phoenix/litellm-mirror
0
0
Fork 1
mirror of https://github.com/BerriAI/litellm.git synced 2025-04-27 11:43:54 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

fix(azure_dall_e_2.py): handle azure not returning a 'retry-after' param

Browse source
This commit is contained in:
Krrish Dholakia 2024-01-22 12:02:02 -08:00
parent 11bd726942
commit 4e89be0e19
2 changed files with 17 additions and 9 deletions
Download patch file Download diff file Expand all files Collapse all files

19
litellm/llms/azure.py
View file

@ -629,12 +629,23 @@ class AzureChatCompletion(BaseLLM):
client_session = litellm.aclient_session or httpx.AsyncClient(
transport=AsyncCustomHTTPTransport(),
)
openai_aclient = AsyncAzureOpenAI(
azure_client = AsyncAzureOpenAI(
http_client=client_session, **azure_client_params
)
else:
openai_aclient = client
response = await openai_aclient.images.generate(**data, timeout=timeout)
azure_client = client
## LOGGING
logging_obj.pre_call(
input=data["prompt"],
api_key=azure_client.api_key,
additional_args={
"headers": {"api_key": azure_client.api_key},
"api_base": azure_client._base_url._uri_reference,
"acompletion": True,
"complete_input_dict": data,
},
)
response = await azure_client.images.generate(**data, timeout=timeout)
stringified_response = response.model_dump()
## LOGGING
logging_obj.post_call(
@ -719,7 +730,7 @@ class AzureChatCompletion(BaseLLM):
input=prompt,
api_key=azure_client.api_key,
additional_args={
"headers": {"Authorization": f"Bearer {azure_client.api_key}"},
"headers": {"api_key": azure_client.api_key},
"api_base": azure_client._base_url._uri_reference,
"acompletion": False,
"complete_input_dict": data,

7
litellm/llms/custom_httpx/azure_dall_e_2.py
View file

@ -43,7 +43,7 @@ class AsyncCustomHTTPTransport(httpx.AsyncHTTPTransport):
request=request,
)
time.sleep(int(response.headers.get("retry-after")) or 10)
await asyncio.sleep(int(response.headers.get("retry-after") or 10))
response = await super().handle_async_request(request)
await response.aread()
@ -95,7 +95,6 @@ class CustomHTTPTransport(httpx.HTTPTransport):
request.method = "GET"
response = super().handle_request(request)
response.read()
timeout_secs: int = 120
start_time = time.time()
while response.json()["status"] not in ["succeeded", "failed"]:
@ -112,11 +111,9 @@ class CustomHTTPTransport(httpx.HTTPTransport):
content=json.dumps(timeout).encode("utf-8"),
request=request,
)
time.sleep(int(response.headers.get("retry-after")) or 10)
time.sleep(int(response.headers.get("retry-after", None) or 10))
response = super().handle_request(request)
response.read()
if response.json()["status"] == "failed":
error_data = response.json()
return httpx.Response(