feat(b17): Playwright Audit-Walk-Video (Stufe 1, #7)

Nimmt einen kompletten Site-Walk als WebKit-Browser-Session inkl. Video auf. Reviewer kann nachträglich exakt nachvollziehen, wie die Engine zum Befund kam. consent-tester: - services/audit_walk_recorder.py: Playwright record_video_dir, iPhone-Viewport-free 1280×800. Goto homepage → Banner-Accept (Best-Effort: 12 Text-Phrasen + 5 CMP-Fallback-Selektoren) → Footer-Links sammeln (compliance-relevant gefiltert) → pro Link navigate + Dwell-Time → JSON-Action-Index mit UTC-Timestamps + SHA-256 vom Video als Manipulation-Schutz. - routes_audit_walk.py: POST /scan-audit-walk; statische Serves für /audit-walks/{walk_id}/video.webm + walk.json. - main.py: Router registriert. backend: - _b17_wiring.py: Triggert /scan-audit-walk, speichert Walk-Metadata in state["audit_walk"]. Render-Block mit HTML-Tabelle aller Actions (HH:MM:SS + Aktion + Detail) + Links zu Video und walk.json. - _orchestrator.py: run_b17 nach run_b16, async-aufgerufen. - mail_render_v2/_compose.py: audit_walk_html im V2-Layout. - test_b17_audit_walk.py: 8 Tests (Render-Pfade + Wiring). Stufe-2 (Akkordeon-Expansion) und Stufe-3 (DSMS-CID-Anchor) folgen separat. Real-World-Smoke gegen Elli: - 581 KB Video, SHA-256 verifizierbar - 3 Footer-Links besucht (Impressum, Datenschutzerkl., Nutzungs-) - 6 Actions im JSON-Index Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-06-07 17:20:13 +02:00
parent 529c032641
commit cb4b352846
7 changed files with 562 additions and 0 deletions
@@ -0,0 +1,95 @@
+"""Tests for B17 Audit-Walk-Wiring (Stufe 1)."""
+
+import asyncio
+from unittest.mock import patch, MagicMock, AsyncMock
+
+import pytest
+
+from compliance.api.agent_check._b17_wiring import _render, run_b17
+
+
+_FAKE_WALK = {
+    "walk_id": "abc123def456",
+    "url": "https://example.com/",
+    "started_at": "2026-06-07T10:00:00+00:00",
+    "completed_at": "2026-06-07T10:00:30+00:00",
+    "engine": "playwright/webkit",
+    "viewport": "1280x800",
+    "actions": [
+        {"timestamp": "2026-06-07T10:00:00+00:00", "action": "goto",
+         "url": "https://example.com/", "status": 200},
+        {"timestamp": "2026-06-07T10:00:02+00:00", "action": "accept_banner",
+         "result": "clicked", "phrase": "alle akzeptieren"},
+        {"timestamp": "2026-06-07T10:00:04+00:00",
+         "action": "discover_footer_links", "count": 3, "links": []},
+        {"timestamp": "2026-06-07T10:00:06+00:00", "action": "navigate",
+         "url": "https://example.com/datenschutz",
+         "anchor_text": "Datenschutz", "status": 200,
+         "title": "Datenschutzerklärung"},
+    ],
+    "video": {
+        "filename": "video.webm",
+        "size_bytes": 512000,
+        "sha256": "a1b2c3d4e5f67890fedcba0987654321ffffeeeeddddccccbbbbaaaa00001111",
+    },
+}
+
+
+class TestRender:
+    def test_renders_walk_id_and_link(self):
+        html = _render(_FAKE_WALK)
+        assert "abc123def456" in html
+        assert "video.webm" in html
+        assert "walk.json" in html
+
+    def test_includes_sha_prefix(self):
+        html = _render(_FAKE_WALK)
+        # First 12 chars of sha
+        assert "a1b2c3d4e5f6" in html
+
+    def test_action_table_lists_all_actions(self):
+        html = _render(_FAKE_WALK)
+        # All four actions appear as <tr>
+        assert html.count("<tr>") >= 4  # incl. header
+
+    def test_nav_count_reflects_navigate_actions(self):
+        html = _render(_FAKE_WALK)
+        # 1 navigate in the fixture
+        assert "1 Compliance-Seiten" in html
+
+
+class TestRunB17:
+    def test_no_request_skipped(self):
+        state = {}
+        asyncio.run(run_b17(state))
+        assert "audit_walk" not in state
+
+    def test_no_url_skipped(self):
+        state = {"req": MagicMock(documents=[MagicMock(url="")])}
+        asyncio.run(run_b17(state))
+        assert "audit_walk" not in state
+
+    def test_consent_tester_failure_skipped(self):
+        req = MagicMock(documents=[MagicMock(url="https://example.com/dse")])
+        state = {"req": req}
+        with patch(
+            "compliance.api.agent_check._b17_wiring.httpx.AsyncClient"
+        ) as mock_client:
+            instance = mock_client.return_value.__aenter__.return_value
+            instance.post = AsyncMock(side_effect=Exception("nope"))
+            asyncio.run(run_b17(state))
+        assert "audit_walk" not in state
+
+    def test_success_populates_state(self):
+        req = MagicMock(documents=[MagicMock(url="https://example.com/dse")])
+        state = {"req": req}
+        resp = MagicMock(status_code=200)
+        resp.json = MagicMock(return_value=_FAKE_WALK)
+        with patch(
+            "compliance.api.agent_check._b17_wiring.httpx.AsyncClient"
+        ) as mock_client:
+            instance = mock_client.return_value.__aenter__.return_value
+            instance.post = AsyncMock(return_value=resp)
+            asyncio.run(run_b17(state))
+        assert state["audit_walk"]["walk_id"] == "abc123def456"
+        assert "video.webm" in state["audit_walk_html"]