"""Prospective full-law check of distinct-neighbor nonadditivity. Only primitive sources, events and numerical policy reach the shared producer. The static prediction is rebuilt for a separate comparison; it never narrows the forward enclosure. Original acquisition and a common complete source are conditional premises, not consequences of these arithmetic checks. """ from dataclasses import asdict from fractions import Fraction as Q from ..physics._sine_class_collective_interface import _bound_collective_interface from ..physics._sine_class_contrast import _contrast_decision from ..physics._sine_class_neighbor_nonadditivity import _bound_neighbor_nonadditivity from ..physics._sine_class_readout_evidence import _reconstruct_class_readout from ..physics.relational_sine_class_readout import _admit_class_readout_inputs from .artifact_io import exact_record from .sine_class_comparison_protocol import canonical_comparison_sources _EPS, _G, _H, _M = Q(1, 10**32), Q(1, 3000), Q(1, 8), Q(7, 10000) def neighbor_forward_sources(): """Select the original (1,2,1) reference and separate acquired-family cover. Absolute continuous phases enclose Theta, not its deviation coordinates. The Cartesian actual cover retains no claim that every corner is acquired; the original component balls, zero sums and preparation remain premises. """ sources = canonical_comparison_sources() return dict( classes=sources["classes"][1], pi_bounds=sources["pi_bounds"], **{ key: sources[key][1] for key in ( "reference_form_bounds", "reference_phase_bounds", "actual_form_bounds", "actual_phase_bounds", ) }, ) def neighbor_forward_inputs(): """Fix a single four-history attempt without executing any flow.""" source = neighbor_forward_sources() return dict( initial_form_bounds=source["reference_form_bounds"], initial_phase_bounds=source["reference_phase_bounds"], first_probe_amplitude=_M, second_probe_amplitude=_M, first_probe_node=4, second_probe_node=22, readout_node=13, delay=Q(0), total_duration=_H, time_step=Q(1, 128), order=16, max_steps=64, ) def neighbor_forward_policy(): """Keep full-law source transfer separate from analytic truncation errors. All four histories share exp(J*t)*z0 exactly. Only the four remaining nonlinear initialization defects survive their signed sum. In particular the zero-input defect is retained. No R5 or cubic feedback error is added to an independent complete-law forward enclosure. """ source = sum( ( _bound_collective_interface( total_input_variation=amplitude, horizon=_H, endpoint_radius=_EPS, ).nonlinear_initialization_error_upper_bound for amplitude in (Q(0), _M, _M, 2 * _M) ), Q(0), ) return dict( endpoint_radius=_EPS, gamma_upper=_G, total_duration=_H, source_error_upper_bound=source, readout_error_bound=Q(1, 10**30), actual_width_allowance=Q(1, 10**30), reading_count=4, independent_additive_reading_count=4, mixed_coefficients=(1, -1, -1, 1), prediction_comparison="closed_interval_overlap_without_intersection", separation="negative_forward_recorded_upper_below_independent_additive_recorded_lower", primary_observation="shared_prefix_cancelled_suffix_increment_mixed_interval", source_transfer="four_nonlinear_initialization_defects_after_exact_common_linear_cancellation", ) def neighbor_forward_prediction(): """Rebuild the static theorem and work/identity checks, not a response.""" return _bound_neighbor_nonadditivity( mediator_class=2, donor_amplitude=_M, receiver_amplitude=_M, horizon=_H, endpoint_radius=_EPS, readout_error_bound=Q(1, 10**30), radius=Q(1, 12), contact_work_allowance=Q(1, 10**12), donor_work_allowance=Q(1, 10**6), receiver_work_allowance=Q(1, 10**6), ) def _pair(value): if not isinstance(value, (tuple, list)) or len(value) != 2: raise ValueError("reference requires an ordered exact endpoint pair") lower, upper = map(exact_record, value) if lower > upper: raise ValueError("reference endpoints must be ordered") return lower, upper def assess_neighbor_forward_bounds(reference_bounds): """Assess an independent scalar enclosure after evidence reconstruction. This arithmetic API does not authenticate the supplied reference. A report consumer must use the reconstruction entry point below. Prediction bands are freshly derived; no stored theoretical or numerical verdict is trusted. """ reference = None if reference_bounds is None else _pair(reference_bounds) policy, prediction = neighbor_forward_policy(), neighbor_forward_prediction() direct = prediction.static_cone.direct_cubic_heat_bounds truncation = ( prediction.complete_cubic_correction_upper_bound + prediction.higher_amplitude_error_upper_bound ) nominal = direct[0] - truncation, direct[1] + truncation predicted_actual = prediction.decision.true_bounds source = policy["source_error_upper_bound"] actual = decision = reference_width = actual_width = None narrow = nominal_overlap = actual_overlap = None status = "unavailable" if reference is not None: actual = reference[0] - source, reference[1] + source reference_width = reference[1] - reference[0] actual_width = actual[1] - actual[0] narrow = actual_width <= policy["actual_width_allowance"] nominal_overlap = max(reference[0], nominal[0]) <= min(reference[1], nominal[1]) actual_overlap = max(actual[0], predicted_actual[0]) <= min( actual[1], predicted_actual[1] ) decision = _contrast_decision( actual, Q(0), policy["readout_error_bound"], reading_count=4 ) status = ( "consistency_conflict" if not nominal_overlap or not actual_overlap else ( "numerically_unresolved" if not narrow else ( "all_conditions_met" if decision.orientation == -1 and decision.null_excluded else "discrimination_unresolved" ) ) ) return dict( reference_bounds=reference, actual_bounds=actual, reference_width=reference_width, actual_width=actual_width, source_error_upper_bound=source, nominal_prediction_bounds=nominal, actual_prediction_bounds=predicted_actual, nominal_prediction_overlap=nominal_overlap, actual_prediction_overlap=actual_overlap, width_within_allowance=narrow, decision=None if decision is None else asdict(decision), additive_recorded_bounds=( -4 * policy["readout_error_bound"], 4 * policy["readout_error_bound"], ), all_conditions_met=status == "all_conditions_met", status=status, ) def assess_neighbor_forward_report(report): """Re-admit full-state evidence, including event and observation selectors. Replaying retained Taylor arithmetic does not regenerate or authenticate its derivative/Picard construction. Complete-law execution and the original source theorem remain explicit premises even when every check passes. """ inputs = neighbor_forward_inputs() selectors = { key: inputs.pop(key) for key in ("first_probe_node", "second_probe_node", "readout_node") } admitted = _admit_class_readout_inputs(**inputs) evidence = _reconstruct_class_readout(report, admitted, **selectors) reference = evidence.mixed_bounds if evidence.complete else None comparison = assess_neighbor_forward_bounds( None if reference is None else (reference.lo, reference.hi) ) return dict( evidence=evidence, comparison=comparison, all_conditions_met=evidence.complete and comparison["all_conditions_met"], status=comparison["status"], )