|
Hi!
rete.run() or rete.executeCommand("(run)") both return only after
all activated rules have fired. If you can infer from this fact, that all
commands have been executed, this is it.
On the
other hand I'm not sure if your pseudo code correctly states what you are trying
to do. rete.run() just runs the engine once, firing 0 to N
rules (see above), and then exits. After having
defined another instance after rete.run() has exited, you need to call
rete.run() again. You might possibly want to run the engine continuously in a
separate thread using Rete::runUntilHalt(). In that case, you'd need some
synchronization mechanism (e.g., the flags you mentioned).
Note
that you can use the following Rete methods directly: Rete::clear(),
Rete::reset(), and Rete::run(). Further, you need to call Funcall::execute(...)
in order execute the (definstance) function call.
Thomas
Jess experts:
How do I know if Jess has completed a scan of all
the rules in its knowledge base?
I am trying to build a command-response
framework, where the commands are (short-lived) beans that are instantiated
prior to a Jess run(), and then are released after the results are obtained.
I'm trying to figure out if there is some way I can tell if Jess is "done"
(perhaps set a field in the bean itself).
Specifically, I want to know if the following
design/pseudocode will work:
public class RuleEngine extends
Thread
{
private Rete
rete;
public
RuleEngine()
{
rete = new Rete();
rete.addUserpackage(new
MiscFunctions());
rete.addUserpackage(new
jess.ReflectFunctions());
//
Read in the
rules
rete.executeCommand("(clear)");
rete.executeCommand("(batch sample.clp)");
}
public void run()
{
// create bean
instance
Funcall f =
new Funcall("definstance",
rete);
f.add(new
Value("Command",
RU.ATOM));
f.add(new
Value(command));
// code to
"lock" the bean
rete.executeCommand("(reset)");
rete.executeCommand("(run)");
// code to
release this instance
// code to
"unlock" bean
{
sleep for some time
}
}
}
| 