lib/validation.py

from __future__ import annotations

import json
from pathlib import Path
from typing import Any

from jsonschema import Draft202012Validator, FormatChecker
from referencing import Registry, Resource

from .claim_ladder import LEVEL_ORDER
from .consent_gate import comparison_gate, public_exhibit_gate
from .temporal_ratchet import validate_temporal_ratchet

SCHEMA_FILES = {
    "creator_trace_capsule": "creator_trace_capsule_schema_v0_1_0.json",
    "temporal_event": "temporal_event_schema_v0_1_0.json",
    "evidence_asset_manifest": "evidence_asset_manifest_schema_v0_1_0.json",
    "consent_receipt": "consent_receipt_schema_v0_1_0.json",
    "causal_arc_packet": "causal_arc_packet_schema_v0_1_0.json",
    "cross_account_comparison_node": "cross_account_comparison_node_schema_v0_1_0.json",
    "metric_receipt": "metric_receipt_schema_v0_2_0.json",
    "holographic_comparison": "holographic_comparison_schema_v0_2_0.json",
}


def _load_schema(schema_dir: Path, filename: str) -> dict[str, Any]:
    return json.loads((schema_dir / filename).read_text(encoding="utf-8"))


def build_registry(schema_dir: str | Path) -> tuple[Registry, dict[str, dict[str, Any]]]:
    root = Path(schema_dir)
    schemas = {key: _load_schema(root, filename) for key, filename in SCHEMA_FILES.items()}
    registry = Registry()

    # Register each schema under its literal filename, declared $id, and the
    # absolute URI produced when the creator capsule resolves local references.
    for key, filename in SCHEMA_FILES.items():
        schema = schemas[key]
        resource = Resource.from_contents(schema)
        uris = {
            filename,
            schema.get("$id", filename),
            f"https://primordial.example/schema/{filename}",
        }
        for uri in uris:
            registry = registry.with_resource(uri, resource)
    return registry, schemas


def schema_errors(instance: Any, schema_name: str, schema_dir: str | Path) -> list[str]:
    registry, schemas = build_registry(schema_dir)
    validator = Draft202012Validator(schemas[schema_name], registry=registry, format_checker=FormatChecker())
    errors = sorted(validator.iter_errors(instance), key=lambda e: list(e.absolute_path))
    rendered: list[str] = []
    for error in errors:
        path = "/".join(str(p) for p in error.absolute_path) or "<root>"
        rendered.append(f"{path}: {error.message}")
    return rendered


def semantic_capsule_errors(capsule: dict[str, Any]) -> list[str]:
    errors: list[str] = []
    events = capsule.get("temporal_events", [])
    errors.extend(validate_temporal_ratchet(events))

    content_topology = capsule.get("content_topology", {})
    if content_topology.get("system_inferred_labels") not in ([], None):
        errors.append("system_inferred_labels must remain empty.")

    asset_ids = {a.get("asset_id") for a in capsule.get("evidence_assets", [])}
    event_ids = {e.get("event_id") for e in events}
    for asset in capsule.get("evidence_assets", []):
        if asset.get("execution_blocked") is not True:
            errors.append(f"{asset.get('asset_id')}: execution_blocked must be true.")
        for linked in asset.get("linked_event_ids", []) or []:
            if linked not in event_ids:
                errors.append(f"{asset.get('asset_id')}: linked event {linked!r} does not exist.")

    for claim in capsule.get("claims", []):
        rank = LEVEL_ORDER.get(claim.get("claim_level", ""), 99)
        if rank >= 5 and claim.get("state") != "BLOCKED":
            errors.append(f"{claim.get('claim_id')}: L5/L6 claims must be BLOCKED in this prototype.")
        if rank >= 2 and claim.get("human_review_required") is not True:
            errors.append(f"{claim.get('claim_id')}: L2-L6 claims require human_review_required=true.")
        for basis in claim.get("basis", []) or []:
            if basis.startswith("ASSET_") and basis not in asset_ids:
                errors.append(f"{claim.get('claim_id')}: claim basis {basis!r} is not a known asset.")

    review_pack = capsule.get("review_pack", {})
    blockers = review_pack.get("closure_blockers", []) or []
    if capsule.get("loop_state") == "CLOSED_FOR_CURRENT_SCOPE" and blockers:
        errors.append("False closure: CLOSED_FOR_CURRENT_SCOPE cannot coexist with closure blockers.")

    consent = capsule.get("consent", {})
    if consent.get("publication_requires_additional_review") is not True:
        errors.append("publication_requires_additional_review must be true.")

    return errors


def validate_capsule(capsule: dict[str, Any], schema_dir: str | Path) -> list[str]:
    return schema_errors(capsule, "creator_trace_capsule", schema_dir) + semantic_capsule_errors(capsule)


def validate_causal_arc(packet: dict[str, Any], schema_dir: str | Path) -> list[str]:
    errors = schema_errors(packet, "causal_arc_packet", schema_dir)
    if packet.get("human_review_required") is not True:
        errors.append("human_review_required must be true.")
    return errors


def gate_receipt(consent_state: str, scope: str) -> dict[str, str | bool]:
    comparison_ok, comparison_reason = comparison_gate(consent_state, scope, human_review_approved=False)
    public_ok, public_reason = public_exhibit_gate(consent_state, scope, publication_review_approved=False)
    return {
        "comparison_allowed_now": comparison_ok,
        "comparison_state": comparison_reason,
        "public_exhibit_allowed_now": public_ok,
        "public_exhibit_state": public_reason,
    }