Improve oemof simulation runs

[henhuy]

Changed following behaviour:

• result is not stored if simulation with same parameters already exists (checked after simulating results - this prevents double entries if simulations with same parameters run in parallel)
• result is not stored if optimization is infeasible

Infeasbile optimization run now throws an APIException error which is handled in digiplan here:
https://github.com/rl-institut-private/digiplan/blob/30ddeb9dc27c5e7042177ddf916c4065eb01ff5f/digiplan/static/js/results.js#L101

Note: I did not test whats happening on client side if simulation finishes with infeasible results! As SIMULATION_FINISHED event is fired with non valid simulation id this will lead to errors. Should be handled somehow - but I had not time to built an infeasible optimization problem...

[nesnoj]

Thank you @henhuy for this nice feature!
I tested and the termination works out when going back to settings. Also, the results are not stored in DB when system is infeasible 👍 .

One question: When going back to settings, the celery logs say

[2023-10-06 09:11:54,072: INFO/ForkPoolWorker-15] Starting simulation for scenario='scenario_2045'...
[2023-10-06 09:12:48,974: INFO/MainProcess] Terminating fcb417d1-3c44-412c-9d92-04829aed3a43 (Signals.SIGTERM)
[2023-10-06 09:13:04,788: INFO/MainProcess] Task django_oemof.simulation.simulate_scenario[110f179b-4c43-4b4b-96dd-7dc0de5e2cff] received

Sound good so far, but sometimes it says

[2023-10-06 09:13:11,219: INFO/ForkPoolWorker-17] Starting simulation for scenario='scenario_2045'...
[2023-10-06 09:13:21,900: INFO/MainProcess] Terminating 110f179b-4c43-4b4b-96dd-7dc0de5e2cff (Signals.SIGTERM)
[2023-10-06 09:13:21,901: ERROR/ForkPoolWorker-17] Constructing component 'TransformerBlock.relation_build' from data=None failed:
SystemExit: -241

or similar output in the 3rd line like Constructing component 'flow' from data=None failed or Constructing component 'GenericStorageBlock.balance' from data=None failed.
This does not seem to occur when the time between clicking on results tab and going back to settings is either quite short (a few seconds) or long (>30s). I guess in between the esys is build and that's why these errors appear....(?)

However, the opt. is terminated properly so I wouldn't bother too much. Or is the SystemExit a problem?

[henhuy]

I think pyomo gets interrupted by termination and thus this errors occurs.
If it does not affect further simulation runs, I wouldn't bother either...

[nesnoj]

I think pyomo gets interrupted by termination and thus this errors occurs. If it does not affect further simulation runs, I wouldn't bother either...

Ok, thx! :)