release 0.2

[jasmainak]

hi guys,

I think it's time for another release. 3 month cycles are usually recommended but 6 months is also fine. I'd like to get it out before getting busy with next grant cycle. I'm thinking that these are the goals for this release:

Priorities

• finish average dipoles PR [MRG] Fix average_dipoles to allow for one trial #368 @kenloi
• finish beta erp with Law 2019 model [MRG] Implement beta event modulated ERP model  #348 @ntolley
• LFP [MRG] Improved LFP calculation #329 @cjayb
• internal data containers: list to numpy when possible convert data containers in ExtracellularArray to Numpy #358 @cjayb and @ntolley
• calcium dynamics [MRG] ENH: calcium dynamics #333 @ntolley
• can we close section points and lengths don't match #330? we need to check that Neuron indeed adjusts L to match points. But also make sure our L and sec_pts agree! @cjayb
• Check that make all use of post-processing for plotting explicit (ignore params) #262 can be closed (no use of implicit post-processing in our examples, perhaps set postproc=False as default in simulate_dipole @cjayb
• [MAINT][WIP] Simulate dipole enh #315 ... I am actually stuck on this one. It's fiendishly difficult and low reward but must be done for the long run. Any help is really appreciated @jasmainak
• optimization PR [MRG] Optimization example script #77 (doesn't use drives API yet) @jasmainak
• LFP follow-up: create API for setting in-plane cell distance [MRG] specify in plane distance of cells #370 (finalise after [MRG] ENH: calcium dynamics #333)

Connectivity API

• Add .drop(probability, src_gids, target_gids) .plot
• Function to search for specific connections in net.connectivity list: conns = pick_connecitivity(conns, src=‘L2_pyramidal')
• Viz function that plots connections weights with color bar as plot_connectivity(net, conn)
• Merge the functions used for drives and cell-cell connections. This is partially done as everything is funneled through net.add_connection, but the next step is allowing the drives to target specific gids in a similar API [MRG+1] unify cell connectivity and drive connectivity #369
  - Make _Connectivity a proper class with public and private attributes
• Pass probability argument to drives allowing for inputs to subsets of cells [MRG] Add probability arguments to drives API #416

Bump to 0.3

• The "baseline renormalisation" issue still freaks the bejeez' out of me (hardcoded values in baseline_renormalize #220). The ad hoc fix I proposed in [WIP] subtract exponential to remove offset (a.k.a baseline_renormalize) #270 isn't the right way to go. We have to dig deeper in the mechanism files to figure out which of them have massive initialisation transients (only L5, not L2)
• dealing with make drive event times independent of gid (randomisation) #234 and Moving towards tests independent of legacy GUI output #244 @cjayb

anything else? It's going to be a bombshell of a release!

[cjayb]

Made some edits/proposals. Perhaps need to prioritise... ;)

[ntolley]

I'd be interested in taking up the calcium dynamics. I am already laying the groundwork with #348 which is nearing completion (just need to take an afternoon and play with the simulations).

I have also made a lot of progress on what is likely the most difficult task remaining for the connectivity API #367.

The remaining stuff for more verbose warnings when adding connections might be a good exercise for @kenloi as a more intermediate level PR. This would involve creating a list of what is considered "valid connections" (example, pyramidal cells shouldn't make GABA synapses), and referencing that list from net.add_connection.

[jasmainak]

@rythorpe any issues that are close to your heart and/or you'd want to tackle? we're thinking of an august release

[ntolley]

Something that's been on my mind for awhile, do you guys thinks it'd be possible to add a sentence or two to the main hnn-core page? At the moment it is extremely cryptic and requires you to go the HNN-GUI repo to figure out that this is a simulator for a neocortical column.

Perhaps

This is a leaner and cleaner version of the code based off the HNN repository, a user friendly tool for the biophysical simulation of a neocortical column. However, a Graphical User Interface is not supported at the moment in this repository.

I can open a PR if you guys agree. Perhaps later it'd also be great to have a picture of the model there as well to make it immediately clear what this repo is for.

[jasmainak]

yes, that's fine! PR is welcome :)

[jasmainak]

okay, I checked some boxes and reordered the issues so it's clearer to what's done and what's left. Expanded the connectivity checklist so it's clear to see what needs to be done there. This issue has also been pinned so it's on top of the priority list.