[BBPBGLIB-1182 Main Olfactory Bulb] Resume a simulation with a changed parameter

[WeinaJi]

Migrated from https://bbpteam.epfl.ch/project/issues/browse/BBPBGLIB-1182

Request from users:

I would need support regarding the simulation I need to run with neurodamus for the Main Olfactory Bulb.

I use modOverride to change the synapses of my model. In my synapses mod files I have a parameter called training. When this parameter is equal to 1, the weight of the synapses change along the simulation. When the parameter is equal to 0, the weight are stuck.
Would it be possible to run a training simulation, then change this parameter to 0 and resume the simulation with the new weights?

Also, for validation purpose, I would need to be able o extract the synapses weight in a human readable format, is it feasible?

[WeinaJi]

I tried to record AmpaNmda.training and AmpaNmda.training_time in the synapse report.

          "AmpaNmda_training": {
               "cells": "Test_Migliore_inhib",
               "type": "synapse",
               "variable_name": "AmpaNmda.training",
               "sections": "all",
               "unit": "uS",
               "dt": 0.1,
               "start_time": 0.0,
               "end_time": 2000.0
          },
          "AmpaNmda_trainingtime": {
               "cells": "Test_Migliore_inhib",
               "type": "synapse",
               "variable_name": "AmpaNmda.training_time",
               "sections": "all",
               "unit": "uS",
               "dt": 0.1,
               "start_time": 0.0,
               "end_time": 2100.0
}

I can see the value of AmpaNmda.training_time is set correctly by the sonata config file via "synapse_configure": "%s.gmax = 1.25 %s.training_time = 100" (I used 100 in my test). But AmpaNmda.training from the report is always 1 for the whole tstop . I expect training=0 after t=100.