m2pack-secure/scripts/validate_audio_scenarios.py

#!/usr/bin/env python3
from __future__ import annotations

import argparse
import json
import re
import sys
from dataclasses import asdict, dataclass
from pathlib import Path

from known_issues import classify_issue_ids, load_known_issue_ids


SOUND_DATA_RE = re.compile(r'^SoundData\d+\s+[0-9.]+\s+"([^"]+)"', re.IGNORECASE)


@dataclass
class AudioCheck:
    file: str
    references: list[str]
    missing_references: list[str]


def parse_args() -> argparse.Namespace:
    parser = argparse.ArgumentParser(
        description="Validate .mss audio script references against the runtime asset set."
    )
    parser.add_argument(
        "--runtime-root",
        type=Path,
        required=True,
        help="Client runtime root containing assets/.",
    )
    parser.add_argument(
        "--json",
        action="store_true",
        help="Emit JSON output.",
    )
    parser.add_argument(
        "--known-issues",
        type=str,
        default=None,
        help="Optional known issues baseline JSON. Defaults to repo known_issues/runtime_known_issues.json if present.",
    )
    parser.add_argument(
        "--strict-known-issues",
        action="store_true",
        help="Also fail when the known-issues baseline contains stale entries not observed anymore.",
    )
    return parser.parse_args()


def normalize_virtual_path(value: str) -> str:
    value = value.strip().strip('"').replace("\\", "/").lower()
    for prefix in ("d:/", "d:"):
        if value.startswith(prefix):
            value = value[len(prefix):]
            break
    return value.lstrip("/")


def build_asset_index(runtime_assets: Path) -> set[str]:
    index: set[str] = set()
    for path in runtime_assets.rglob("*"):
        if not path.is_file():
            continue
        try:
            rel = path.relative_to(runtime_assets)
        except ValueError:
            continue
        rel_posix = rel.as_posix().lower()
        index.add(rel_posix)
        if len(rel.parts) >= 2:
            index.add(Path(*rel.parts[1:]).as_posix().lower())
    return index


def parse_mss_references(path: Path) -> list[str]:
    references: list[str] = []
    for raw_line in path.read_text(encoding="utf-8", errors="ignore").splitlines():
        line = raw_line.strip()
        match = SOUND_DATA_RE.match(line)
        if match:
            references.append(match.group(1))
    return references


def validate_mss(path: Path, runtime_assets: Path, asset_index: set[str]) -> AudioCheck:
    rel = path.relative_to(runtime_assets).as_posix()
    missing: list[str] = []
    references = parse_mss_references(path)

    for reference in references:
        normalized = normalize_virtual_path(reference)
        if normalized and normalized not in asset_index:
            missing.append(reference)

    return AudioCheck(file=rel, references=references, missing_references=sorted(set(missing)))


def main() -> int:
    args = parse_args()
    runtime_root = args.runtime_root.resolve()
    runtime_assets = runtime_root / "assets"
    if not runtime_assets.is_dir():
        raise SystemExit(f"assets dir not found: {runtime_assets}")

    asset_index = build_asset_index(runtime_assets)
    checks: list[AudioCheck] = []
    failures: list[str] = []
    issue_map: dict[str, str] = {}
    referenced_audio_files: set[str] = set()

    for path in sorted(runtime_assets.rglob("*.mss")):
        check = validate_mss(path, runtime_assets, asset_index)
        checks.append(check)
        for missing in check.missing_references:
            normalized = normalize_virtual_path(missing)
            issue_id = f"audio:reference:{check.file}:{normalized}"
            message = f"{check.file}: missing reference {missing}"
            failures.append(message)
            issue_map[issue_id] = message
        for reference in check.references:
            normalized = normalize_virtual_path(reference)
            if normalized:
                referenced_audio_files.add(normalized)

    observed_issue_ids = set(issue_map.keys())
    known_path, known_issue_ids = load_known_issue_ids(__file__, "audio", args.known_issues)
    known_observed, unexpected_issue_ids, stale_known_issue_ids = classify_issue_ids(observed_issue_ids, known_issue_ids)

    result = {
        "ok": not unexpected_issue_ids and (not args.strict_known_issues or not stale_known_issue_ids),
        "checked_files": len(checks),
        "referenced_audio_files": len(referenced_audio_files),
        "failures": failures,
        "issue_ids": sorted(observed_issue_ids),
        "known_issue_ids": sorted(known_observed),
        "unexpected_issue_ids": sorted(unexpected_issue_ids),
        "stale_known_issue_ids": sorted(stale_known_issue_ids),
        "unexpected_failures": [issue_map[issue_id] for issue_id in sorted(unexpected_issue_ids)],
        "stale_known_failures": sorted(stale_known_issue_ids),
        "known_issues_path": str(known_path) if known_path else None,
        "checks": [asdict(check) for check in checks],
    }

    if args.json:
        print(json.dumps(result, indent=2))
    else:
        print(
            f"ok={result['ok']} checked_files={result['checked_files']} "
            f"referenced_audio_files={result['referenced_audio_files']}"
        )
        for failure in result["unexpected_failures"]:
            print(f"FAIL: {failure}")
        for issue_id in result["stale_known_issue_ids"]:
            print(f"STALE: {issue_id}")

    return 0 if result["ok"] else 1


if __name__ == "__main__":
    sys.exit(main())