phoenix/litellm-mirror
0
0
Fork 1
mirror of https://github.com/BerriAI/litellm.git synced 2025-04-25 18:54:30 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

fix(route_llm_request.py): move to using common router, even for clie… (#8966)

Browse source 
* fix(route_llm_request.py): move to using common router, even for client-side credentials

ensures fallbacks / cooldown logic still works

* test(test_route_llm_request.py): add unit test for route request

* feat(router.py): generate unique model id when clientside credential passed in

Prevents cooldowns for api key 1 from impacting api key 2

* test(test_router.py): update testing to ensure original litellm params not mutated

* fix(router.py): upsert clientside call into llm router model list

enables cooldown logic to work accurately

* fix: fix linting error

* test(test_router_utils.py): add direct test for new util on router
This commit is contained in:
Krish Dholakia 2025-03-03 22:57:08 -08:00 • committed by GitHub
parent 94563ab1e7
commit 2fc6262675
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
9 changed files with 273 additions and 36 deletions
Download patch file Download diff file Expand all files Collapse all files

57
litellm/router.py
View file

@ -67,6 +67,10 @@ from litellm.router_utils.batch_utils import (
replace_model_in_jsonl,
)
from litellm.router_utils.client_initalization_utils import InitalizeOpenAISDKClient
from litellm.router_utils.clientside_credential_handler import (
get_dynamic_litellm_params,
is_clientside_credential,
)
from litellm.router_utils.cooldown_cache import CooldownCache
from litellm.router_utils.cooldown_handlers import (
DEFAULT_COOLDOWN_TIME_SECONDS,
@ -1067,20 +1071,61 @@ class Router:
elif k == "metadata":
kwargs[k].update(v)
def _handle_clientside_credential(
self, deployment: dict, kwargs: dict
) -> Deployment:
"""
Handle clientside credential
"""
model_info = deployment.get("model_info", {}).copy()
litellm_params = deployment["litellm_params"].copy()
dynamic_litellm_params = get_dynamic_litellm_params(
litellm_params=litellm_params, request_kwargs=kwargs
)
metadata = kwargs.get("metadata", {})
model_group = cast(str, metadata.get("model_group"))
_model_id = self._generate_model_id(
model_group=model_group, litellm_params=dynamic_litellm_params
)
original_model_id = model_info.get("id")
model_info["id"] = _model_id
model_info["original_model_id"] = original_model_id
deployment_pydantic_obj = Deployment(
model_name=model_group,
litellm_params=LiteLLM_Params(**dynamic_litellm_params),
model_info=model_info,
)
self.upsert_deployment(
deployment=deployment_pydantic_obj
) # add new deployment to router
return deployment_pydantic_obj
def _update_kwargs_with_deployment(self, deployment: dict, kwargs: dict) -> None:
"""
2 jobs:
- Adds selected deployment, model_info and api_base to kwargs["metadata"] (used for logging)
- Adds default litellm params to kwargs, if set.
"""
model_info = deployment.get("model_info", {}).copy()
deployment_model_name = deployment["litellm_params"]["model"]
deployment_api_base = deployment["litellm_params"].get("api_base")
if is_clientside_credential(request_kwargs=kwargs):
deployment_pydantic_obj = self._handle_clientside_credential(
deployment=deployment, kwargs=kwargs
)
model_info = deployment_pydantic_obj.model_info.model_dump()
deployment_model_name = deployment_pydantic_obj.litellm_params.model
deployment_api_base = deployment_pydantic_obj.litellm_params.api_base
kwargs.setdefault("metadata", {}).update(
{
"deployment": deployment["litellm_params"]["model"],
"model_info": deployment.get("model_info", {}),
"api_base": deployment.get("litellm_params", {}).get("api_base"),
"deployment": deployment_model_name,
"model_info": model_info,
"api_base": deployment_api_base,
}
)
kwargs["model_info"] = deployment.get("model_info", {})
kwargs["model_info"] = model_info
kwargs["timeout"] = self._get_timeout(
kwargs=kwargs, data=deployment["litellm_params"]
)
@ -3601,6 +3646,7 @@ class Router:
- True if the deployment should be put in cooldown
- False if the deployment should not be put in cooldown
"""
verbose_router_logger.debug("Router: Entering 'deployment_callback_on_failure'")
try:
exception = kwargs.get("exception", None)
exception_status = getattr(exception, "status_code", "")
@ -3642,6 +3688,9 @@ class Router:
return result
else:
verbose_router_logger.debug(
"Router: Exiting 'deployment_callback_on_failure' without cooldown. No model_info found."
)
return False
except Exception as e: