feat: no-answer retry, agent status machine, escalation events/timeline (ТЗ §13-15,22,25-27,37, AC-08,16-19)

Phase 2 of the L1->L2 routing engine (Phase 1: MR!4).

- ami_loop() now also captures native AMI DialEnd/Hangup frames (not
  only UserEvent), needed to detect that an escalated agent did not
  answer. No dialplan change required - Redirect already routes the
  client channel into an existing Dial()-based transfer context, so
  Asterisk emits these events on its own; the listener just wasn't
  reading them before.
- retry_escalation_no_answer(): on NOANSWER/BUSY/CANCEL/CHANUNAVAIL/
  CONGESTION, releases the non-answering agent, excludes it, and
  reserves+redirects to the next available agent via the routing
  engine's existing exclude_agent_ids support. Exhausted pool marks
  the escalation failed and leaves the call with the AI instead of
  dropping the client (ТЗ §32).
- Agent status now actually moves through
  RESERVED -> RINGING -> TALKING -> AFTER_CALL_WORK -> AVAILABLE
  instead of staying stuck on RESERVED for the whole call; a new
  acw_sweep_loop background thread (same pattern as the existing
  failed_retry_loop) times out AFTER_CALL_WORK back to AVAILABLE.
- escalations gains attempt_count/real_agent_id/attempted_agent_ids_json
  (migration 0033); fixes a latent bug where assigned_agent_id stored
  the SIP extension instead of the real agent_id despite routing-service
  already returning it in RoutingAgentReserveOut.
- Every transition now records an interaction timeline entry and
  publishes the ТЗ §25 event catalog (AgentReserved/AgentRinging/
  AgentNoAnswer/AgentConnected/TransferCompleted/TransferFailed)
  through the existing emit_voice_event/EventOutboxRow idempotent path.

Not in this MR (see plan): SLA config, Callback, L3 (needs real
technical agents from the business), metrics.

tests/test_routing_engine.py

@@ -1,3 +1,4 @@
+from datetime import datetime, timedelta, timezone
 import json
 
 from sqlalchemy import select
@@ -134,3 +135,50 @@ def test_reserve_agent_excludes_disabled_and_excluded_ids():
         assert reserved.agent_id != excluded.agent_id
     finally:
         session.close()
+
+
+def test_set_agent_status_transitions_ringing_to_talking():
+    init_sql_schema()
+    session = get_session()
+    try:
+        agent = _make_agent(session, level="L2")
+        routing_engine.reserve_agent(session, call_id="call-status", level="L2", tenant_id=None, required_skills=[])
+
+        ringing = routing_engine.set_agent_status(session, agent_id=agent.agent_id, status="RINGING")
+        assert ringing is not None
+        assert ringing.status == "RINGING"
+
+        talking = routing_engine.set_agent_status(session, agent_id=agent.agent_id, status="TALKING")
+        assert talking.status == "TALKING"
+
+        missing = routing_engine.set_agent_status(session, agent_id="unknown-agent", status="AVAILABLE")
+        assert missing is None
+    finally:
+        session.close()
+
+
+def test_sweep_after_call_work_releases_only_expired_agents():
+    init_sql_schema()
+    session = get_session()
+    try:
+        stale = _make_agent(session, level="L2", status="AFTER_CALL_WORK")
+        fresh = _make_agent(session, level="L2", status="AFTER_CALL_WORK")
+
+        stale_ts = (datetime.now(timezone.utc) - timedelta(seconds=120)).replace(microsecond=0).isoformat()
+        stale.updated_at = stale_ts
+        stale.current_call_id = "call-stale"
+        session.commit()
+
+        swept = routing_engine.sweep_after_call_work(session, older_than_seconds=30)
+        swept_ids = {a.agent_id for a in swept}
+
+        assert stale.agent_id in swept_ids
+        assert fresh.agent_id not in swept_ids
+
+        session.refresh(stale)
+        session.refresh(fresh)
+        assert stale.status == "AVAILABLE"
+        assert stale.current_call_id is None
+        assert fresh.status == "AFTER_CALL_WORK"
+    finally:
+        session.close()