Clarification and updating docstrings for the behaviour of `add_projections` argument in instantiate_gids()

[darshanmandge]

• add_projections: bool | list[str] = False : add_projection argument of instantiate_gids() function in bluecellulab can take 4 types of inputs:

  • False (default) or None: Enable all external projections and internal projections (connections). The projections can be from any of these node types: biophysical, virtual, single_compartment, point_neuron, astrocyte and vasculature.
  • True: will keep only "virtual" node type connections
  • str: single string of the node population whose projections will be enabled
  • list of str: list of names of node populations whose projections will be enabled

• References: See here and here.

• We need to update the docstring for the argument as well.

[ilkilic]

I have opened a draft PR that clarifies the semantics of add_projections for SONATA circuits and removes ambiguity between intrinsic connectivity and projections.

The proposal makes projection handling edge-population based, updates the docstrings and tests accordingly and slightly changes the behavior of add_projections=True to mean intrinsic + all projections (instead of projection-only).

This is intended as a first step toward clearer, more explicit semantics and not as a full Neurodamus alignment yet