#!/usr/bin/env python3
"""Check Phase 3A source boundaries, generated provenance, and secret canaries."""

from __future__ import annotations

import hashlib
import json
import pathlib
import re
import sys


ROOT = pathlib.Path(__file__).resolve().parents[1]
FORBIDDEN_WIRE_FIELDS = {
    "provider_url",
    "vm_address",
    "machine_id",
    "machine_address",
    "direct_host",
    "private_key",
    "pairing_key",
    "rtsp_url",
    "gamestream",
    "apollo_url",
    "windows_password",
}
SECRET_PATTERNS = (
    re.compile(rb"-----BEGIN [A-Z ]+PRIVATE KEY-----"),
    re.compile(rb"\bAKIA[0-9A-Z]{16}\b"),
    re.compile(rb"\bgh[pousr]_[A-Za-z0-9]{20,}\b"),
    re.compile(rb"\bsk-[A-Za-z0-9]{20,}\b"),
)
ALLOWED_SECRET_PROPERTIES = {
    ("ProviderSessionWork", "client_private_key_pem"),
}


def fail(message: str) -> None:
    raise ValueError(message)


def check_generated_provenance() -> None:
    schema_path = ROOT / "schemas/control-v1.schema.json"
    generator_path = ROOT / "tools/generate.py"
    manifest = json.loads((ROOT / "gen/manifest.json").read_text(encoding="utf-8"))
    schema_hash = hashlib.sha256(schema_path.read_bytes()).hexdigest()
    generator_hash = hashlib.sha256(generator_path.read_bytes()).hexdigest()
    if manifest.get("schema_sha256") != schema_hash:
        fail("generated manifest schema hash is stale")
    if manifest.get("generator_sha256") != generator_hash:
        fail("generated manifest generator hash is stale")
    if manifest.get("protocol_version") != (ROOT / "VERSION").read_text(encoding="utf-8").strip():
        fail("generated manifest protocol version is stale")


def check_manifest_schema() -> None:
    schema = json.loads((ROOT / "schemas/control-v1.schema.json").read_text(encoding="utf-8"))
    definitions = schema.get("$defs", {})
    for name, definition in definitions.items():
        for field in definition.get("properties", {}):
            if any(forbidden in field.lower() for forbidden in FORBIDDEN_WIRE_FIELDS):
                if (name, field) not in ALLOWED_SECRET_PROPERTIES:
                    fail(f"{name} exposes forbidden wire field {field}")


def check_proto_boundaries(path: pathlib.Path) -> None:
    message = ""
    depth = 0
    for line in path.read_text(encoding="utf-8").splitlines():
        match = re.match(r"\s*message\s+([A-Za-z0-9_]+)\s*\{", line)
        if match and depth == 0:
            message = match.group(1)
        if any(field in line.lower() for field in FORBIDDEN_WIRE_FIELDS):
            allowed = (
                message == "ProviderSessionWork"
                and re.fullmatch(r"\s*string\s+client_private_key_pem\s*=\s*[0-9]+;\s*", line)
            )
            if not allowed:
                fail(f"{path.relative_to(ROOT)} exposes forbidden wire field in {message or 'file scope'}")
        depth += line.count("{") - line.count("}")
        if depth == 0:
            message = ""


def check_text_boundaries() -> None:
    paths = [
        ROOT / "openapi/control-v1.yaml",
        ROOT / "proto/versevdi/control/v1/control.proto",
        ROOT / "frames/datagram-v1.md",
        ROOT / "frames/datagram-v2.md",
        ROOT / "frames/registry.json",
        ROOT / "registries/features.json",
        ROOT / "registries/datagrams.json",
    ]
    for path in paths:
        text = path.read_text(encoding="utf-8").lower()
        for field in FORBIDDEN_WIRE_FIELDS:
            if field in text:
                fail(f"{path.relative_to(ROOT)} contains forbidden wire field {field}")

    check_proto_boundaries(ROOT / "proto/versevdi/tunnel/v1/tunnel.proto")


def check_secret_canaries() -> None:
    for path in ROOT.rglob("*"):
        if not path.is_file() or ".git" in path.parts or path.name in {"LICENSE"}:
            continue
        data = path.read_bytes()
        for pattern in SECRET_PATTERNS:
            if pattern.search(data):
                fail(f"secret canary matched in {path.relative_to(ROOT)}")


def main() -> int:
    check_generated_provenance()
    check_manifest_schema()
    check_text_boundaries()
    check_secret_canaries()
    print("Protocol scope and provenance validation passed")
    return 0


if __name__ == "__main__":
    try:
        raise SystemExit(main())
    except (OSError, ValueError, json.JSONDecodeError) as exc:
        print(f"scope: {exc}", file=sys.stderr)
        raise SystemExit(1)