wx44wx/pipecat
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Correct docstrings and comments regarding incomplete_long_timeout duration, 10 sec

Browse Source
This commit is contained in:
Mark Backman
2026-05-07 09:14:53 -04:00
committed by Aleix Conchillo Flaqué
parent b78cecf7b2
commit 457a68ce64
3 changed files with 6 additions and 6 deletions
Download Patch File Download Diff File Expand all files Collapse all files

6
examples/turn-management/turn-management-filter-incomplete-turns.py
View File

@@ -10,7 +10,7 @@ Demonstrates LLM-based turn completion detection to suppress bot responses when
the user was cut off mid-thought. The LLM outputs one of three markers: the user was cut off mid-thought. The LLM outputs one of three markers:
- ✓ (complete): User finished their thought, respond normally - ✓ (complete): User finished their thought, respond normally
- ○ (incomplete short): User was cut off, wait ~5s then prompt - ○ (incomplete short): User was cut off, wait ~5s then prompt
- ◐ (incomplete long): User needs time to think, wait ~15s then prompt - ◐ (incomplete long): User needs time to think, wait ~10s then prompt
When incomplete is detected, the bot's response is suppressed. After the timeout When incomplete is detected, the bot's response is suppressed. After the timeout
expires, the LLM is automatically prompted to re-engage the user. expires, the LLM is automatically prompted to re-engage the user.
@@ -91,7 +91,7 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
# with one of: # with one of:
# ✓ = complete (respond normally) # ✓ = complete (respond normally)
# ○ = incomplete short (wait 5s, then prompt) # ○ = incomplete short (wait 5s, then prompt)
# ◐ = incomplete long (wait 15s, then prompt) # ◐ = incomplete long (wait 10s, then prompt)
user_aggregator, assistant_aggregator = LLMContextAggregatorPair( user_aggregator, assistant_aggregator = LLMContextAggregatorPair(
context, context,
user_params=LLMUserAggregatorParams( user_params=LLMUserAggregatorParams(
@@ -100,7 +100,7 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
# Optional: customize turn completion behavior # Optional: customize turn completion behavior
# config=UserTurnCompletionConfig( # config=UserTurnCompletionConfig(
# incomplete_short_timeout=5.0, # incomplete_short_timeout=5.0,
# incomplete_long_timeout=15.0, # incomplete_long_timeout=10.0,
# incomplete_short_prompt="Custom prompt...", # incomplete_short_prompt="Custom prompt...",
# incomplete_long_prompt="Custom prompt...", # incomplete_long_prompt="Custom prompt...",
# instructions="Custom turn completion instructions...", # instructions="Custom turn completion instructions...",

4
src/pipecat/turns/user_turn_completion_mixin.py
View File

@@ -187,8 +187,8 @@ class UserTurnCompletionLLMServiceMixin(FrameProcessor):
It processes turn completion markers to enable smarter conversation flow: It processes turn completion markers to enable smarter conversation flow:
- ✓ (COMPLETE): Push response normally - ✓ (COMPLETE): Push response normally
- ○ (INCOMPLETE SHORT): Suppress response, wait ~5s, then prompt - ○ (INCOMPLETE SHORT): Suppress response, wait 5s, then prompt
- ◐ (INCOMPLETE LONG): Suppress response, wait ~15s, then prompt - ◐ (INCOMPLETE LONG): Suppress response, wait 10s, then prompt
When incomplete timeouts expire, the mixin automatically prompts the LLM When incomplete timeouts expire, the mixin automatically prompts the LLM
with a contextual follow-up message to re-engage the user. with a contextual follow-up message to re-engage the user.

2
src/pipecat/turns/user_turn_strategies.py
View File

@@ -138,7 +138,7 @@ class FilterIncompleteUserTurnStrategies(UserTurnStrategies):
user_turn_strategies=FilterIncompleteUserTurnStrategies( user_turn_strategies=FilterIncompleteUserTurnStrategies(
config=UserTurnCompletionConfig( config=UserTurnCompletionConfig(
incomplete_short_timeout=5.0, incomplete_short_timeout=5.0,
incomplete_long_timeout=15.0, incomplete_long_timeout=10.0,
), ),
) )
""" """