End-to-End Wiring Audit¶

Date: 2026-03-29 — point-in-time snapshot. Method: Manual import + functional verification of every module, one by one Result (as of 2026-03-29): every module then in the package imported (one expected optional-dependency skip), all pipelines functional, 0 broken wires.

This is a dated snapshot. The package has grown substantially since; the module counts below are the figures recorded on 2026-03-29, not the current total. The live, continuously-enforced wiring guarantees are the governance gates (tools/check_module_linkage.py, the capability manifest, and the product- boundary check), which run on every commit. Re-run this manual audit to refresh the snapshot.

1. Module Tree Import (2026-03-29 snapshot)¶

Every Python module then in the package was walked via pkgutil.walk_packages and imported. One expected failure:

Module	Status	Reason
`grpc_gen.spo_pb2_grpc`	SKIP	`grpc` not installed (optional dependency)
All other 152 modules	OK	—

2. Core Public API (init.py, 15 symbols)¶

All symbols in __all__ resolve:

AuditLogger, BifurcationDiagram, BindingSpec, BoundaryObserver,
ControlAction, CouplingBuilder, PhaseExtractor, PhaseState,
RegimeManager, SPOError, StuartLandauEngine, SupervisorPolicy,
UPDEEngine, find_critical_coupling, lyapunov_spectrum, trace_sync_transition

Status: OK

3. nn/ Module (61 lazy-loaded symbols, 16 submodules)¶

All 61 symbols in nn.__all__ resolve via lazy import (__getattr__). All 16 submodule files import directly:

Submodule	Symbols	Status
`functional.py`	kuramoto_step/rk4/forward, masked variants, winfree, simplicial, stuart_landau, order_parameter, plv, saf, coupling_laplacian	OK
`kuramoto_layer.py`	KuramotoLayer	OK
`stuart_landau_layer.py`	StuartLandauLayer	OK
`simplicial_layer.py`	SimplicialKuramotoLayer	OK
`theta_neuron.py`	ThetaNeuronLayer, theta_neuron_step/rk4/forward	OK
`ude.py`	UDEKuramotoLayer, CouplingResidual	OK
`inverse.py`	analytical_inverse, hybrid_inverse, infer_coupling, coupling_correlation	OK
`oim.py`	oim_step/forward/solve, extract_coloring, coloring_violations/energy	OK
`bold.py`	bold_from_neural, bold_signal, balloon_windkessel_step	OK
`reservoir.py`	reservoir_drive, ridge_readout, reservoir_predict	OK
`chimera.py`	local_order_parameter, chimera_index, detect_chimera	OK
`spectral.py`	laplacian_spectrum, algebraic_connectivity, eigenratio, sync_threshold	OK
`training.py`	sync_loss, trajectory_loss, train_step, train, generate_kuramoto_data	OK
`__init__.py`	Lazy dispatch for all 61 symbols	OK

4. Core Pipeline (binding → coupling → engine → monitor → audit)¶

Tested with power_grid domainpack (N=12):

Step	Module	Input	Output	Status
1. Load spec	`binding.loader`	`binding_spec.yaml`	`BindingSpec(name=power_grid, 4 layers)`	OK
2. Build coupling	`coupling.knm`	`base_strength, decay_alpha`	`K shape=(12,12), symmetric, non-negative`	OK
3. Run engine	`upde.engine`	`phases, omegas, K, 200 steps`	`R=0.3896`	OK
4. Monitor	`monitor.boundaries`	`spec.boundaries`	`BoundaryObserver`	OK
5. Imprint	`imprint.update`	`ImprintModel + ImprintState`	`m_k mean=0.039`	OK
6. Audit	`runtime.audit_logger`	`path`	`AuditLogger`	OK

5. nn/ Pipeline (inverse + forward, JAX)¶

Tested on the SAME power_grid data as core pipeline:

Step	Module	Result	Status
7. Inverse	`nn.inverse.analytical_inverse`	`K_est, corr=0.529`	OK
8. Forward	`nn.functional.kuramoto_forward`	`R_jax=0.3896`	OK
9. Parity	NumPy vs JAX	`\|R_np - R_jax\| = 0.0000`	EXACT MATCH

6. Domainpack Load Snapshot¶

The 2026-03-29 wiring audit snapshot loaded 33 domainpacks without error:

agent_coordination, autonomous_vehicles, brain_connectome,
cardiac_rhythm, chemical_reactor, circadian_biology, epidemic_sir,
financial_markets, firefly_swarm, fusion_equilibrium,
gene_oscillator, identity_coherence, laser_array,
manufacturing_spc, metaphysics_demo, minimal_domain,
musical_acoustics, network_security, neuroscience_eeg,
plasma_control, power_grid, quantum_simulation, queuewaves,
robotic_cpg, rotating_machinery, satellite_constellation,
sleep_architecture, social_opinion, swarm_robotics,
traffic_flow, vortex_shedding, weather_teleconnection, wildlife_migration

33/33 snapshot OK

7. Advanced Modules¶

Category	Modules	Status
Supervisor	`regimes`, `policy`, `policy_rules`, `events`, `petri_net`, `petri_adapter`, `predictive`	OK
UPDE engines	`engine` (Euler/RK4/RK45), `stuart_landau`, `delay`, `inertial`, `jax_engine`	OK
UPDE analysis	`order_params`, `metrics`, `pac`, `bifurcation`, `reduction` (Ott-Antonsen)	OK
SSGF	`closure`, `free_energy`, `ethical`, `carrier`, `costs`, `pgbo`, `topological_integration`	OK
Drivers	`psi_physical`, `psi_informational`, `psi_symbolic`	OK
Adapters	`remanentia_bridge`, `synapse_channel_bridge`, `synapse_coupling_bridge`	OK
Reporting	`plots` (CoherencePlot)	OK
Autotune	`pipeline`, `coupling_est`, `freq_id`, `phase_extract`	OK
CLI	`cli.main` (Click Group)	OK

8. Cross-Backend Parity¶

Engine	R (power_grid, 200 steps)	Status
NumPy UPDEEngine (RK4)	0.3896	—
JAX kuramoto_forward (RK4)	0.3896	EXACT MATCH

Difference: 0.0000 (float32 sufficient for this configuration).

9. Functional Verification Summary¶

Test	Modules touched	Status
Core imports	`__init__` (15 symbols)	OK
nn/ imports	`nn/__init__` (61 symbols)	OK
nn/ direct imports	16 submodules	OK
Binding pipeline	`binding.loader`, `binding.types`	OK
Coupling builder	`coupling.knm` → symmetric, non-negative, zero-diagonal K	OK
UPDE engine	`upde.engine` (RK4, 500 steps) → R=1.0 for identical ω	OK
Stuart-Landau	`upde.stuart_landau` → r≈sqrt(μ), phases synchronise	OK
Imprint	`imprint.update`, `imprint.state` → modulate_coupling	OK
Actuation	`supervisor.ControlAction` → knob, scope, value, justification	OK
PI/S Drivers	`drivers.psi_physical/informational/symbolic`	OK
33-domainpack snapshot	All load without error	OK
CLI	`cli.main` Click Group	OK
Adapters	3 bridge modules	OK
Advanced supervisor	PetriNet, EventBus, Policy, Predictive	OK
Advanced engines	Delay, Inertial, JAX, OA reduction, PAC	OK
SSGF	7 submodules (closure, FE, ethical, carrier, costs, pgbo, topological_integration)	OK
Full module tree	152/153 (1 expected skip: grpc)	OK
End-to-end wiring	binding→coupling→engine→monitor→imprint→audit→nn/	OK
NumPy↔JAX parity	R difference = 0.0000 on power_grid	OK

Total: 0 broken wires. All modules functional and correctly interconnected.