[BBPBGLIB-762] The ability to update single cell features mid simulation

[WeinaJi]

This is a duplicate of the JIRA ticket BBPBGLIB-762.

In the case of thalamus simulations, it is very interesting to see the transition from a wake-like to a sleep-like state, which in part amounts to increasing potassium leak conductance (e.g. from 0.5gk_leak to 1.5gk_leak) during a simulation.

Current solution working for both NEURON and CORENEURON:
We will implement a new parameter "delay" in the "modifications" section under "conditions". An example to update the "gk_leak" of a certain mechanism(TC_??) will be like:

"conditions": {
    "v_init": -65,
    "randomize_gaba_rise_time": true,
    "modifications": [
       {
            "name": "gk_leak_t15",
             "node_set": "small",
            "type": "ConfigureAllSections",
            "section_configure": "%s.gk_leak_TC_?? *= 0.5"
        },
        {
            "name": "gk_max_t20",
            "node_set": "small",
            "type": "ConfigureAllSections",
            "section_configure": "%s.gk_leak_TC_?? *= 2",
            "delay" : 20
        }
     ]
}

[WeinaJi]

Required changes:

1. neurodamus branch: https://github.com/BlueBrain/neurodamus/tree/weji/polina_copy
2. libsonata branch for adding "delay" key in the simulation config https://github.com/BlueBrain/libsonata/tree/weji/simconf_modifications
3. neuron branch for update coreneuron report in the middle of simulation https://github.com/neuronsimulator/nrn/tree/weji/update_reports_v2