So, if I have to go the API route, I need some guidance to understand:
1) How can I embed process builder inside my app (jars that can be used, parameters that need to be configured etc). As I mentioned earlier, using an iFrame and deploying work bench parallely is not an option for me.
since this is web application most likely it won't work in standalone mode. It must be deployed to app server
2) How can I run the path finder algorithms on my existing process BPMNs to get the list of paths and related path element counts? Any direction to get the path finder utility and its documentation on usage will be really helpful.
path finder is part of jbpm-simulation and can be used in standalone mode. See these test cases to see examples on how to use it.
Thanks for the pointer. The new simulation API did help me in finding the paths for the processes. However, if there is a loop in the process, I see the elements only once for the looped section. I was planning on using the paths for generating project schedules considering the time needed on each node. But with looping structure, I do not get the repeated nodes in the path, so I cannot add times for them again as they need to be processed for the second time. Any suggetions on this?
On the workbench designer side, I believe that the solution will be to share the repository between Workbench and my application to access the knowledge generated by the workbench. Publishing to my app knowledge base will have to be manually instantiated from my app in such a case.
Thanks for you help.
path finder does not really care how many iterations will go over the loop as it just identified there is a loop as one of potential paths and that's it. for going over the loop itself you need to either simulate the process by specifying what are the expected wait times in the nodes to calculate the complete time needed to complete the instance.