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

Add on_conversation_item_created and on_conversation_item_updated events to OpenAIRealtimeBetaLLMService.

Browse Source 
The hope is that this will expose to the user conversation item ids at relevant times for them to use with the new `conversation.item.retrieve` introspection message.
This commit is contained in:
Paul Kompfner
2025-03-17 16:52:53 -04:00
parent 9840abd85b
commit 39ca607bbb
1 changed files with 9 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files

9
src/pipecat/services/openai_realtime_beta/openai.py
View File

@@ -117,6 +117,9 @@ class OpenAIRealtimeBetaLLMService(LLMService):
self._messages_added_manually = {} self._messages_added_manually = {}
self._user_and_response_message_tuple = None self._user_and_response_message_tuple = None
self._register_event_handler("on_conversation_item_created")
self._register_event_handler("on_conversation_item_updated")
def can_generate_metrics(self) -> bool: def can_generate_metrics(self) -> bool:
return True return True
@@ -413,6 +416,8 @@ class OpenAIRealtimeBetaLLMService(LLMService):
# receive a BotStoppedSpeakingFrame from the output transport. # receive a BotStoppedSpeakingFrame from the output transport.
async def _handle_evt_conversation_item_created(self, evt): async def _handle_evt_conversation_item_created(self, evt):
await self._call_event_handler("on_conversation_item_created", evt.item.id, evt.item)
# This will get sent from the server every time a new "message" is added # This will get sent from the server every time a new "message" is added
# to the server's conversation state, whether we create it via the API # to the server's conversation state, whether we create it via the API
# or the server creates it from LLM output. # or the server creates it from LLM output.
@@ -437,6 +442,8 @@ class OpenAIRealtimeBetaLLMService(LLMService):
) )
async def handle_evt_input_audio_transcription_completed(self, evt): async def handle_evt_input_audio_transcription_completed(self, evt):
await self._call_event_handler("on_conversation_item_updated", evt.item_id, None)
if self._send_transcription_frames: if self._send_transcription_frames:
await self.push_frame( await self.push_frame(
# no way to get a language code? # no way to get a language code?
@@ -473,6 +480,8 @@ class OpenAIRealtimeBetaLLMService(LLMService):
) )
return return
# response content # response content
for item in evt.response.output:
await self._call_event_handler("on_conversation_item_updated", item.id, item)
pair = self._user_and_response_message_tuple pair = self._user_and_response_message_tuple
if pair: if pair:
user, assistant = pair user, assistant = pair