feat(governance): close the 6 data-disease gaps — DQ monitoring, ownership, access posture, lineage & observability

Adds native Command Center features (no new containers) integrated as sub-tabs
in the existing Data Explorer and Data Quality views:

- Continuous Data Quality (dq_monitor.py): live completeness/uniqueness/validity/
  freshness scorecards via Trino with rolling trends → DataQuality "Live Monitoring".
- Ownership & stewardship (catalog_governance.py): owner/steward/tier matrix,
  orphan detection, business glossary; local store best-effort synced to
  OpenMetadata (owner PATCH) → Data Explorer "Ownership".
- Access & policy posture: per-dataset compliance combining PII masking, ownership,
  live DQ and observability alerts vs data contracts → Data Explorer "Access & Policies".
- Lineage (lineage.py): staged source→CDC→Spark→S3→Iceberg→Trino→serving graph with
  live row counts and column-level PII/masking tracing → Data Explorer "Lineage".
- Observability (observability.py): volume/freshness/schema-drift monitoring with
  alerts → Data Explorer "Observability".
- Shared lake_meta.py dataset registry + bounded Trino client; fast native row-count
  and PK-indexed freshness so monitors stay cheap on 25-54M-row tables.
- LLM context (lab_context.py) enriched with DQ scores, ownership and active alerts.

api/lab_context.py

@@ -671,6 +671,21 @@ def collect_governance(log: TerminalLogFn | None = None) -> dict[str, Any]:
         "mysql_hr.hr.employee_events -> iceberg.curated_masked.employee_events_masked (PII masked)",
         "hdfs:/data/historical/sales_orders -> iceberg.hadoop.historical_sales_hdfs",
     ]
+    try:
+        from dq_monitor import summary_for_llm as _dq
+        out["data_quality"] = _dq()
+    except Exception as exc:
+        out["data_quality"] = {"error": str(exc)}
+    try:
+        from observability import summary_for_llm as _obs
+        out["observability"] = _obs()
+    except Exception as exc:
+        out["observability"] = {"error": str(exc)}
+    try:
+        from catalog_governance import summary_for_llm as _own
+        out["ownership"] = _own()
+    except Exception as exc:
+        out["ownership"] = {"error": str(exc)}
     return out
 
 
@@ -700,6 +715,27 @@ def _section_governance(g: dict[str, Any]) -> list[str]:
     lines.append("  Lineage:")
     for ln in g.get("lineage") or []:
         lines.append(f"    - {ln}")
+    dq = g.get("data_quality") or {}
+    if isinstance(dq, dict) and "error" not in dq:
+        lines.append(f"  Data quality (continuous, live tables): platform score {dq.get('platform_dq_score')}")
+        for d in dq.get("datasets") or []:
+            iss = f" issues: {', '.join(d['issues'][:3])}" if d.get("issues") else ""
+            lines.append(f"    - {d['key']}: score {d.get('score')}{iss}")
+    own = g.get("ownership") or {}
+    if isinstance(own, dict) and "error" not in own:
+        orph = own.get("orphan_datasets") or []
+        lines.append(f"  Ownership: {len(own.get('owners', {}))} assigned"
+                     + (f", orphans (no owner): {', '.join(orph)}" if orph else ", no orphans"))
+        for k, v in (own.get("owners") or {}).items():
+            if v.get("owner"):
+                lines.append(f"    - {k}: owner={v.get('owner')} steward={v.get('steward') or '—'} tier={v.get('tier') or '—'}")
+    obs = g.get("observability") or {}
+    if isinstance(obs, dict) and "error" not in obs:
+        ac = obs.get("active_alerts") or {}
+        lines.append(f"  Observability alerts: {ac.get('total', 0)} active "
+                     f"(critical={ac.get('critical', 0)}, warning={ac.get('warning', 0)})")
+        for a in (obs.get("alerts") or [])[:5]:
+            lines.append(f"    - [{a['severity']}] {a['dataset']}: {a['message']}")
     return lines