mirror of
https://github.com/meta-llama/llama-stack.git
synced 2025-12-14 19:32:37 +00:00
bug fix
This commit is contained in:
parent
4017d5a7b9
commit
a707d2aaa9
1 changed files with 17 additions and 11 deletions
|
|
@ -345,18 +345,24 @@ class OpenAIResponsesImpl:
|
|||
failed_response = stream_chunk.response
|
||||
yield stream_chunk
|
||||
|
||||
# Store the response if requested
|
||||
if store and final_response and failed_response is None:
|
||||
await self._store_response(
|
||||
response=final_response,
|
||||
input=all_input,
|
||||
messages=orchestrator.final_messages,
|
||||
)
|
||||
# Store and sync immediately after yielding terminal events
|
||||
# This ensures the storage/syncing happens even if the consumer breaks early
|
||||
if (
|
||||
stream_chunk.type in {"response.completed", "response.incomplete"}
|
||||
and store
|
||||
and final_response
|
||||
and failed_response is None
|
||||
):
|
||||
await self._store_response(
|
||||
response=final_response,
|
||||
input=all_input,
|
||||
messages=orchestrator.final_messages,
|
||||
)
|
||||
|
||||
if conversation and final_response:
|
||||
# for Conversations, we need to use the original_input if it's available, otherwise use input
|
||||
sync_input = original_input if original_input is not None else input
|
||||
await self._sync_response_to_conversation(conversation, sync_input, final_response)
|
||||
if stream_chunk.type in {"response.completed", "response.incomplete"} and conversation and final_response:
|
||||
# for Conversations, we need to use the original_input if it's available, otherwise use input
|
||||
sync_input = original_input if original_input is not None else input
|
||||
await self._sync_response_to_conversation(conversation, sync_input, final_response)
|
||||
|
||||
async def delete_openai_response(self, response_id: str) -> OpenAIDeleteResponseObject:
|
||||
return await self.responses_store.delete_response_object(response_id)
|
||||
|
|
|
|||
Loading…
Add table
Add a link
Reference in a new issue