to accept all forms of possible
> sequence input - whatever is decided on as the standard way of
> referencing sequence data in BJ3. The original plan for BJ3 was to allow
> String/CharSequence and List<Symbol> (see
> http://www.biojava.org/wiki/BioJava3:HowTo )

Good point. I'll work on this. The List<Symbol> is a bit tricky: one would
need to create a timed sequence so that the program would not flood the
service. My own .02 cents: this should be done at the program level, not the
class level. The class should need to be "preoccupied" by a single request.
 
> setAdvancedOptions() should not accept a String, but rather a Properties
> or a Map<String,String>, where the keys of the Map/Properties are
> restricted to a range of acceptable values determined (and published,
> maybe as an enum?) by each of the implementation classes (e.g.
> RemoteQBlastService). The implementation class then uses this to
> construct the call string. The reason for doing it this way is that (a)
> it allows the parameters to be verified by checking them against a known
> list of allowable key/values, and (b) it allows for non-URL based remote
> requests to be constructed from the values, e.g. SOAP calls.

Mots definitely a good thing!
 
> I would also replace the static int HTML/TEXT/XML with an enum as
> numeric constants are sometimes a Bad Thing.
> 
> The setProgram() method in my mind is specific to Blast, as opposed to
> being a generic Pairwise Alignment concept. Therefore it might be better
> to move this to a Blast-specific sub-interface or make it only appear in
> the implementation classes that refer to Blast.

I am to used to using BLAST only...
 
> Finally, the JavaDocs for the various set() methods are incorrect -
> they're all mostly the same in fact! :)

I am looking and not seeing the same thing... Ok, I'll check triple ;-)
 
> But overall it looks good.

Thanks. I'll be working on this in the next few days.

Best regards

Sylvain

===================================================================

 Sylvain Foisy, Ph. D.
 Consultant Bio-informatique / Bioinformatics
 Diploide.net - TI pour la vie / IT for Life

 Courriel: sylvain.foisy at diploide.net
 Web: http://www.diploide.net
 Tel: (514) 893-4363
===================================================================