import hashlib, json
from pathlib import Path
ROOT = Path(__file__).resolve().parents[1]

def test_registry_hashes_match_raw_files():
    reg = json.loads((ROOT / "data/metadata/source_registry.json").read_text())
    assert len(reg) >= 8
    for k, rec in reg.items():
        p = ROOT / rec["local_path"]
        if not p.exists():   # raw files are not committed; skip on a fresh clone
            continue
        h = hashlib.sha256(p.read_bytes()).hexdigest()
        assert h == rec["sha256"], k