From d657c8375a5c6f083c92d0add26706eb4527b8ca Mon Sep 17 00:00:00 2001 From: Nina Seidel Date: Tue, 24 Mar 2026 10:15:00 +0000 Subject: [PATCH] Test PII anonymisation --- tests/test_clean.py | 9 +++++++++ 1 file changed, 9 insertions(+) diff --git a/tests/test_clean.py b/tests/test_clean.py index 03c7598..96b6900 100644 --- a/tests/test_clean.py +++ b/tests/test_clean.py @@ -21,3 +21,12 @@ def test_cast_types(): df = pd.DataFrame({"count": ["1", "2", "3"]}) result = cast_types(df, {"count": int}) assert result["count"].dtype == int + + +def test_strip_pii_hashes_values(): + import hashlib + df = pd.DataFrame({"email": ["alice@example.com"]}) + from src.transform.clean import strip_pii + result = strip_pii(df, ["email"]) + expected = hashlib.sha256(b"alice@example.com").hexdigest() + assert result["email"].values[0] == expected