[MOBY-dev] [moby] Re: Collections vs. Simples in registry service searches, a bug?

[Mark Wilkinson]

Hi Paul, 

I'm a bit snowed-under with four GC Tech Development applications in
preparation.  Could you check the code yourself and see if there's an
obvious reason for the behavior you are observing?

I'm not sure if searches for collections was ever implemented... ever?!?
but it was almost 5 years ago now, so my memory is a bit rusty ;-)

M




On Fri, 2006-12-15 at 14:05 -0700, Paul Gordon wrote:
> Hi all,
> 
> While I agree that makes sense for a *client* such as MOWServ to do this 
> type of splitting if it wants, my major problem is that the *registry* 
> treats both equally.  Collections should only be registered as input 
> types where a collection is actually used as a single entity, e.g. 
> multiple sequence alignments.  If you want to allow multiple 
> DNASequences to be fed into a request for a BLAST service, the service 
> should take a single DNASequence, and you can put multiple mobyData 
> blocks in the request. 
> 
> Currently, I must check through the services returned by the registry 
> and exclude those that take DNASequence Collections in order to avoid 
> errors in my Seahawk client...this seems very wrong.  Mark? Eddie?
> > Hi
> > some time ago we start developing a strategy to deal with collection. An 
> > initial idea was that MOWServ would allows call a 
> > simple-input/simple-output service with a collection (MOWServ should 
> > split the collection, invoke iteratively the service with each simple 
> > and collect all the results providing a collection as output (all 
> > simple/simple, simple/collec, collec/simple and collec/colecc cases were 
> > analysed properly).
> >
> > Under this idea, in MOWServ when you perform a query asking for services 
> > that take a simple as input, also, those services that uses a collection 
> > (of the same type, of course) are returned.
> >
> > (however we have not make a final decission and implementation, and we 
> > are re-analysing our ideas)
> >
> > regards,
> > O. (and sorry for delay in posting this comment but have a mountain of 
> > pending emails)
> >
> >
> >
> >
> > Paul Gordon escribió:
> >   
> >> Hi all,
> >>
> >> Is there any reason that a service taking a *collection* of DNASequences 
> >> as input is returned when I perform a query asking for services that 
> >> take a *simple* DNASequence as input?  e.g. the GenerateCodonPairTable 
> >> service takes a collection and is returned on a simple DNASequence 
> >> search. Here is the relevant part of the RDF from the registry:
> >>
> >>                      <mygrid:inputParameter>
> >>                          <mygrid:parameter>
> >>                               <mygrid:objectType>
> >>                                    <mobyObject:DNASequence/>
> >>                               </mygrid:objectType>
> >>                               
> >> <mygrid:hasParameterNameText>inputSequences</mygrid:hasParameterNameText>
> >>                               <mygrid:hasParameterType>
> >>                                    <mygrid:collectionParameter/>
> >>                               </mygrid:hasParameterType>
> >>                          </mygrid:parameter>
> >>                     </mygrid:inputParameter>
> >>
> >> And the request is:
> >>
> >> <findService><inputObjects>
> >> <Input>
> >> <Simple articleName="not_important">
> >> <objectType>DNASequence</objectType>
> >> <Namespace>seahawk</Namespace></Simple>
> >> </Input>
> >> </inputObjects>
> >> <outputObjects>
> >>  <Output>
> >> </Output>
> >> </outputObjects>
> >> <serviceType></serviceType>
> >> <Category>moby</Category>
> >> <authURI></authURI>
> >> <expandObjects>1</expandObjects>
> >> <expandServices>1</expandServices>
> >> <authoritative>0</authoritative>
> >> <keywords>
> >> </keywords>
> >> </findService>
> >> _______________________________________________
> >> MOBY-dev mailing list
> >> MOBY-dev at lists.open-bio.org
> >> http://lists.open-bio.org/mailman/listinfo/moby-dev
> >>
> >>   
> >>     
> > _______________________________________________
> > MOBY-dev mailing list
> > MOBY-dev at lists.open-bio.org
> > http://lists.open-bio.org/mailman/listinfo/moby-dev
> >
> >
> >   
> 
> _______________________________________________
> MOBY-dev mailing list
> MOBY-dev at lists.open-bio.org
> http://lists.open-bio.org/mailman/listinfo/moby-dev