Warn when VAD stop_secs misconfiguration may degrade turn detection

Add warnings in SpeechTimeoutUserTurnStopStrategy and
TurnAnalyzerUserTurnStopStrategy when stop_secs differs from the
recommended default (0.2s) or when stop_secs >= STT p99 latency,
which collapses the STT wait timeout to 0s. Document the stop_secs=0.2
assumption in stt_latency.py.
This commit is contained in:
Mark Backman
2026-03-23 17:57:51 -04:00
parent 12dc429761
commit 483b643b07
4 changed files with 135 additions and 0 deletions

View File

@@ -6,6 +6,7 @@
import asyncio
import unittest
from unittest.mock import patch
from pipecat.frames.frames import (
InterimTranscriptionFrame,
@@ -538,6 +539,84 @@ class TestSpeechTimeoutUserTurnStopStrategy(unittest.IsolatedAsyncioTestCase):
self.assertEqual(stop_count, 2)
class TestSpeechTimeoutStopSecsWarnings(unittest.IsolatedAsyncioTestCase):
"""Tests for stop_secs misconfiguration warnings."""
async def asyncSetUp(self) -> None:
self.task_manager = TaskManager()
self.task_manager.setup(TaskManagerParams(loop=asyncio.get_running_loop()))
async def _create_strategy(self, stt_timeout=0.35):
strategy = SpeechTimeoutUserTurnStopStrategy(user_speech_timeout=AGGREGATION_TIMEOUT)
await strategy.setup(self.task_manager)
await strategy.process_frame(
STTMetadataFrame(service_name="test", ttfs_p99_latency=stt_timeout)
)
return strategy
@patch("pipecat.turns.user_stop.speech_timeout_user_turn_stop_strategy.logger")
async def test_warns_on_non_default_stop_secs(self, mock_logger):
# Use high stt_timeout so only Warning A fires (stop_secs < stt_timeout)
strategy = await self._create_strategy(stt_timeout=1.0)
await strategy.process_frame(VADUserStartedSpeakingFrame())
await strategy.process_frame(VADUserStoppedSpeakingFrame(stop_secs=0.5))
mock_logger.warning.assert_called_once()
self.assertIn("differs from the recommended default", mock_logger.warning.call_args[0][0])
@patch("pipecat.turns.user_stop.speech_timeout_user_turn_stop_strategy.logger")
async def test_warns_on_stop_secs_gte_stt_timeout(self, mock_logger):
strategy = await self._create_strategy(stt_timeout=0.35)
await strategy.process_frame(VADUserStartedSpeakingFrame())
await strategy.process_frame(VADUserStoppedSpeakingFrame(stop_secs=0.5))
# Both warnings fire: non-default stop_secs AND stop_secs >= stt_timeout
self.assertEqual(mock_logger.warning.call_count, 2)
self.assertIn("collapsed to 0s", mock_logger.warning.call_args_list[1][0][0])
@patch("pipecat.turns.user_stop.speech_timeout_user_turn_stop_strategy.logger")
async def test_warns_only_once(self, mock_logger):
# Use high stt_timeout so only Warning A fires
strategy = await self._create_strategy(stt_timeout=1.0)
# First VAD stop — triggers warning
await strategy.process_frame(VADUserStartedSpeakingFrame())
await strategy.process_frame(VADUserStoppedSpeakingFrame(stop_secs=0.5))
self.assertEqual(mock_logger.warning.call_count, 1)
# Second VAD stop — no duplicate warning
await strategy.process_frame(VADUserStartedSpeakingFrame())
await strategy.process_frame(VADUserStoppedSpeakingFrame(stop_secs=0.5))
self.assertEqual(mock_logger.warning.call_count, 1)
@patch("pipecat.turns.user_stop.speech_timeout_user_turn_stop_strategy.logger")
async def test_warning_resets_on_new_stt_metadata(self, mock_logger):
# Use high stt_timeout so only Warning A fires
strategy = await self._create_strategy(stt_timeout=1.0)
await strategy.process_frame(VADUserStartedSpeakingFrame())
await strategy.process_frame(VADUserStoppedSpeakingFrame(stop_secs=0.5))
self.assertEqual(mock_logger.warning.call_count, 1)
# New STTMetadataFrame resets the warned flag
await strategy.process_frame(STTMetadataFrame(service_name="test", ttfs_p99_latency=1.0))
await strategy.process_frame(VADUserStartedSpeakingFrame())
await strategy.process_frame(VADUserStoppedSpeakingFrame(stop_secs=0.5))
self.assertEqual(mock_logger.warning.call_count, 2)
@patch("pipecat.turns.user_stop.speech_timeout_user_turn_stop_strategy.logger")
async def test_no_warning_on_default_stop_secs(self, mock_logger):
strategy = await self._create_strategy()
await strategy.process_frame(VADUserStartedSpeakingFrame())
await strategy.process_frame(VADUserStoppedSpeakingFrame(stop_secs=0.2))
mock_logger.warning.assert_not_called()
class TestExternalUserTurnStopStrategy(unittest.IsolatedAsyncioTestCase):
async def test_external_strategy(self):
strategy = ExternalUserTurnStopStrategy()