[kepler-users] About creating a custom actor and using it.

[Derik Barseghian]

Hi Bina,
Hopefully I've addressed some of these questions in my last response: 
http://lists.nceas.ucsb.edu/kepler/pipermail/kepler-users/2012-March/002183.html

Other responses inline:

On Mar 6, 2012, at 9:45 AM, Bina Philip wrote:

> Hello,
> 
> I wanted to know how the over-all process of creating a custom actor would work. If I were to make a custom actor in kepler then according to the documentation I would have to check out the code in the trunk for kepler and then after creating a module for my actor I would have to start coding for my actor (like mentioned in the HelloWorld example). If I have some additional external jars I need for this actor of mine I just import those external jars into my module's build path and start coding for my actor. Please correct my understanding if I am incorrect until this point.
> 

That's right. Though it's recommended that 3rd party module developers work off the last release branch (2.3 currently) rather than trunk.

> Now after this I had a question from the Users perspective, if there is another member in my team who download Kepler to use it (without any source code check out etc) and wants to use the actor I just created for Kepler. How can he/she use this actor and gain access to this? 

Users that download the released version of kepler (e.g. install from a 2.3.0 installer) will only be able to get your actor module once you publish it. They will get download it using the Module Manager from within Kepler.

> 
> I also wanted to know the contents of the kar file does this contain any modules within it that have the jars related to that kar? When somebody tries to open a kar file in kepler what goes on in the background while the workflow runs? How are the jars related to this kar identified from within the workflow system? I am trying to see if I can make a custom actor and supply my team member with a kar file for the actor, can he/she import this kar file in kepler and run this actor (without getting any source code checked out and adding any additional jars?)
> 

When you publish your module it's placed in the repository as a zip file, not a kar file.
E.g.: https://code.kepler-project.org/code/kepler/releases/released/dataturbine-2.2.0/
The zip file contains everything your module needs, e.g. the actor code, the jars it needs, etc.

In the past we discussed distributing modules as kars, but this didn't get implemented. Currently a kar typically only contains an actor, a workflow, a workflow + report design, or a workflow + report design + report instance + workflow run metadata file.

> Kindly let me know.
> 
> Best
> 
> -- 
> Regards,
> Bina
> Indiana University Bloomington
> Dept Of Computer Science (Master's).
> Contact:- 812-327-4780
> 
> _______________________________________________
> Kepler-users mailing list
> Kepler-users at kepler-project.org
> http://lists.nceas.ucsb.edu/kepler/mailman/listinfo/kepler-users