1# Author: Simon-Pierre Boucher — simon-pierre.boucher@uqo.ca2"""Evaluation layer (WP 26): metrics, paired inference, calibration, stratification,3faithfulness audit and LaTeX table writers. Definitions: docs/evaluation_protocol.md."""4from __future__ import annotations56from .metrics import per_record, summarise, summarise_by7from .inference import block_bootstrap_contrast, compare_arms, diebold_mariano, holm8from .calibration import apply_conformal, conformalise, coverage_by, pit_from_quantiles, pit_histogram, reliability, sharpness9from .stratify import add_strata, stratified_metrics10from .faithfulness import audit_run11from .tables import write_main_results, write_contrasts, write_coverage, write_strata, write_cost, write_faithfulness1213__all__ = [14 "per_record", "summarise", "summarise_by",15 "block_bootstrap_contrast", "compare_arms", "diebold_mariano", "holm",16 "apply_conformal", "conformalise", "coverage_by", "pit_from_quantiles", "pit_histogram", "reliability", "sharpness",17 "add_strata", "stratified_metrics",18 "audit_run",19 "write_main_results", "write_contrasts", "write_coverage", "write_strata", "write_cost", "write_faithfulness",20]21