SniperGold_ML/engine/run_complete.py

"""Run completion protocol (Layer 16, spec 22, RC_V1).

RUN_COMPLETE.json is produced ONLY after all chunks are committed, the
aggregator is flushed, evidence is computed, and the independent verifier
returned ACCEPTED. It contains no self-referential hashes; machine
verification re-checks every hash and file from disk.
"""

import os
import json

from . import versions
from .util import atomic_write_json


def build_run_complete(out_root, ckpt, evidence, verifier_report_path,
                       verifier_verdict, layer_hashes, outputs):
    payload = {
        "schema_version": versions.RUN_COMPLETE_SCHEMA_VERSION,
        "status": "COMPLETED",
        "run_id": ckpt["run_id"],
        "source_identity": ckpt["source_identity"],
        "versions": ckpt["versions"],
        "rows": {
            "parsed": ckpt["cumulative"]["rows_parsed"],
            "canonical": ckpt["cumulative"]["rows_canonical"],
            "bytes_read": ckpt["cumulative"]["bytes_read"],
            "malformed_total": sum(ckpt["cumulative"]["malformed"].values()),
        },
        "malformed": ckpt["cumulative"]["malformed"],
        "outputs": outputs,
        "checkpoint": {
            "last_completed_chunk": ckpt["last_completed_chunk"],
            "next_chunk": ckpt["next_chunk"],
            "status": ckpt["status"],
        },
        "hashes": {
            "evidence_manifest_hash": evidence.get("evidence_manifest_hash"),
            "layer_hashes": layer_hashes,
        },
        "verifier_report": verifier_report_path,
        "verifier_verdict": verifier_verdict,
        "completion_ts_utc": _utc_now_iso(),
    }
    atomic_write_json(_run_complete_of(out_root), payload)
    return payload


def _run_complete_of(out_root):
    return os.path.join(out_root, "RUN_COMPLETE.json")


def load_run_complete(out_root):
    with open(_run_complete_of(out_root), "r", encoding="utf-8") as fh:
        return json.load(fh)


def verify_run_complete(out_root, evidence_payload):
    """Machine re-check (spec 22): status COMPLETED and the recorded evidence
    manifest hash must equal the on-disk evidence manifest hash. Returns
    (ok, diffs)."""
    try:
        rc = load_run_complete(out_root)
    except Exception as e:
        return False, ["RUN_COMPLETE.json unreadable: %s" % e]
    diffs = []
    if rc.get("status") != "COMPLETED":
        diffs.append("status != COMPLETED")
    if rc.get("hashes", {}).get("evidence_manifest_hash") != \
            evidence_payload.get("evidence_manifest_hash"):
        diffs.append("evidence_manifest_hash mismatch")
    return (not diffs), diffs


def _utc_now_iso():
    import datetime
    return datetime.datetime.now(datetime.timezone.utc).isoformat().replace("+00:00", "Z")