Add a get_run_state function at application level

[optislangdeveloper]

📝 Description of the feature

If you use

osl.application.project.start(wait_for_finished=False)

you want to have a way of checking if the run finished.

My current approach of polling
while True: status = osl.project.root_system.get_status() if status == "Processing done": print(f'status: {status}') break elif status == "Check failed": print(f'status: {status}') break

does work, but it comes to timing issues when I activate/deactivate certain nodes and call again

osl.application.project.start(wait_for_finished=False)

The issue I noticed is, that the second start call does nothing - my assumption, the first start is not yet at end even though the status is already Processing done.

💡 Steps for implementing the feature

I want to call a function that returns the run state of the application / system based on the same notification system that is used when I call

osl.application.project.start(wait_for_finished=True)

🔗 Useful links and references

proxy_solver_amop_opt_on_amop_validation.txt

[PanekOndrej]

There is implemented check, whether project is not already running at the beginning of the start method and if so, we don't send START command at all. That's probably why the second start does nothing.

if self.__get_project_status() == "PROCESSING":
            already_running = True
            self._logger.debug("Status PROCESSING")

if not already_running:
            self.send_command(
                command=commands.start(self.__password),
                timeout=self.timeouts_register.get_value(current_func_name),
                max_request_attempts=self.max_request_attempts_register.get_value(
                    current_func_name
                ),
            )

[optislangdeveloper]

@PanekOndrej

Does it mean, that instead of
status = osl.project.root_system.get_status()
I can just use
status = osl.project.get_status()
?

[PanekOndrej]

@PanekOndrej

Does it mean, that instead of status = osl.project.root_system.get_status() I can just use status = osl.project.get_status() ?

There is a slight difference in these 2:

while osl.project.root_system.get_status behaves as a Node and therefore returns status from get_actor_info query,
osl.project.get_status returns status from get_basic_project_info query.

But I believe it should provide the same outcome.

[optislangdeveloper]

Okay, let me try to explain it.

When I query root_system.get_status() and I get "Processing done" an immediate call to start() fails w/ no error message every now and then. I believe, b/c

• the condition you described in the comment above
• the system is not done in optiSLang and the call is ignored there

I would suggest, that you always register the listener for EXECUTION_FINISHED as in the start(wait_for_finished=True) case and that the EXECUTION_FINISHED state can be queries for the start(wait_for_finished=False) case.

@rfahlberg what do you think?

[optislangdeveloper]

The issue is not only with subsequent start(wait_for_finished=False) calls but also appears WITH throwing an error for example for save()

[PanekOndrej]

I investigated this a little while creating rdo example. Problem is that update of project status (get_basic_project_info query) is delayed after root_system status update (get_actor_info query).

When start command is called multiple times with not enough time in between the starts to let optiSLang update it's states, following start command were sometimes falling into already_running condition of start command and therefore, start command was not sent to the server.

• Preferably use project status when polling as this one used in start command.

• I think we should raise level of log from debug to warning in this case.

I also observed weird behavior when running RDO example:

• It sequentially activates individual nodes and runs them separately one by one
• First blocked is finished (based on the polling mechanism)
• Status changed back to 'PROCESSING'

2025-01-28 15:14:59,446 - INFO - optiSLang_2469694455632 - proxy_solver_amop_opt_on_amop_validation.wait_before_starting_project - Project is still `PROCESSING`, wait until status changes.
2025-01-28 15:14:59,549 - INFO - optiSLang_2469694455632 - proxy_solver_amop_opt_on_amop_validation.wait_before_starting_project - Wait completed, current project status is: `Processing done`
2025-01-28 15:14:59,553 - WARNING - optiSLang_2469694455632 - osl_server.start - Project is already PROCESSING, `start` command was not sent.

• Question: Does the status change in between execution of individual nodes (or even designs) or always remains in 'Runing' (root system) / 'PROCESSING' (project) state until all active nodes are finished?