Add WP39 clean DOCX export policy tests

tests/test_clean_docx_export_policy.py

@@ -0,0 +1,96 @@
+from __future__ import annotations
+
+from pathlib import Path
+
+
+REPO_ROOT = Path(__file__).resolve().parents[1]
+POLICY_DOC = REPO_ROOT / "CLEAN_DOCX_EXPORT_POLICY.md"
+
+
+def _policy_text() -> str:
+    return POLICY_DOC.read_text(encoding="utf-8")
+
+
+def test_policy_forbids_clean_docx_claim_for_current_exports():
+    text = _policy_text()
+    lowered = text.lower()
+
+    assert "current docx output must not be described as a clean docx export" in lowered
+    assert "not allowed yet" in lowered
+    assert "clean docx" in lowered
+    assert "fully cleaned docx" in lowered
+    assert "metadata-free docx" in lowered
+    assert "comments removed" in lowered
+    assert "tracked changes removed" in lowered
+
+
+def test_policy_preserves_current_export_semantics_and_no_blocking():
+    text = _policy_text()
+    lowered = text.lower()
+
+    assert "current export behavior remains unchanged" in lowered
+    assert "no export is blocked by wp39" in lowered
+    assert "no content is removed by wp39" in lowered
+    assert "no export file bytes are changed by wp39" in lowered
+    assert "blocking must not be silently added" in lowered
+    assert "without a separate workpackage" in lowered
+
+
+def test_policy_defines_warning_report_rules_for_docx_hygiene_findings():
+    text = _policy_text()
+    lowered = text.lower()
+
+    assert "no supported hidden-content findings were detected" in lowered
+    assert "not a clean-docx guarantee" in lowered
+    assert "headers or footers" in lowered
+    assert "comments or margin notes" in lowered
+    assert "tracked changes can preserve" in lowered
+    assert "docx hygiene risk could not be fully assessed" in lowered
+    assert "high-risk warning + audit finding" in lowered
+
+
+def test_policy_sets_minimum_requirements_before_clean_export_claim():
+    text = _policy_text()
+    lowered = text.lower()
+
+    assert "minimum requirements before a clean-docx claim" in lowered
+    assert "explicit clean-export mode" in lowered
+    assert "comments, tracked changes, headers and footers" in lowered
+    assert "metadata" in lowered
+    assert "footnotes" in lowered
+    assert "endnotes" in lowered
+    assert "custom xml" in lowered
+    assert "residual placeholders" in lowered
+    assert "no hidden recovery" in lowered
+    assert "must not claim a docx export is clean" in lowered
+
+
+def test_policy_explicitly_does_not_implement_cleaner_or_ui_behavior():
+    text = _policy_text()
+    lowered = text.lower()
+
+    for required in [
+        "does not:",
+        "implement a docx cleaner",
+        "remove comments",
+        "accept or remove tracked changes",
+        "remove metadata",
+        "block export",
+        "change export semantics",
+        "change docx reinsert behavior",
+        "change streamlit ui",
+        "change scrub key schema",
+        "add dependencies",
+        "add real data",
+        "add cloud processing",
+    ]:
+        assert required in lowered
+
+
+def test_policy_uses_no_real_personal_data_examples():
+    text = _policy_text()
+
+    assert "Jan Jansen" not in text
+    assert "Piet de Vries" not in text
+    assert "123456782" not in text
+    assert "BSN 123" not in text