[MOBY-l] RE: [moby] debugging workflow

Mark Wilkinson mwilkinson at mrl.ubc.ca
Wed Feb 2 15:52:37 UTC 2005


Hi Matthew, 

Collections are one of the more tricky bits of MOBY messages to
understand, and unfortunately I can't quite get enough info from your
message to give you a straight answer.

Case 1)  My service takes an indefined number of inputs and treats them
all equivalently.  These would come in as a Collection.  e.g. a multiple
sequence alignment service would consume a Collection of sequences

Case 2)  My servce takes a defined, but multiple number of identical
inputs, and treats them differently.  The inputs would come in as two
(named!!) Simples.  e.g. the blast2seq program takes exactly two
sequences, so each of these would be a Simple containing a Sequence
object, and each Simple would be named something like "sequence_a" and
"sequence_b".

Case 3)  My service takes a defined, multiple number of disparate inputs
and treats them individually differently.  The inputs come into the
service as multiple, named, Simple and or Collection objects.  e.g. a
service that takes as its input a FASTA Object and a Sequence to blast
against that FASTA file.  The FASTA Object would be one Simple, and the
Sequence would be another Simple.  Alternately, the service might
consume a Collection of Sequence objects as one input, plus a Simple
Sequence object as its second input, to achieve similar results.

Hope that helps!

(hope you don't mind me forwarding this to the mailing list - it is
always better to have these conversations in the public eye to help
others who might be having similar problems.

M



On Wed, 2005-02-02 at 06:43, matthew couchman (JIC) wrote:
> Hi Mark,
> 
> thanks for this tip. I'm afraid i've got another question now! How do
> you go about setting up a service that takes more than one object as a
> parameter? Should they be part of a Collection?
> 
> Thanks,
> 
> 		Matt.
> 
> 
> -----Original Message-----
> From: Mark Wilkinson [mailto:mwilkinson at mrl.ubc.ca] 
> Sent: 21 January 2005 21:41
> To: matthew couchman (JIC)
> Subject: RE: [moby] debugging workflow
> 
> 
> Hi Matthew,
> 
> In the Perl/scripts folder of the CVS there is a script called
> "debugYourService" that does what you want :-)
> 
> M
> 
> 
> On Fri, 2005-01-21 at 05:44, matthew couchman (JIC) wrote:
> > Hi Mark,
> > 
> > i've been trying to knock up just a small script that will take a file
> > containing Moby inputs and send it to a service. I thought it would be
> > useful for testing things like this workflow problem using the
> > intermediate input results from the Taverna output. Unfortunately I
> > haven't got very far - I can't seem to get the input parsed properly.
> > Have you got any scripts that do this sort of thing? Sorry to bother
> > you, I realise you must be pretty busy at the moment :)
> > 
> > Cheers,
> > 
> > 		Matt.
> > 
> > -----Original Message-----
> > From: Mark Wilkinson [mailto:mwilkinson at mrl.ubc.ca] 
> > Sent: 20 January 2005 18:05
> > To: Beatrice Schildknecht
> > Cc: Rebecca Ernst; matthew couchman (JIC)
> > Subject: Re: [moby] debugging workflow
> > 
> > 
> > On Thu, 2005-01-20 at 09:05, Beatrice Schildknecht wrote:
> > > Aaargh...Just logging in to a flood of emails.
> > 
> > :-)  Merry Christmas!
> > 
> > M
-- 
Mark Wilkinson
Assistant Professor (Bioinformatics)
Dept. Medical Genetics, UBC, Canada




More information about the moby-l mailing list