[MAINT][WIP] Simulate dipole enh

[jasmainak]

Just opening the PR with the intent of finishing it later

[jasmainak]

I could use a bit of help in this @rythorpe or #314

[rythorpe]

What's the goal of this PR besides to add tstop and dt?

While we're at it, wdyt of implementing the capability of a burn-in period?

[jasmainak]

Burn-in would be a bit more complicated I suppose? You'd need to make sure that downstream analysis pipelines are aware of what the burn-in is and adjust the plots accordingly

[jasmainak]

@rythorpe the goal is to get rid of code like this:

hnn-core/examples/plot_simulate_mpi_backend.py

Lines 38 to 40 in 8c07186

	params.update({
	'tstop': 310.0,
	})

as much as possible

[jasmainak]

@cjayb would it be acceptable to you if we made tstop compulsory argument to add_xxx_drives ? Currently, we were secretly passing tstop to Network instead of simulate_dipole which seems a bit semantically wrong to me. I think we might be able to make it more intelligent down the line but it's too tangled for me right now to make any sense of things unless I make it compulsory

[cjayb]

which seems a bit semantically wrong to me

Agreed! tstop is a parameter of the simulation, not the network. Poisson and bursty drives already take tstop, but the implementation there is still dependent on the idea that the end-time of a simulation is known at the time the add_XXX-method is called. I don't think this needs to be so. We can do validation/sanity checks on the drive parameters after simulate is called. I think tstop=None still makes sense as a good default for Poisson and bursty, no?

[jasmainak]

closed by #397