Hi Girish  (brought back to the list)

I am bit surprised by your request, because as you probably know, we are 
just now finalizing our provenance component implementation, which at 
the moment consists of
1- an event model for raw provenance events sent by the Taverna 
provenance layer
2- a (relational) data model for storing provenance events in a DB  
(mySQL or Derby are supported directly at the moment)
3- a query processing algorithm with a simple, XML-based provenance 
query language
4- a Java API that lets third parties access all the available 
provenance functionality
5- a UI component that lets workbench users access intermediate values 
produced by a workflow run, which are taken from the provenance DB (but, 
properly speaking, this is not provenance).

There is currently no proper UI to visualize provenance dependency 
graphs, and any help on that would be greatly appreciated.

All of this is being finalized, and hopefully released at the beginning 
of November, and I will then provide the associated documentation, 
specifically for (1), (3), and (4). I have no plans to officially 
document the DB schema, which is designed to be for internal use, just 
because the intention is to let third parties extract provenance through 
the API. But if a dedicated developer is keen to /contribute to the 
existing implementation/, as part of a joint effort, I have no objection 
to sharing the necessary additional info.
 For info, the next short-term step is to expose part of the DB as an 
RDF view (automatically generated).

If you could clarify what you need to do with provenance exactly, maybe 
we could be of further help. In any event, as Ian explained, if you 
really want to start from scratch then understanding the event model 
(see ProvenanceItem) is the right place to start. I suspect you would be 
re-tracing many of my steps! So it may be useful to talk

thanks, -Paolo

> Hi Ian ,
>
>  
>
> Thanks for your reply. I have some additional questions :
>
>  
>
> Is there any documentation or a technical report available which talks 
> about the various events that are sent out when the workflow gets 
> executed ? And also the event contents ?
>
>  
>
> I assume right now all these Provenance Items are getting stored in a 
> database. Is there any documentation available which talks about the 
> schema ?
>
>  
>
>  
>
> Thanks,
>
> Girish
>
>  
>
> *From:* Ian Dunlop [mailto:[email protected]]
> *Sent:* Monday, October 12, 2009 11:06 AM
> *To:* List for general discussion and hacking of the Taverna project
> *Subject:* Re: [Taverna-hackers] Provenance Plugin for Taverna
>
>  
>
> Hello,
>
> The provenance capture in T2 captures 'everything' that happens when 
> workflows are invoked.  The events that flow through your workflow are 
> captured as ProvenanceItems and sent across to the 
> ProvenanceReporter/Provenanceconnector plugin for processing.  It is 
> here where you can change what happens to these events. 
> You would need to create your own ProvenanceConnector, have a look at 
> the DerbyProvenanceConnector in the SVN for ideas and also how to 
> plugin your own. It's really just a case of subclassing 
> ProvenanceConnector and overriding the addProvenanceItem method.  You 
> will also need the magic SPI files in 
> src/main/resources/META-INF/services to make sure your version can be 
> found by taverna.  Taverna creates them through a factory so you also 
> need to create that class. 
>
> Cheers,
>
> Ian
>
> 2009/10/12 Alan Williams <[email protected] 
> <mailto:[email protected]>>
>
> Girish Subramanian wrote:
> > Hi all,
>
> Hello,
>
>
> > I am planning to write a provenance plugin for Taverna.
>
> Have you looked at the provenance capture that is already in Taverna?
> In 2.1 beta 2 the information about the run is captured in a database
> and it is this data which is used to produce the "intermediate results"
> information.
>
> For 2.1, Paolo Missier, Ian Dunlop and Stuart Owen have been working on
> improving the provenance capture.  This includes OPM support and an API
> to allow out-of-Taverna access to the provenance data.
>
>
> > I have the following
> > questions :
> >
> > 1. The SVN repository that I am using is
> > http://taverna.googlecode.com/svn/taverna/ . I guess this is the latest
> > ones. Please correct me if I am wrong.
>
> Yes, that is the code repository.
>
>
> > 2. For my provenance collection part I need some callback mechanism 
> which
> > gets invoked , before invoking or executing any of the nodes in the
> > workflows. Do we have such a framework built in Taverna engine, else 
> what is
> > the work around ? Can someone guide me through the classes that I should
> > look into to get this done .
>
> The relevant packages are under net.sf.taverna.t2.provenance.
> net.sf.taverna.t2.provenance.lineageservice.Provenance takes events and
> passes them to net.sf.taverna.t2.provenance.lineageservice.EventProcessor
>
> You would probably want to implement your own EventProcessor.  Paolo and
> Ian can give more details.
>
>
> > 3. Is there any documentation which points out how to create a new 
> plugin
> > for tavern.
>
> There will be archetypes available, certainly for activity plugins, 
> for 2.1.
>
> > Thanks,
> > Girish Subramanian
>
> Alan
>
>
> ------------------------------------------------------------------------------
> Come build with us! The BlackBerry(R) Developer Conference in SF, CA
> is the only developer event you need to attend this year. Jumpstart your
> developing skills, take BlackBerry mobile applications to market and stay
> ahead of the curve. Join us from November 9 - 12, 2009. Register now!
> http://p.sf.net/sfu/devconference
> _______________________________________________
> taverna-hackers mailing list
> [email protected] 
> <mailto:[email protected]>
> Web site: http://www.taverna.org.uk
> Mailing lists: http://www.taverna.org.uk/taverna-mailing-lists/
> Developers Guide: http://www.mygrid.org.uk/tools/developer-information
>
>
>
>
> -- 
> Ian Dunlop
> myGrid Team
> School of Computer Science
> University of Manchester
>

-- 
-----------  ~oo~  --------------
Dr. Paolo Missier
Information Management Group -  School of Computer Science, University of 
Manchester, UK
[email protected]  http://www.cs.man.ac.uk/~pmissier
-----------  ~oo~  --------------
HAPPLE (vb.) -  To annoy people by finishing their sentences for them and then 
telling them what they really meant to say.
(from The Meaning of Liff, Douglas Adams and John Lloyd)


------------------------------------------------------------------------------
Come build with us! The BlackBerry(R) Developer Conference in SF, CA
is the only developer event you need to attend this year. Jumpstart your
developing skills, take BlackBerry mobile applications to market and stay 
ahead of the curve. Join us from November 9 - 12, 2009. Register now!
http://p.sf.net/sfu/devconference
_______________________________________________
taverna-hackers mailing list
[email protected]
Web site: http://www.taverna.org.uk
Mailing lists: http://www.taverna.org.uk/taverna-mailing-lists/
Developers Guide: http://www.mygrid.org.uk/tools/developer-information

Reply via email to