feat: Faceted Search — Dropdown-Counts passen sich aktiven Filtern an

Backend: controls-meta akzeptiert alle Filter-Parameter und berechnet
Faceted Counts (jede Dimension zaehlt mit allen ANDEREN Filtern).
Neue Facets: severity, verification_method, category, evidence_type,
release_state — zusaetzlich zu domains, sources, type_counts.

Frontend: loadMeta laedt bei jeder Filteraenderung neu, alle Dropdowns
zeigen kontextsensitive Zahlen. Proxy leitet Filter an controls-meta weiter.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

backend-compliance/tests/test_canonical_control_routes.py

@@ -443,18 +443,22 @@ class TestControlsMeta:
         db.__enter__ = MagicMock(return_value=db)
         db.__exit__ = MagicMock(return_value=False)
 
-        # 4 sequential execute() calls
-        total_r = MagicMock(); total_r.scalar.return_value = 100
-        domain_r = MagicMock(); domain_r.fetchall.return_value = []
-        source_r = MagicMock(); source_r.fetchall.return_value = []
-        nosrc_r = MagicMock(); nosrc_r.scalar.return_value = 20
-        db.execute.side_effect = [total_r, domain_r, source_r, nosrc_r]
+        # Faceted meta does many execute() calls — use a default mock
+        scalar_r = MagicMock()
+        scalar_r.scalar.return_value = 100
+        scalar_r.fetchall.return_value = []
+        db.execute.return_value = scalar_r
         mock_cls.return_value = db
 
         resp = _client.get("/api/compliance/v1/canonical/controls-meta")
         assert resp.status_code == 200
         data = resp.json()
         assert data["total"] == 100
-        assert data["no_source_count"] == 20
         assert isinstance(data["domains"], list)
         assert isinstance(data["sources"], list)
+        assert "type_counts" in data
+        assert "severity_counts" in data
+        assert "verification_method_counts" in data
+        assert "category_counts" in data
+        assert "evidence_type_counts" in data
+        assert "release_state_counts" in data