pyphi.macro.search.competing_systems#

pyphi.macro.search.competing_systems(substrate, unit, micro_history, *, bounds=SearchBounds(max_constituents=4, max_update_grain=1, max_depth=1, mappings='FAMILIES', exhaustive_cap=8, apportionment='NONE', max_background=0), parallel=None, prune=None, progress=None)[source]#

f(U^J, W^J, τ_J) materialized within the unit’s footprint (Eq. 17).

τ_J is the unit’s micro grain: every unit of a returned system has a micro grain of at most τ_J.

prune="certified" skips partition sweeps whose outcome is certified by the intrinsic-information requirement; "off" evaluates everything; None (default) selects "certified" exactly when the active system measure applies the requirement. The returned systems are identical under every mode.

parallel and progress govern the evaluation of candidate systems; None defers to config.infrastructure.parallel_macro_system_evaluation.

Parameters:

substrate (Substrate)
unit (MacroUnit)
bounds (SearchBounds)
parallel (bool | None)
prune (str | None)
progress (bool | None)

Return type:

tuple[MacroSystem, …]