digiops/call-center
2
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

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

Browse Source 
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.
This commit is contained in:
arys
2026-08-30 14:19:23 +05:00
parent 010a8dcab6
commit 1dcfaf46cf
19 changed files with 792 additions and 36 deletions
Download Patch File Download Diff File Expand all files Collapse all files
services/asterisk_bridge_service/bridge_processing.py
+153 -7
View File
@@ -4,6 +4,7 @@ from datetime import datetime, timezone
import hashlib
import json
import logging
import os
from pathlib import Path
import threading
from typing import Any
@@ -884,18 +885,34 @@ def process_call_ended(
open_escalation = session.execute(
select(EscalationRow).where(
EscalationRow.call_id == row.call_id,
EscalationRow.status.in_(["requested", "ringing"]),
EscalationRow.status.in_(["requested", "ringing", "connected"]),
)
).scalar_one_or_none()
was_talking = open_escalation is not None and open_escalation.status == "connected"
if open_escalation is not None:
open_escalation.status = "completed" if answered else "failed"
open_escalation.completed_at = now
if answered and not open_escalation.connected_at:
open_escalation.connected_at = now
try:
bridge._release_routing_agent(row.call_id)
except Exception:
pass
if was_talking and open_escalation and open_escalation.real_agent_id:
try:
bridge._set_routing_agent_status(open_escalation.real_agent_id, "AFTER_CALL_WORK")
except Exception:
logger.warning("bridge.agent_acw_status_failed call_id=%s agent_id=%s", row.call_id, open_escalation.real_agent_id)
else:
try:
bridge._release_routing_agent(row.call_id)
except Exception:
pass
if open_escalation is not None:
try:
bridge._append_interaction_timeline(
interaction_id=link.interaction_id,
action="escalation.completed" if answered else "escalation.failed_client_disconnected",
metadata={"call_id": row.call_id, "escalation_id": open_escalation.escalation_id, "agent_id": open_escalation.real_agent_id},
)
except Exception:
pass
try:
bridge._notify_voice_ai_telephony_event(
voice_session_id=link.voice_session_id,
@@ -969,6 +986,48 @@ def process_operator_connected(
link.telephony_status = "connected"
link.connected_at = now
link.updated_at = now
open_escalation = session.execute(
select(EscalationRow).where(
EscalationRow.call_id == row.call_id,
EscalationRow.status == "ringing",
)
).scalar_one_or_none()
if open_escalation is not None:
open_escalation.status = "connected"
open_escalation.connected_at = now
if open_escalation.real_agent_id:
try:
bridge._set_routing_agent_status(open_escalation.real_agent_id, "TALKING")
except Exception:
logger.warning("bridge.agent_talking_status_failed call_id=%s agent_id=%s", row.call_id, open_escalation.real_agent_id)
try:
bridge._append_interaction_timeline(
interaction_id=link.interaction_id,
action="escalation.agent_connected",
metadata={"call_id": row.call_id, "agent_id": open_escalation.real_agent_id, "escalation_id": open_escalation.escalation_id},
)
except Exception:
pass
try:
bridge._emit_voice_event(
event_type="AgentConnected",
call_id=row.call_id,
interaction_id=link.interaction_id,
payload={"escalation_id": open_escalation.escalation_id, "agent_id": open_escalation.real_agent_id},
)
except Exception:
pass
try:
bridge._emit_voice_event(
event_type="TransferCompleted",
call_id=row.call_id,
interaction_id=link.interaction_id,
payload={"escalation_id": open_escalation.escalation_id, "agent_id": open_escalation.real_agent_id},
)
except Exception:
pass
_upsert_voice_reporting_fact(
session,
link,
@@ -1173,6 +1232,55 @@ def process_recording_ready(
local_path.unlink(missing_ok=True)
_NO_ANSWER_DIAL_STATUSES = {"NOANSWER", "BUSY", "CANCEL", "CHANUNAVAIL", "CONGESTION"}
_NO_ANSWER_HANGUP_CAUSES = {"17", "18", "19", "21", "34", "38"}
def process_agent_dial_outcome(session, row: AsteriskEventLogRow, payload: dict[str, Any]) -> None:
bridge = _bridge_app()
call_id = row.call_id
if not call_id:
bridge._mark_log_forwarded(session, row, interaction_id=None)
return
dial_status = str(payload.get("DialStatus") or "").strip().upper()
hangup_cause = str(payload.get("Cause") or "").strip()
is_no_answer_outcome = (
(row.ami_event_name == "DialEnd" and dial_status in _NO_ANSWER_DIAL_STATUSES)
or (row.ami_event_name == "Hangup" and hangup_cause in _NO_ANSWER_HANGUP_CAUSES)
)
if not is_no_answer_outcome:
bridge._mark_log_forwarded(session, row, interaction_id=None)
return
escalation = session.execute(
select(EscalationRow).where(
EscalationRow.call_id == call_id,
EscalationRow.status == "ringing",
)
).scalar_one_or_none()
if escalation is None:
bridge._mark_log_forwarded(session, row, interaction_id=None)
return
dialed_channel = payload.get("DestChannel") if row.ami_event_name == "DialEnd" else payload.get("Channel")
dialed_extension = bridge._extract_extension_from_channel(dialed_channel)
if dialed_extension and escalation.assigned_agent_id and dialed_extension != escalation.assigned_agent_id:
bridge._mark_log_forwarded(session, row, interaction_id=None)
return
try:
bridge._retry_escalation_no_answer(
session,
call_id=call_id,
dial_outcome=dial_status or f"hangup_cause_{hangup_cause}",
)
except Exception:
logger.exception("bridge.agent_dial_outcome_retry_failed call_id=%s", call_id)
bridge._mark_log_forwarded(session, row, interaction_id=None)
def process_bridge_row(session, row: AsteriskEventLogRow) -> AsteriskEventLogRow:
bridge = _bridge_app()
payload = json.loads(row.payload_json or "{}")
@@ -1209,6 +1317,8 @@ def process_bridge_row(session, row: AsteriskEventLogRow) -> AsteriskEventLogRow
bridge._process_call_ended(session, row, payload)
elif row.ami_event_name == f"{bridge._ami_prefix()}RecordingReady":
bridge._process_recording_ready(session, row, payload)
elif row.ami_event_name in {"DialEnd", "Hangup"}:
bridge._process_agent_dial_outcome(session, row, payload)
else:
row.forward_status = "received"
row.updated_at = utc_now_iso()
@@ -1216,9 +1326,9 @@ def process_bridge_row(session, row: AsteriskEventLogRow) -> AsteriskEventLogRow
return row
def record_ami_payload(payload: dict[str, Any]) -> None:
def record_ami_payload(payload: dict[str, Any], *, event_name: str | None = None) -> None:
bridge = _bridge_app()
event_name = str(payload.get("UserEvent") or "").strip()
event_name = str(event_name or payload.get("UserEvent") or "").strip()
call_id = bridge._extract_call_id(payload)
linked_id = bridge._extract_linked_id(payload, call_id)
if not event_name or not call_id:
@@ -1262,6 +1372,42 @@ def retry_failed_events_once() -> None:
bridge._process_claimed_bridge_event(bridge_event_id)
def acw_duration_seconds() -> int:
raw = str(os.environ.get("ACW_DURATION_SECONDS", "30")).strip()
try:
return max(int(raw), 1)
except ValueError:
return 30
def acw_sweep_interval_seconds() -> float:
return max(min(float(acw_duration_seconds()) / 2, 15.0), 5.0)
def acw_sweep_once() -> None:
from services.routing_service import engine as routing_engine
session = get_session()
try:
swept = routing_engine.sweep_after_call_work(session, older_than_seconds=acw_duration_seconds())
for agent in swept:
logger.warning("bridge.agent_acw_swept agent_id=%s", agent.agent_id)
finally:
session.close()
def acw_sweep_loop(stop_event: threading.Event | None = None) -> None:
bridge = _bridge_app()
active_stop_event = stop_event or bridge._background_stop_event()
while not active_stop_event.is_set():
try:
acw_sweep_once()
except Exception:
logger.exception("bridge.acw_sweep_failed")
if active_stop_event.wait(acw_sweep_interval_seconds()):
break
def failed_retry_loop(stop_event: threading.Event | None = None) -> None:
bridge = _bridge_app()
active_stop_event = stop_event or bridge._background_stop_event()