[kepler-dev] ProvenanceListener issues

Tristan King tristan.king at jcu.edu.au
Wed Nov 14 16:41:04 PST 2007


Hi Dan,

I played around with building my own recorder in jython, based off  
your sql recorder (but specific to my portal's database), but i got a  
big too deep without understanding exactly what was going on and  
decided to back out :)

So what i do now, is create an instance of ProvenanceListener, and  
inject it into the workflow i'm about to run. then set the recorder to  
the text recording type, and set the output to a file (specific to the  
job).
then after my job is run i can view the data from the portal.

I'll look into building my own recorder again soon. and actually start  
from scratch rather than just stealing your code :). then hopefully  
more complex views can be created.

Unfortunately your current sql schema doesn't fit into the same model  
as the portal so it isn't as easy a task to embed as i hoped. but with  
the ability to build my own recorder it shouldn't be a big problem.

Cheers
-Tristan

On 15/11/2007, at 9:08 AM, Daniel Crawl wrote:

>
> Hi Tristan,
>
> Tristan King wrote:
>> So, my changes would be:
>> * add the TextFileRecording class in as a fall back recording type if
>> the config file is unavailable
>> * add public (or at least protected) modifiers to add (and remove?)
>> recording types.
>>
> Thanks for the suggestions. I've implemented and committed them
> to CVS.
>
> I am interested to know how you are using ProvenanceListener. Did
> create a new Recording subclass?
>
> --dan



More information about the Kepler-dev mailing list