8:44 a.m.
The way you are inserting the process instance is probably fine, I'm
just wondering why you need the process instance in your DRL. As there
might be alternatives where you don't need to use DRL constraint but
maybe code constraints or event nodes to reach the same behaviour.
Kris
Quoting PremKumar s <premkumar.sivanandan(a)gmail.com>:
Kris,
I am new to drools flow am not aware of other better ways to
pass
ProcessInstance to drools rule (DRL) When conditions. Based on few
examples in the forum and sample testing code with out inserting
processInstance to working memory rule constraints not able get to
process
instance in drools rule drl file.
If would be great if could suggest better approach to pass the
process
instance to drools rule (DRL) conditions. I with try that out and
update
you.
Thanks
Prem
On Fri, Oct 9, 2009 at 7:12 AM, Kris Verlaenen <
kris.verlaenen(a)cs.kuleuven.be> wrote:
> You can register a custom strategy like this:
>
> Environment env = KnowledgeBaseFactory.newEnvironment();
> env.set(EnvironmentName.ENTITY_MANAGER_FACTORY, emf);
> env.set(EnvironmentName.TRANSACTION_MANAGER,
> TransactionManagerServices.getTransactionManager());
> ObjectMarshallingStrategy[] objectMarshalingStrategies = new
> ObjectMarshallingStrategy[] {
> new ProcessInstanceMarshallingStrategy(env)
> };
> env.set(EnvironmentName.OBJECT_MARSHALLING_STRATEGIES,
> objectMarshalingStrategies);
> StatefulKnowledgeSession ksession =
> JPAKnowledgeService.newStatefulKnowledgeSession(kbase, null, env);
>
> ... and then define the strategy as ...
>
> private static class ProcessInstanceMarshallingStrategy implements
> ObjectMarshallingStrategy {
> private Environment environment;
> public ProcessInstanceMarshallingStrategy(Environment environment)
{
> this.environment = environment;
> }
> public boolean accept(Object o) {
> return o instanceof ProcessInstance;
> }
> public Object read(ObjectInputStream ois) throws IOException,
> ClassNotFoundException {
> long id = ois.readLong();
> return ((ProcessRuntime)
> environment.get("ksession")).getProcessInstance(id);
> }
> public void write(ObjectOutputStream oos, Object o) throws
IOException {
> oos.writeLong(((ProcessInstance) o).getId());
> }
> }
>
> Note that this strategy only works for active process instances (so
you
> need to make sure they are removed from the working memory before
ending
> the process instance, for example by using an on-entry/exit action.
I
> think we will probably build in this strategy by default in the
future.
>
> Could you clarify why you need the processInstance in the working
> memory? As there might be an easier alternative?
>
> Kris
>
>
> Quoting PremKumar s <premkumar.sivanandan(a)gmail.com>:
>
> > Thanks Kris for your input. It would be great if you could share
us
> > on
> > sample code snippet how to acheive this
ObjectMarshallingStrategy
> > for
> > ProcessInstances .
> >
> > Also confirm us by using this approach the process is avialable
for
> > drools
> > rule constraints to take decisions.
> >
> > Thanks
> > Prem
> >
> > On Thu, Oct 8, 2009 at 4:07 PM, Kris Verlaenen <
> > kris.verlaenen(a)cs.kuleuven.be> wrote:
> >
> > > This is indeed a special case. Process instances are already
> > stored as
> > > part of the session runtime state, so no need to store them
again
> > as
> > > part of the working memory. You need to register a custom
> > > ObjectMarshallingStrategy for ProcessInstances in the working
> > memory
> > > that simply retrieves them from the session state.
> > >
> > > Kris
> > >
> > > Quoting PremKumar s <premkumar.sivanandan(a)gmail.com>:
> > >
> > > > Hi,
> > > >
> > > >
> > > >
> > > > I am evaluating the drools expert with drool flow
using
> > > > TimeManagement
> > > > System. I am inserting the process instance into working
memory
> > of
> > > > drools
> > > > [session.insert(workflowProcessInstance)] So that the rules
can
> > use
> > > > the
> > > > process instance as part of Rule constraints to enable the
rules
> > to
> > > > make
> > > > more sophisticated decisions based on the state of the
current
> > > > process
> > > > instance.
> > > >
> > > >
> > > >
> > > > Rule snippet
> > > >
> > > >
> > > >
> > > > rule "Submit Time" ruleflow-group "Submit Time"
> > > > dialect "java"
> > > > when
> > > >
> > > > processInstance:WorkflowProcessInstance()
> > > >
> > > > tkRequest:TKRequest ()
> > > > tkRequestDAO:TKRequestDAO()
> > > > then
> > > >
> > > > // check the current processInstance
state
> > as
> > > > Active
> > > > then do that
> > > >
> > > > //Do custom operation, status change,
> > > > persistence
> > > > update
> > > >
> > > > end
> > > >
> > > >
> > > >
> > > > The approach is works fine with using
> > > > KnowledgeBuilder.newStatefulKnowledgeSession() that is no
> > > > persistence. But
> > > > once i have enabled the Persistence using
> > > > JPAKnowledgeService.newStatefulKnowledgeSession(know, null,
> > > > getEnvironment()) it throws stack over flow error due the
> > > > workflowProcessInstance is not serialized while trying to
persist
> > the
> > > > newly
> > > > added workflowProcessInstance to working memory.
> > > >
> > > >
> > > >
> > > > Once i have commented the
> > > > session.insert(workflowProcessInstance)
> > > > the stack overflow error is gone but none of my rules are
not
> > > > executed.
> > > >
> > > >
> > > >
> > > > Can some one provide guidance on how to pass this
> > > > workflowProcessInstance to
> > > > rule constriant with persistence JPAKnowledgeSession
enabled.
> > > >
> > > >
> > > >
> > > > Thanks
> > > >
> > > > Prem
> > > >
> > >
> > >
> > >
> > >
> > > Disclaimer: http://www.kuleuven.be/cwis/email_disclaimer.htm
> > >
> >
>
>
>
>
> Disclaimer: http://www.kuleuven.be/cwis/email_disclaimer.htm
>
Disclaimer: http://www.kuleuven.be/cwis/email_disclaimer.htm