phoenix/litellm-mirror
0
0
Fork 1
mirror of https://github.com/BerriAI/litellm.git synced 2025-04-26 19:24:27 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

feat(utils.py): enable returning complete response when stream=true

Browse source
This commit is contained in:
Krrish Dholakia 2023-11-09 09:17:43 -08:00
parent 3dae4e9cda
commit 8ee4b1f603
4 changed files with 22 additions and 7 deletions
Download patch file Download diff file Expand all files Collapse all files

12
litellm/utils.py
View file

@ -949,16 +949,18 @@ def client(original_function):
end_time = datetime.datetime.now()
if "stream" in kwargs and kwargs["stream"] == True:
# TODO: Add to cache for streaming
return result
if "complete_response" in kwargs and kwargs["complete_response"] == True:
chunks = []
for idx, chunk in enumerate(result):
chunks.append(chunk)
return litellm.stream_chunk_builder(chunks)
else:
return result
# [OPTIONAL] ADD TO CACHE
if litellm.caching or litellm.caching_with_models or litellm.cache != None: # user init a cache object
litellm.cache.add_cache(result, *args, **kwargs)
# [OPTIONAL] Return LiteLLM call_id
if litellm.use_client == True:
result['litellm_call_id'] = litellm_call_id
# LOG SUCCESS - handle streaming success logging in the _next_ object, remove `handle_success` once it's deprecated
logging_obj.success_handler(result, start_time, end_time)