#!/usr/bin/env python3
"""Audit and gate the frozen KP-2 full ResNet-20 validation baseline."""
import argparse
import json
import math
import os


SPLIT_HASH = "8328b206a97c420e49e54e3eca4abe3274c4756b084355784ea3fb8059e4515b"


def main():
    parser = argparse.ArgumentParser()
    parser.add_argument("--selection", default="results/kp_short_gate.json")
    parser.add_argument(
        "--bp_selection", default="results/oral_a_bp_selection.json")
    parser.add_argument("--input", default="results/kp_full/kp.json")
    parser.add_argument("--out", default="results/kp_full_gate.json")
    args = parser.parse_args()
    with open(args.selection) as handle:
        selection = json.load(handle)
    if selection.get("protocol") != "kolen_pollack_short_v1":
        raise ValueError("unexpected KP-1 selection protocol")
    if selection.get("status") != "passed":
        raise ValueError("KP-1 did not open KP-2")
    with open(args.input) as handle:
        record = json.load(handle)
    run_args = record["args"]
    expected = {
        "mode": "kp", "depth": 20, "width": 16, "seed": 0,
        "loader_seed": 0, "batch_size": 128, "epochs": 200,
        "train_limit": 0, "val_examples": 5000, "split_seed": 2027,
        "eval_split": "validation", "eval_every": 0,
        "augment_train": 1, "lr": 0.1, "output_lr": 0.1,
        "lr_schedule": "step", "lr_milestones": "100,150",
        "lr_gamma": 0.1, "warmup_epochs": 0, "momentum": 0.9,
        "weight_decay": 1e-4, "normalization": "batchnorm",
        "a_scale": 1.0, "alignment_probe": 32,
    }
    for key, value in expected.items():
        if run_args.get(key) != value:
            raise ValueError(f"KP-2 {key} drift")
    if record["provenance"]["git_tracked_dirty"]:
        raise ValueError("tracked-dirty KP-2 result")
    if record["split"]["validation_index_sha256"] != SPLIT_HASH:
        raise ValueError("KP-2 split drift")
    protocol = record["evaluation_protocol"]
    if protocol["test_evaluations"] or protocol["test_used_for_selection"]:
        raise ValueError("KP-2 touched test")
    if record.get("calibration_metric_space") != (
            "reciprocal_local_activity_products"):
        raise ValueError("KP-2 metric-space drift")

    with open(args.bp_selection) as handle:
        bp_selection = json.load(handle)
    if bp_selection.get("status") != "passed_primary":
        raise ValueError("matched full BP reference is not frozen")
    with open(bp_selection["selected"]["path"]) as handle:
        bp = json.load(handle)
    bp_accuracy = float(bp["final"]["accuracy"])
    bp_macs = int(bp["work"]["total_macs_estimate"])

    tracking = [row.get("feedback_tracking") for row in record["epochs"]]
    if len(tracking) != 200 or any(value is None for value in tracking):
        raise ValueError("KP-2 tracking trajectory is incomplete")
    trajectory_values = []
    for row, values in zip(record["epochs"], tracking):
        trajectory_values.extend([
            float(row["train_loss"]),
            float(values["mean_feedback_forward_cosine"]),
            float(values["mean_feedback_forward_relative_error"]),
            float(values["min_feedback_forward_cosine"]),
            float(values["max_feedback_forward_relative_error"]),
        ])
    diagnostics = record["diagnostics"]
    accuracy = float(record["final"]["accuracy"])
    loss = float(record["final"]["loss"])
    early = float(diagnostics["early_third_mean"])
    final_cosine = float(diagnostics["mean_feedback_forward_cosine"])
    late_cosine = sum(float(value["mean_feedback_forward_cosine"])
                      for value in tracking[150:]) / 50
    total_macs = int(record["work"]["total_macs_estimate"])
    queries = int(record["work"]["logical_batch_loss_queries"])
    finite = (bool(record["final"]["finite"])
              and all(math.isfinite(value) for value in
                      trajectory_values + [accuracy, loss, early, final_cosine]))
    checks = {
        "record_and_trajectory_finite": finite,
        "accuracy_at_least_0.88": accuracy >= 0.88,
        "early_alignment_at_least_0.80": early >= 0.80,
        "final_feedback_cosine_at_least_0.95": final_cosine >= 0.95,
        "epoch151_to200_feedback_cosine_at_least_0.95": late_cosine >= 0.95,
        "zero_task_loss_queries": queries == 0,
        "macs_at_most_1.40x_bp": total_macs <= 1.40 * bp_macs,
    }
    status = "passed" if all(checks.values()) else "failed"
    metrics = {
        "accuracy": accuracy, "loss": loss, "bp_accuracy": bp_accuracy,
        "early_third_alignment": early,
        "final_mean_feedback_forward_cosine": final_cosine,
        "epoch151_to200_mean_feedback_forward_cosine": late_cosine,
        "final_mean_feedback_forward_relative_error": float(
            diagnostics["mean_feedback_forward_relative_error"]),
        "total_macs": total_macs, "bp_total_macs": bp_macs,
        "mac_ratio_to_bp": total_macs / bp_macs,
        "logical_batch_loss_queries": queries,
        "peak_memory_allocated_bytes": int(
            record["hardware"]["peak_memory_allocated_bytes"]),
        "wall_s": float(record["timing"]["total_timed_wall_s"]),
        "source_commit": record["provenance"]["git_commit"],
    }
    output = {
        "protocol": "kolen_pollack_full_v1", "status": status,
        "checks": checks, "metrics": metrics,
        "innovation_experiment_opened": status == "passed",
        "confirmation_test_seeds_touched": False,
        "review_score_before": 5, "review_score_after": 5,
        "score_change_rule": "inherited KP baseline cannot raise score",
    }
    os.makedirs(os.path.dirname(os.path.abspath(args.out)), exist_ok=True)
    with open(args.out, "w") as handle:
        json.dump(output, handle, indent=2, sort_keys=True)
        handle.write("\n")
    print(json.dumps(output, indent=2))


if __name__ == "__main__":
    main()