Potential issues around intervention markers

[blanchco]

Link to project: https://app.staging.terarium.ai/projects/ac6fe22c-8f07-4acc-968f-6594671db2f6/workflow/49119674-bf50-4786-a343-7a06da6fc998?operator=383928ff-55b0-4ca8-a967-8ba9db53a7d5

There could be an issue with intervention policies, policy seems to be applied at x + 1 time and doesn't reach the stated policy number at least in the following example.

I have intervention policies applied on state D at time 10 at the value 3500 and time 40 at the value 3800:

...
        {"D_state": 7039.8759765625, "timepoint_id": 9, "sample_id": 0},
        {"D_state": 7510.7158203125, "timepoint_id": 10, "sample_id": 0},
        {"D_state": 3974.94580078125, "timepoint_id": 11, "sample_id": 0},
...
        {"D_state": 13050.0166015625, "timepoint_id": 39, "sample_id": 0},
        {"D_state": 13199.0947265625, "timepoint_id": 40, "sample_id": 0},
        {"D_state": 3940.3818359375, "timepoint_id": 41, "sample_id": 0},
...

As you can see from the above the interventions seem to be applied after the the timestep and don't seem to reach the stating values.

pyciemss worker logs:

INFO:root:{'name': '4ea5a239-c245-4314-9848-695fafdb3f6b', 'execution_payload': {'engine': 'CIEMSS', 'user_id': 'not_provided', 'model_config_id': '97fa3c4d-0624-439a-8336-1dd7b49ec829', 'timespan': {'start': 0.0, 'end': 100.0}, 'policy_intervention_id': None, 'logging_step_size': 1.0, 'extra': {'num_samples': 100, 'inferred_parameters': None, 'solver_method': 'dopri5', 'solver_step_size': 1.0}}, 'result_files': [], 'type': 'simulate', 'status': 'queued', 'engine': 'CIEMSS', 'workflow_id': '4ea5a239-c245-4314-9848-695fafdb3f6b'}
DEBUG:urllib3.connectionpool:Starting new HTTP connection (1): hmi-server:3000
DEBUG:urllib3.connectionpool:http://hmi-server:3000 "POST /simulations HTTP/1.1" 201 None
INFO:root:{'id': '4745f151-0574-4f0b-b0f6-705ac6226445', 'created_on': '2024-10-25T19:00:47.781+00:00', 'updated_on': '2024-10-25T19:00:47.781+00:00', 'name': '4ea5a239-c245-4314-9848-695fafdb3f6b', 'file_names': [], 'temporary': False, 'public_asset': False, 'execution_payload': {'engine': 'CIEMSS', 'user_id': 'not_provided', 'model_config_id': '97fa3c4d-0624-439a-8336-1dd7b49ec829', 'timespan': {'start': 0.0, 'end': 100.0}, 'policy_intervention_id': None, 'logging_step_size': 1.0, 'extra': {'num_samples': 100, 'inferred_parameters': None, 'solver_method': 'dopri5', 'solver_step_size': 1.0}}, 'result_files': [], 'type': 'SIMULATION', 'status': 'QUEUED', 'progress': 0.0, 'engine': 'CIEMSS', 'updates': []}
DEBUG:rq.queue:Pushed job 4745f151-0574-4f0b-b0f6-705ac6226445 into default
INFO:     10.244.12.49:58736 - "POST /simulate HTTP/1.1" 200 OK
INFO:root:{'name': '4c4b7008-a034-4a1d-aeda-bb97a1580154', 'execution_payload': {'engine': 'CIEMSS', 'user_id': 'not_provided', 'model_config_id': '97fa3c4d-0624-439a-8336-1dd7b49ec829', 'timespan': {'start': 0.0, 'end': 100.0}, 'policy_intervention_id': '251a6578-e270-4a0b-8ee4-1e54bd30558c', 'logging_step_size': 1.0, 'extra': {'num_samples': 100, 'inferred_parameters': None, 'solver_method': 'dopri5', 'solver_step_size': 1.0}}, 'result_files': [], 'type': 'simulate', 'status': 'queued', 'engine': 'CIEMSS', 'workflow_id': '4c4b7008-a034-4a1d-aeda-bb97a1580154'}
DEBUG:urllib3.connectionpool:Starting new HTTP connection (1): hmi-server:3000
DEBUG:urllib3.connectionpool:http://hmi-server:3000 "POST /simulations HTTP/1.1" 201 None
INFO:root:{'id': '3ae95217-a486-4972-82bb-900e64001d02', 'created_on': '2024-10-25T19:00:47.978+00:00', 'updated_on': '2024-10-25T19:00:47.978+00:00', 'name': '4c4b7008-a034-4a1d-aeda-bb97a1580154', 'file_names': [], 'temporary': False, 'public_asset': False, 'execution_payload': {'engine': 'CIEMSS', 'user_id': 'not_provided', 'model_config_id': '97fa3c4d-0624-439a-8336-1dd7b49ec829', 'timespan': {'start': 0.0, 'end': 100.0}, 'policy_intervention_id': '251a6578-e270-4a0b-8ee4-1e54bd30558c', 'logging_step_size': 1.0, 'extra': {'num_samples': 100, 'inferred_parameters': None, 'solver_method': 'dopri5', 'solver_step_size': 1.0}}, 'result_files': [], 'type': 'SIMULATION', 'status': 'QUEUED', 'progress': 0.0, 'engine': 'CIEMSS', 'updates': []}
DEBUG:rq.queue:Pushed job 3ae95217-a486-4972-82bb-900e64001d02 into default

[Tom-Szendrey]

I will create a notebook to test this
I suspect that this is an issue for the ciemss team as our converter service/models/converters.py just uses the timestep provided so im not sure why it would be off by one

[Tom-Szendrey]

ciemss/pyciemss#624