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(azure.py,-openai.py): correctly raise errors if streaming calls fail

Browse source
This commit is contained in:
Krrish Dholakia 2023-12-27 15:08:37 +05:30
parent 9ba520cc8b
commit c9fdbaf898
6 changed files with 110 additions and 24 deletions
Download patch file Download diff file Expand all files Collapse all files

100
litellm/tests/test_router_fallbacks.py
View file

@ -21,6 +21,13 @@ class MyCustomHandler(CustomLogger):
def log_pre_api_call(self, model, messages, kwargs):
print(f"Pre-API Call")
print(
f"previous_models: {kwargs['litellm_params']['metadata']['previous_models']}"
)
self.previous_models += len(
kwargs["litellm_params"]["metadata"]["previous_models"]
) # {"previous_models": [{"model": litellm_model_name, "exception_type": AuthenticationError, "exception_string": <complete_traceback>}]}
print(f"self.previous_models: {self.previous_models}")
def log_post_api_call(self, kwargs, response_obj, start_time, end_time):
print(
@ -34,13 +41,6 @@ class MyCustomHandler(CustomLogger):
print(f"On Stream")
def log_success_event(self, kwargs, response_obj, start_time, end_time):
print(
f"previous_models: {kwargs['litellm_params']['metadata']['previous_models']}"
)
self.previous_models += len(
kwargs["litellm_params"]["metadata"]["previous_models"]
) # {"previous_models": [{"model": litellm_model_name, "exception_type": AuthenticationError, "exception_string": <complete_traceback>}]}
print(f"self.previous_models: {self.previous_models}")
print(f"On Success")
async def async_log_success_event(self, kwargs, response_obj, start_time, end_time):
@ -396,3 +396,89 @@ async def test_dynamic_fallbacks_async():
# asyncio.run(test_dynamic_fallbacks_async())
@pytest.mark.asyncio
async def test_async_fallbacks_streaming():
litellm.set_verbose = False
model_list = [
{ # list of model deployments
"model_name": "azure/gpt-3.5-turbo", # openai model name
"litellm_params": { # params for litellm completion/embedding call
"model": "azure/chatgpt-v-2",
"api_key": "bad-key",
"api_version": os.getenv("AZURE_API_VERSION"),
"api_base": os.getenv("AZURE_API_BASE"),
},
"tpm": 240000,
"rpm": 1800,
},
{ # list of model deployments
"model_name": "azure/gpt-3.5-turbo-context-fallback", # openai model name
"litellm_params": { # params for litellm completion/embedding call
"model": "azure/chatgpt-v-2",
"api_key": os.getenv("AZURE_API_KEY"),
"api_version": os.getenv("AZURE_API_VERSION"),
"api_base": os.getenv("AZURE_API_BASE"),
},
"tpm": 240000,
"rpm": 1800,
},
{
"model_name": "azure/gpt-3.5-turbo", # openai model name
"litellm_params": { # params for litellm completion/embedding call
"model": "azure/chatgpt-functioncalling",
"api_key": "bad-key",
"api_version": os.getenv("AZURE_API_VERSION"),
"api_base": os.getenv("AZURE_API_BASE"),
},
"tpm": 240000,
"rpm": 1800,
},
{
"model_name": "gpt-3.5-turbo", # openai model name
"litellm_params": { # params for litellm completion/embedding call
"model": "gpt-3.5-turbo",
"api_key": os.getenv("OPENAI_API_KEY"),
},
"tpm": 1000000,
"rpm": 9000,
},
{
"model_name": "gpt-3.5-turbo-16k", # openai model name
"litellm_params": { # params for litellm completion/embedding call
"model": "gpt-3.5-turbo-16k",
"api_key": os.getenv("OPENAI_API_KEY"),
},
"tpm": 1000000,
"rpm": 9000,
},
]
router = Router(
model_list=model_list,
fallbacks=[{"azure/gpt-3.5-turbo": ["gpt-3.5-turbo"]}],
context_window_fallbacks=[
{"azure/gpt-3.5-turbo-context-fallback": ["gpt-3.5-turbo-16k"]},
{"gpt-3.5-turbo": ["gpt-3.5-turbo-16k"]},
],
set_verbose=False,
)
customHandler = MyCustomHandler()
litellm.callbacks = [customHandler]
user_message = "Hello, how are you?"
messages = [{"content": user_message, "role": "user"}]
try:
response = await router.acompletion(**kwargs, stream=True)
print(f"customHandler.previous_models: {customHandler.previous_models}")
await asyncio.sleep(
0.05
) # allow a delay as success_callbacks are on a separate thread
assert customHandler.previous_models == 1 # 0 retries, 1 fallback
router.reset()
except litellm.Timeout as e:
pass
except Exception as e:
pytest.fail(f"An exception occurred: {e}")
finally:
router.reset()