phoenix/litellm
1
0
Fork 0
forked from phoenix/litellm-mirror
Code Pull requests Activity

Merge pull request #1557 from BerriAI/litellm_emit_spend_logs

Browse source 
feat(utils.py): emit response cost as part of logs
This commit is contained in:
Krish Dholakia 2024-01-22 21:02:40 -08:00 • committed by GitHub
commit bedb08bdef
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
10 changed files with 191 additions and 46 deletions
Download patch file Download diff file Expand all files Collapse all files

38
litellm/proxy/proxy_server.py
View file

@ -570,13 +570,8 @@ async def track_cost_callback(
litellm_params = kwargs.get("litellm_params", {}) or {}
proxy_server_request = litellm_params.get("proxy_server_request") or {}
user_id = proxy_server_request.get("body", {}).get("user", None)
if "complete_streaming_response" in kwargs:
# for tracking streaming cost we pass the "messages" and the output_text to litellm.completion_cost
completion_response = kwargs["complete_streaming_response"]
response_cost = litellm.completion_cost(
completion_response=completion_response
)
if "response_cost" in kwargs:
response_cost = kwargs["response_cost"]
user_api_key = kwargs["litellm_params"]["metadata"].get(
"user_api_key", None
)
@ -585,31 +580,6 @@ async def track_cost_callback(
"user_api_key_user_id", None
)
verbose_proxy_logger.info(
f"streaming response_cost {response_cost}, for user_id {user_id}"
)
if user_api_key and (
prisma_client is not None or custom_db_client is not None
):
await update_database(
token=user_api_key,
response_cost=response_cost,
user_id=user_id,
kwargs=kwargs,
completion_response=completion_response,
start_time=start_time,
end_time=end_time,
)
elif kwargs["stream"] == False: # for non streaming responses
response_cost = litellm.completion_cost(
completion_response=completion_response
)
user_api_key = kwargs["litellm_params"]["metadata"].get(
"user_api_key", None
)
user_id = user_id or kwargs["litellm_params"]["metadata"].get(
"user_api_key_user_id", None
)
verbose_proxy_logger.info(
f"response_cost {response_cost}, for user_id {user_id}"
)
@ -625,6 +595,10 @@ async def track_cost_callback(
start_time=start_time,
end_time=end_time,
)
else:
raise Exception(
f"Model not in litellm model cost map. Add custom pricing - https://docs.litellm.ai/docs/proxy/custom_pricing"
)
except Exception as e:
verbose_proxy_logger.debug(f"error in tracking cost callback - {str(e)}")