mirror of
https://github.com/BerriAI/litellm.git
synced 2025-04-26 03:04:13 +00:00
Merge pull request #3936 from BerriAI/litellm_assistants_api_proxy
feat(proxy_server.py): add assistants api endpoints to proxy server
This commit is contained in:
Krish Dholakia
GitHub
commit
08bae3185a
6 changed files with 1741 additions and 50 deletions
|
|
@ -5306,6 +5306,765 @@ async def audio_transcriptions(
|
|||
)
|
||||
|
||||
|
||||
######################################################################
|
||||
|
||||
# /v1/assistant Endpoints
|
||||
|
||||
|
||||
######################################################################
|
||||
|
||||
|
||||
@router.get(
|
||||
"/v1/assistants",
|
||||
dependencies=[Depends(user_api_key_auth)],
|
||||
tags=["assistants"],
|
||||
)
|
||||
@router.get(
|
||||
"/assistants",
|
||||
dependencies=[Depends(user_api_key_auth)],
|
||||
tags=["assistants"],
|
||||
)
|
||||
async def get_assistants(
|
||||
request: Request,
|
||||
fastapi_response: Response,
|
||||
user_api_key_dict: UserAPIKeyAuth = Depends(user_api_key_auth),
|
||||
):
|
||||
global proxy_logging_obj
|
||||
data: Dict = {}
|
||||
try:
|
||||
# Use orjson to parse JSON data, orjson speeds up requests significantly
|
||||
body = await request.body()
|
||||
data = orjson.loads(body)
|
||||
|
||||
# Include original request and headers in the data
|
||||
data["proxy_server_request"] = { # type: ignore
|
||||
"url": str(request.url),
|
||||
"method": request.method,
|
||||
"headers": dict(request.headers),
|
||||
"body": copy.copy(data), # use copy instead of deepcopy
|
||||
}
|
||||
|
||||
if data.get("user", None) is None and user_api_key_dict.user_id is not None:
|
||||
data["user"] = user_api_key_dict.user_id
|
||||
|
||||
if "metadata" not in data:
|
||||
data["metadata"] = {}
|
||||
data["metadata"]["user_api_key"] = user_api_key_dict.api_key
|
||||
data["metadata"]["user_api_key_metadata"] = user_api_key_dict.metadata
|
||||
_headers = dict(request.headers)
|
||||
_headers.pop(
|
||||
"authorization", None
|
||||
) # do not store the original `sk-..` api key in the db
|
||||
data["metadata"]["headers"] = _headers
|
||||
data["metadata"]["user_api_key_alias"] = getattr(
|
||||
user_api_key_dict, "key_alias", None
|
||||
)
|
||||
data["metadata"]["user_api_key_user_id"] = user_api_key_dict.user_id
|
||||
data["metadata"]["user_api_key_team_id"] = getattr(
|
||||
user_api_key_dict, "team_id", None
|
||||
)
|
||||
data["metadata"]["global_max_parallel_requests"] = general_settings.get(
|
||||
"global_max_parallel_requests", None
|
||||
)
|
||||
data["metadata"]["user_api_key_team_alias"] = getattr(
|
||||
user_api_key_dict, "team_alias", None
|
||||
)
|
||||
data["metadata"]["endpoint"] = str(request.url)
|
||||
|
||||
### TEAM-SPECIFIC PARAMS ###
|
||||
if user_api_key_dict.team_id is not None:
|
||||
team_config = await proxy_config.load_team_config(
|
||||
team_id=user_api_key_dict.team_id
|
||||
)
|
||||
if len(team_config) == 0:
|
||||
pass
|
||||
else:
|
||||
team_id = team_config.pop("team_id", None)
|
||||
data["metadata"]["team_id"] = team_id
|
||||
data = {
|
||||
**team_config,
|
||||
**data,
|
||||
} # add the team-specific configs to the completion call
|
||||
|
||||
# for now use custom_llm_provider=="openai" -> this will change as LiteLLM adds more providers for acreate_batch
|
||||
if llm_router is None:
|
||||
raise HTTPException(
|
||||
status_code=500, detail={"error": CommonProxyErrors.no_llm_router.value}
|
||||
)
|
||||
response = await llm_router.aget_assistants(
|
||||
custom_llm_provider="openai", client=None, **data
|
||||
)
|
||||
|
||||
### ALERTING ###
|
||||
data["litellm_status"] = "success" # used for alerting
|
||||
|
||||
### RESPONSE HEADERS ###
|
||||
hidden_params = getattr(response, "_hidden_params", {}) or {}
|
||||
model_id = hidden_params.get("model_id", None) or ""
|
||||
cache_key = hidden_params.get("cache_key", None) or ""
|
||||
api_base = hidden_params.get("api_base", None) or ""
|
||||
|
||||
fastapi_response.headers.update(
|
||||
get_custom_headers(
|
||||
user_api_key_dict=user_api_key_dict,
|
||||
model_id=model_id,
|
||||
cache_key=cache_key,
|
||||
api_base=api_base,
|
||||
version=version,
|
||||
model_region=getattr(user_api_key_dict, "allowed_model_region", ""),
|
||||
)
|
||||
)
|
||||
|
||||
return response
|
||||
except Exception as e:
|
||||
data["litellm_status"] = "fail" # used for alerting
|
||||
await proxy_logging_obj.post_call_failure_hook(
|
||||
user_api_key_dict=user_api_key_dict, original_exception=e, request_data=data
|
||||
)
|
||||
traceback.print_exc()
|
||||
if isinstance(e, HTTPException):
|
||||
raise ProxyException(
|
||||
message=getattr(e, "message", str(e.detail)),
|
||||
type=getattr(e, "type", "None"),
|
||||
param=getattr(e, "param", "None"),
|
||||
code=getattr(e, "status_code", status.HTTP_400_BAD_REQUEST),
|
||||
)
|
||||
else:
|
||||
error_traceback = traceback.format_exc()
|
||||
error_msg = f"{str(e)}"
|
||||
raise ProxyException(
|
||||
message=getattr(e, "message", error_msg),
|
||||
type=getattr(e, "type", "None"),
|
||||
param=getattr(e, "param", "None"),
|
||||
code=getattr(e, "status_code", 500),
|
||||
)
|
||||
|
||||
|
||||
@router.post(
|
||||
"/v1/threads",
|
||||
dependencies=[Depends(user_api_key_auth)],
|
||||
tags=["assistants"],
|
||||
)
|
||||
@router.post(
|
||||
"/threads",
|
||||
dependencies=[Depends(user_api_key_auth)],
|
||||
tags=["assistants"],
|
||||
)
|
||||
async def create_threads(
|
||||
request: Request,
|
||||
fastapi_response: Response,
|
||||
user_api_key_dict: UserAPIKeyAuth = Depends(user_api_key_auth),
|
||||
):
|
||||
global proxy_logging_obj
|
||||
data: Dict = {}
|
||||
try:
|
||||
# Use orjson to parse JSON data, orjson speeds up requests significantly
|
||||
body = await request.body()
|
||||
data = orjson.loads(body)
|
||||
|
||||
# Include original request and headers in the data
|
||||
data["proxy_server_request"] = { # type: ignore
|
||||
"url": str(request.url),
|
||||
"method": request.method,
|
||||
"headers": dict(request.headers),
|
||||
"body": copy.copy(data), # use copy instead of deepcopy
|
||||
}
|
||||
|
||||
if data.get("user", None) is None and user_api_key_dict.user_id is not None:
|
||||
data["user"] = user_api_key_dict.user_id
|
||||
|
||||
if "litellm_metadata" not in data:
|
||||
data["litellm_metadata"] = {}
|
||||
data["litellm_metadata"]["user_api_key"] = user_api_key_dict.api_key
|
||||
data["litellm_metadata"]["user_api_key_metadata"] = user_api_key_dict.metadata
|
||||
_headers = dict(request.headers)
|
||||
_headers.pop(
|
||||
"authorization", None
|
||||
) # do not store the original `sk-..` api key in the db
|
||||
data["litellm_metadata"]["headers"] = _headers
|
||||
data["litellm_metadata"]["user_api_key_alias"] = getattr(
|
||||
user_api_key_dict, "key_alias", None
|
||||
)
|
||||
data["litellm_metadata"]["user_api_key_user_id"] = user_api_key_dict.user_id
|
||||
data["litellm_metadata"]["user_api_key_team_id"] = getattr(
|
||||
user_api_key_dict, "team_id", None
|
||||
)
|
||||
data["litellm_metadata"]["global_max_parallel_requests"] = general_settings.get(
|
||||
"global_max_parallel_requests", None
|
||||
)
|
||||
data["litellm_metadata"]["user_api_key_team_alias"] = getattr(
|
||||
user_api_key_dict, "team_alias", None
|
||||
)
|
||||
data["litellm_metadata"]["endpoint"] = str(request.url)
|
||||
|
||||
### TEAM-SPECIFIC PARAMS ###
|
||||
if user_api_key_dict.team_id is not None:
|
||||
team_config = await proxy_config.load_team_config(
|
||||
team_id=user_api_key_dict.team_id
|
||||
)
|
||||
if len(team_config) == 0:
|
||||
pass
|
||||
else:
|
||||
team_id = team_config.pop("team_id", None)
|
||||
data["litellm_metadata"]["team_id"] = team_id
|
||||
data = {
|
||||
**team_config,
|
||||
**data,
|
||||
} # add the team-specific configs to the completion call
|
||||
|
||||
# for now use custom_llm_provider=="openai" -> this will change as LiteLLM adds more providers for acreate_batch
|
||||
if llm_router is None:
|
||||
raise HTTPException(
|
||||
status_code=500, detail={"error": CommonProxyErrors.no_llm_router.value}
|
||||
)
|
||||
response = await llm_router.acreate_thread(
|
||||
custom_llm_provider="openai", client=None, **data
|
||||
)
|
||||
|
||||
### ALERTING ###
|
||||
data["litellm_status"] = "success" # used for alerting
|
||||
|
||||
### RESPONSE HEADERS ###
|
||||
hidden_params = getattr(response, "_hidden_params", {}) or {}
|
||||
model_id = hidden_params.get("model_id", None) or ""
|
||||
cache_key = hidden_params.get("cache_key", None) or ""
|
||||
api_base = hidden_params.get("api_base", None) or ""
|
||||
|
||||
fastapi_response.headers.update(
|
||||
get_custom_headers(
|
||||
user_api_key_dict=user_api_key_dict,
|
||||
model_id=model_id,
|
||||
cache_key=cache_key,
|
||||
api_base=api_base,
|
||||
version=version,
|
||||
model_region=getattr(user_api_key_dict, "allowed_model_region", ""),
|
||||
)
|
||||
)
|
||||
|
||||
return response
|
||||
except Exception as e:
|
||||
data["litellm_status"] = "fail" # used for alerting
|
||||
await proxy_logging_obj.post_call_failure_hook(
|
||||
user_api_key_dict=user_api_key_dict, original_exception=e, request_data=data
|
||||
)
|
||||
traceback.print_exc()
|
||||
if isinstance(e, HTTPException):
|
||||
raise ProxyException(
|
||||
message=getattr(e, "message", str(e.detail)),
|
||||
type=getattr(e, "type", "None"),
|
||||
param=getattr(e, "param", "None"),
|
||||
code=getattr(e, "status_code", status.HTTP_400_BAD_REQUEST),
|
||||
)
|
||||
else:
|
||||
error_traceback = traceback.format_exc()
|
||||
error_msg = f"{str(e)}"
|
||||
raise ProxyException(
|
||||
message=getattr(e, "message", error_msg),
|
||||
type=getattr(e, "type", "None"),
|
||||
param=getattr(e, "param", "None"),
|
||||
code=getattr(e, "status_code", 500),
|
||||
)
|
||||
|
||||
|
||||
@router.get(
|
||||
"/v1/threads/{thread_id}",
|
||||
dependencies=[Depends(user_api_key_auth)],
|
||||
tags=["assistants"],
|
||||
)
|
||||
@router.get(
|
||||
"/threads/{thread_id}",
|
||||
dependencies=[Depends(user_api_key_auth)],
|
||||
tags=["assistants"],
|
||||
)
|
||||
async def get_thread(
|
||||
request: Request,
|
||||
thread_id: str,
|
||||
fastapi_response: Response,
|
||||
user_api_key_dict: UserAPIKeyAuth = Depends(user_api_key_auth),
|
||||
):
|
||||
global proxy_logging_obj
|
||||
data: Dict = {}
|
||||
try:
|
||||
|
||||
# Include original request and headers in the data
|
||||
data["proxy_server_request"] = { # type: ignore
|
||||
"url": str(request.url),
|
||||
"method": request.method,
|
||||
"headers": dict(request.headers),
|
||||
"body": copy.copy(data), # use copy instead of deepcopy
|
||||
}
|
||||
|
||||
if data.get("user", None) is None and user_api_key_dict.user_id is not None:
|
||||
data["user"] = user_api_key_dict.user_id
|
||||
|
||||
if "metadata" not in data:
|
||||
data["metadata"] = {}
|
||||
data["metadata"]["user_api_key"] = user_api_key_dict.api_key
|
||||
data["metadata"]["user_api_key_metadata"] = user_api_key_dict.metadata
|
||||
_headers = dict(request.headers)
|
||||
_headers.pop(
|
||||
"authorization", None
|
||||
) # do not store the original `sk-..` api key in the db
|
||||
data["metadata"]["headers"] = _headers
|
||||
data["metadata"]["user_api_key_alias"] = getattr(
|
||||
user_api_key_dict, "key_alias", None
|
||||
)
|
||||
data["metadata"]["user_api_key_user_id"] = user_api_key_dict.user_id
|
||||
data["metadata"]["user_api_key_team_id"] = getattr(
|
||||
user_api_key_dict, "team_id", None
|
||||
)
|
||||
data["metadata"]["global_max_parallel_requests"] = general_settings.get(
|
||||
"global_max_parallel_requests", None
|
||||
)
|
||||
data["metadata"]["user_api_key_team_alias"] = getattr(
|
||||
user_api_key_dict, "team_alias", None
|
||||
)
|
||||
data["metadata"]["endpoint"] = str(request.url)
|
||||
|
||||
### TEAM-SPECIFIC PARAMS ###
|
||||
if user_api_key_dict.team_id is not None:
|
||||
team_config = await proxy_config.load_team_config(
|
||||
team_id=user_api_key_dict.team_id
|
||||
)
|
||||
if len(team_config) == 0:
|
||||
pass
|
||||
else:
|
||||
team_id = team_config.pop("team_id", None)
|
||||
data["metadata"]["team_id"] = team_id
|
||||
data = {
|
||||
**team_config,
|
||||
**data,
|
||||
} # add the team-specific configs to the completion call
|
||||
|
||||
# for now use custom_llm_provider=="openai" -> this will change as LiteLLM adds more providers for acreate_batch
|
||||
if llm_router is None:
|
||||
raise HTTPException(
|
||||
status_code=500, detail={"error": CommonProxyErrors.no_llm_router.value}
|
||||
)
|
||||
response = await llm_router.aget_thread(
|
||||
custom_llm_provider="openai", thread_id=thread_id, client=None, **data
|
||||
)
|
||||
|
||||
### ALERTING ###
|
||||
data["litellm_status"] = "success" # used for alerting
|
||||
|
||||
### RESPONSE HEADERS ###
|
||||
hidden_params = getattr(response, "_hidden_params", {}) or {}
|
||||
model_id = hidden_params.get("model_id", None) or ""
|
||||
cache_key = hidden_params.get("cache_key", None) or ""
|
||||
api_base = hidden_params.get("api_base", None) or ""
|
||||
|
||||
fastapi_response.headers.update(
|
||||
get_custom_headers(
|
||||
user_api_key_dict=user_api_key_dict,
|
||||
model_id=model_id,
|
||||
cache_key=cache_key,
|
||||
api_base=api_base,
|
||||
version=version,
|
||||
model_region=getattr(user_api_key_dict, "allowed_model_region", ""),
|
||||
)
|
||||
)
|
||||
|
||||
return response
|
||||
except Exception as e:
|
||||
data["litellm_status"] = "fail" # used for alerting
|
||||
await proxy_logging_obj.post_call_failure_hook(
|
||||
user_api_key_dict=user_api_key_dict, original_exception=e, request_data=data
|
||||
)
|
||||
traceback.print_exc()
|
||||
if isinstance(e, HTTPException):
|
||||
raise ProxyException(
|
||||
message=getattr(e, "message", str(e.detail)),
|
||||
type=getattr(e, "type", "None"),
|
||||
param=getattr(e, "param", "None"),
|
||||
code=getattr(e, "status_code", status.HTTP_400_BAD_REQUEST),
|
||||
)
|
||||
else:
|
||||
error_traceback = traceback.format_exc()
|
||||
error_msg = f"{str(e)}"
|
||||
raise ProxyException(
|
||||
message=getattr(e, "message", error_msg),
|
||||
type=getattr(e, "type", "None"),
|
||||
param=getattr(e, "param", "None"),
|
||||
code=getattr(e, "status_code", 500),
|
||||
)
|
||||
|
||||
|
||||
@router.post(
|
||||
"/v1/threads/{thread_id}/messages",
|
||||
dependencies=[Depends(user_api_key_auth)],
|
||||
tags=["assistants"],
|
||||
)
|
||||
@router.post(
|
||||
"/threads/{thread_id}/messages",
|
||||
dependencies=[Depends(user_api_key_auth)],
|
||||
tags=["assistants"],
|
||||
)
|
||||
async def add_messages(
|
||||
request: Request,
|
||||
thread_id: str,
|
||||
fastapi_response: Response,
|
||||
user_api_key_dict: UserAPIKeyAuth = Depends(user_api_key_auth),
|
||||
):
|
||||
global proxy_logging_obj
|
||||
data: Dict = {}
|
||||
try:
|
||||
# Use orjson to parse JSON data, orjson speeds up requests significantly
|
||||
body = await request.body()
|
||||
data = orjson.loads(body)
|
||||
|
||||
# Include original request and headers in the data
|
||||
data["proxy_server_request"] = { # type: ignore
|
||||
"url": str(request.url),
|
||||
"method": request.method,
|
||||
"headers": dict(request.headers),
|
||||
"body": copy.copy(data), # use copy instead of deepcopy
|
||||
}
|
||||
|
||||
if data.get("user", None) is None and user_api_key_dict.user_id is not None:
|
||||
data["user"] = user_api_key_dict.user_id
|
||||
|
||||
if "litellm_metadata" not in data:
|
||||
data["litellm_metadata"] = {}
|
||||
data["litellm_metadata"]["user_api_key"] = user_api_key_dict.api_key
|
||||
data["litellm_metadata"]["user_api_key_metadata"] = user_api_key_dict.metadata
|
||||
_headers = dict(request.headers)
|
||||
_headers.pop(
|
||||
"authorization", None
|
||||
) # do not store the original `sk-..` api key in the db
|
||||
data["litellm_metadata"]["headers"] = _headers
|
||||
data["litellm_metadata"]["user_api_key_alias"] = getattr(
|
||||
user_api_key_dict, "key_alias", None
|
||||
)
|
||||
data["litellm_metadata"]["user_api_key_user_id"] = user_api_key_dict.user_id
|
||||
data["litellm_metadata"]["user_api_key_team_id"] = getattr(
|
||||
user_api_key_dict, "team_id", None
|
||||
)
|
||||
data["litellm_metadata"]["global_max_parallel_requests"] = general_settings.get(
|
||||
"global_max_parallel_requests", None
|
||||
)
|
||||
data["litellm_metadata"]["user_api_key_team_alias"] = getattr(
|
||||
user_api_key_dict, "team_alias", None
|
||||
)
|
||||
data["litellm_metadata"]["endpoint"] = str(request.url)
|
||||
|
||||
### TEAM-SPECIFIC PARAMS ###
|
||||
if user_api_key_dict.team_id is not None:
|
||||
team_config = await proxy_config.load_team_config(
|
||||
team_id=user_api_key_dict.team_id
|
||||
)
|
||||
if len(team_config) == 0:
|
||||
pass
|
||||
else:
|
||||
team_id = team_config.pop("team_id", None)
|
||||
data["litellm_metadata"]["team_id"] = team_id
|
||||
data = {
|
||||
**team_config,
|
||||
**data,
|
||||
} # add the team-specific configs to the completion call
|
||||
|
||||
# for now use custom_llm_provider=="openai" -> this will change as LiteLLM adds more providers for acreate_batch
|
||||
if llm_router is None:
|
||||
raise HTTPException(
|
||||
status_code=500, detail={"error": CommonProxyErrors.no_llm_router.value}
|
||||
)
|
||||
response = await llm_router.a_add_message(
|
||||
custom_llm_provider="openai", thread_id=thread_id, client=None, **data
|
||||
)
|
||||
|
||||
### ALERTING ###
|
||||
data["litellm_status"] = "success" # used for alerting
|
||||
|
||||
### RESPONSE HEADERS ###
|
||||
hidden_params = getattr(response, "_hidden_params", {}) or {}
|
||||
model_id = hidden_params.get("model_id", None) or ""
|
||||
cache_key = hidden_params.get("cache_key", None) or ""
|
||||
api_base = hidden_params.get("api_base", None) or ""
|
||||
|
||||
fastapi_response.headers.update(
|
||||
get_custom_headers(
|
||||
user_api_key_dict=user_api_key_dict,
|
||||
model_id=model_id,
|
||||
cache_key=cache_key,
|
||||
api_base=api_base,
|
||||
version=version,
|
||||
model_region=getattr(user_api_key_dict, "allowed_model_region", ""),
|
||||
)
|
||||
)
|
||||
|
||||
return response
|
||||
except Exception as e:
|
||||
data["litellm_status"] = "fail" # used for alerting
|
||||
await proxy_logging_obj.post_call_failure_hook(
|
||||
user_api_key_dict=user_api_key_dict, original_exception=e, request_data=data
|
||||
)
|
||||
traceback.print_exc()
|
||||
if isinstance(e, HTTPException):
|
||||
raise ProxyException(
|
||||
message=getattr(e, "message", str(e.detail)),
|
||||
type=getattr(e, "type", "None"),
|
||||
param=getattr(e, "param", "None"),
|
||||
code=getattr(e, "status_code", status.HTTP_400_BAD_REQUEST),
|
||||
)
|
||||
else:
|
||||
error_traceback = traceback.format_exc()
|
||||
error_msg = f"{str(e)}"
|
||||
raise ProxyException(
|
||||
message=getattr(e, "message", error_msg),
|
||||
type=getattr(e, "type", "None"),
|
||||
param=getattr(e, "param", "None"),
|
||||
code=getattr(e, "status_code", 500),
|
||||
)
|
||||
|
||||
|
||||
@router.get(
|
||||
"/v1/threads/{thread_id}/messages",
|
||||
dependencies=[Depends(user_api_key_auth)],
|
||||
tags=["assistants"],
|
||||
)
|
||||
@router.get(
|
||||
"/threads/{thread_id}/messages",
|
||||
dependencies=[Depends(user_api_key_auth)],
|
||||
tags=["assistants"],
|
||||
)
|
||||
async def get_messages(
|
||||
request: Request,
|
||||
thread_id: str,
|
||||
fastapi_response: Response,
|
||||
user_api_key_dict: UserAPIKeyAuth = Depends(user_api_key_auth),
|
||||
):
|
||||
global proxy_logging_obj
|
||||
data: Dict = {}
|
||||
try:
|
||||
# Include original request and headers in the data
|
||||
data["proxy_server_request"] = { # type: ignore
|
||||
"url": str(request.url),
|
||||
"method": request.method,
|
||||
"headers": dict(request.headers),
|
||||
"body": copy.copy(data), # use copy instead of deepcopy
|
||||
}
|
||||
|
||||
if data.get("user", None) is None and user_api_key_dict.user_id is not None:
|
||||
data["user"] = user_api_key_dict.user_id
|
||||
|
||||
if "metadata" not in data:
|
||||
data["metadata"] = {}
|
||||
data["metadata"]["user_api_key"] = user_api_key_dict.api_key
|
||||
data["metadata"]["user_api_key_metadata"] = user_api_key_dict.metadata
|
||||
_headers = dict(request.headers)
|
||||
_headers.pop(
|
||||
"authorization", None
|
||||
) # do not store the original `sk-..` api key in the db
|
||||
data["metadata"]["headers"] = _headers
|
||||
data["metadata"]["user_api_key_alias"] = getattr(
|
||||
user_api_key_dict, "key_alias", None
|
||||
)
|
||||
data["metadata"]["user_api_key_user_id"] = user_api_key_dict.user_id
|
||||
data["metadata"]["user_api_key_team_id"] = getattr(
|
||||
user_api_key_dict, "team_id", None
|
||||
)
|
||||
data["metadata"]["global_max_parallel_requests"] = general_settings.get(
|
||||
"global_max_parallel_requests", None
|
||||
)
|
||||
data["metadata"]["user_api_key_team_alias"] = getattr(
|
||||
user_api_key_dict, "team_alias", None
|
||||
)
|
||||
data["metadata"]["endpoint"] = str(request.url)
|
||||
|
||||
### TEAM-SPECIFIC PARAMS ###
|
||||
if user_api_key_dict.team_id is not None:
|
||||
team_config = await proxy_config.load_team_config(
|
||||
team_id=user_api_key_dict.team_id
|
||||
)
|
||||
if len(team_config) == 0:
|
||||
pass
|
||||
else:
|
||||
team_id = team_config.pop("team_id", None)
|
||||
data["metadata"]["team_id"] = team_id
|
||||
data = {
|
||||
**team_config,
|
||||
**data,
|
||||
} # add the team-specific configs to the completion call
|
||||
|
||||
# for now use custom_llm_provider=="openai" -> this will change as LiteLLM adds more providers for acreate_batch
|
||||
if llm_router is None:
|
||||
raise HTTPException(
|
||||
status_code=500, detail={"error": CommonProxyErrors.no_llm_router.value}
|
||||
)
|
||||
response = await llm_router.aget_messages(
|
||||
custom_llm_provider="openai", thread_id=thread_id, client=None, **data
|
||||
)
|
||||
|
||||
### ALERTING ###
|
||||
data["litellm_status"] = "success" # used for alerting
|
||||
|
||||
### RESPONSE HEADERS ###
|
||||
hidden_params = getattr(response, "_hidden_params", {}) or {}
|
||||
model_id = hidden_params.get("model_id", None) or ""
|
||||
cache_key = hidden_params.get("cache_key", None) or ""
|
||||
api_base = hidden_params.get("api_base", None) or ""
|
||||
|
||||
fastapi_response.headers.update(
|
||||
get_custom_headers(
|
||||
user_api_key_dict=user_api_key_dict,
|
||||
model_id=model_id,
|
||||
cache_key=cache_key,
|
||||
api_base=api_base,
|
||||
version=version,
|
||||
model_region=getattr(user_api_key_dict, "allowed_model_region", ""),
|
||||
)
|
||||
)
|
||||
|
||||
return response
|
||||
except Exception as e:
|
||||
data["litellm_status"] = "fail" # used for alerting
|
||||
await proxy_logging_obj.post_call_failure_hook(
|
||||
user_api_key_dict=user_api_key_dict, original_exception=e, request_data=data
|
||||
)
|
||||
traceback.print_exc()
|
||||
if isinstance(e, HTTPException):
|
||||
raise ProxyException(
|
||||
message=getattr(e, "message", str(e.detail)),
|
||||
type=getattr(e, "type", "None"),
|
||||
param=getattr(e, "param", "None"),
|
||||
code=getattr(e, "status_code", status.HTTP_400_BAD_REQUEST),
|
||||
)
|
||||
else:
|
||||
error_traceback = traceback.format_exc()
|
||||
error_msg = f"{str(e)}"
|
||||
raise ProxyException(
|
||||
message=getattr(e, "message", error_msg),
|
||||
type=getattr(e, "type", "None"),
|
||||
param=getattr(e, "param", "None"),
|
||||
code=getattr(e, "status_code", 500),
|
||||
)
|
||||
|
||||
|
||||
@router.get(
|
||||
"/v1/threads/{thread_id}/runs",
|
||||
dependencies=[Depends(user_api_key_auth)],
|
||||
tags=["assistants"],
|
||||
)
|
||||
@router.get(
|
||||
"/threads/{thread_id}/runs",
|
||||
dependencies=[Depends(user_api_key_auth)],
|
||||
tags=["assistants"],
|
||||
)
|
||||
async def run_thread(
|
||||
request: Request,
|
||||
thread_id: str,
|
||||
fastapi_response: Response,
|
||||
user_api_key_dict: UserAPIKeyAuth = Depends(user_api_key_auth),
|
||||
):
|
||||
global proxy_logging_obj
|
||||
data: Dict = {}
|
||||
try:
|
||||
body = await request.body()
|
||||
data = orjson.loads(body)
|
||||
# Include original request and headers in the data
|
||||
data["proxy_server_request"] = { # type: ignore
|
||||
"url": str(request.url),
|
||||
"method": request.method,
|
||||
"headers": dict(request.headers),
|
||||
"body": copy.copy(data), # use copy instead of deepcopy
|
||||
}
|
||||
|
||||
if data.get("user", None) is None and user_api_key_dict.user_id is not None:
|
||||
data["user"] = user_api_key_dict.user_id
|
||||
|
||||
if "litellm_metadata" not in data:
|
||||
data["litellm_metadata"] = {}
|
||||
data["litellm_metadata"]["user_api_key"] = user_api_key_dict.api_key
|
||||
data["litellm_metadata"]["user_api_key_metadata"] = user_api_key_dict.metadata
|
||||
_headers = dict(request.headers)
|
||||
_headers.pop(
|
||||
"authorization", None
|
||||
) # do not store the original `sk-..` api key in the db
|
||||
data["litellm_metadata"]["headers"] = _headers
|
||||
data["litellm_metadata"]["user_api_key_alias"] = getattr(
|
||||
user_api_key_dict, "key_alias", None
|
||||
)
|
||||
data["litellm_metadata"]["user_api_key_user_id"] = user_api_key_dict.user_id
|
||||
data["litellm_metadata"]["user_api_key_team_id"] = getattr(
|
||||
user_api_key_dict, "team_id", None
|
||||
)
|
||||
data["litellm_metadata"]["global_max_parallel_requests"] = general_settings.get(
|
||||
"global_max_parallel_requests", None
|
||||
)
|
||||
data["litellm_metadata"]["user_api_key_team_alias"] = getattr(
|
||||
user_api_key_dict, "team_alias", None
|
||||
)
|
||||
data["litellm_metadata"]["endpoint"] = str(request.url)
|
||||
|
||||
### TEAM-SPECIFIC PARAMS ###
|
||||
if user_api_key_dict.team_id is not None:
|
||||
team_config = await proxy_config.load_team_config(
|
||||
team_id=user_api_key_dict.team_id
|
||||
)
|
||||
if len(team_config) == 0:
|
||||
pass
|
||||
else:
|
||||
team_id = team_config.pop("team_id", None)
|
||||
data["litellm_metadata"]["team_id"] = team_id
|
||||
data = {
|
||||
**team_config,
|
||||
**data,
|
||||
} # add the team-specific configs to the completion call
|
||||
|
||||
# for now use custom_llm_provider=="openai" -> this will change as LiteLLM adds more providers for acreate_batch
|
||||
if llm_router is None:
|
||||
raise HTTPException(
|
||||
status_code=500, detail={"error": CommonProxyErrors.no_llm_router.value}
|
||||
)
|
||||
response = await llm_router.arun_thread(
|
||||
custom_llm_provider="openai", thread_id=thread_id, client=None, **data
|
||||
)
|
||||
|
||||
### ALERTING ###
|
||||
data["litellm_status"] = "success" # used for alerting
|
||||
|
||||
### RESPONSE HEADERS ###
|
||||
hidden_params = getattr(response, "_hidden_params", {}) or {}
|
||||
model_id = hidden_params.get("model_id", None) or ""
|
||||
cache_key = hidden_params.get("cache_key", None) or ""
|
||||
api_base = hidden_params.get("api_base", None) or ""
|
||||
|
||||
fastapi_response.headers.update(
|
||||
get_custom_headers(
|
||||
user_api_key_dict=user_api_key_dict,
|
||||
model_id=model_id,
|
||||
cache_key=cache_key,
|
||||
api_base=api_base,
|
||||
version=version,
|
||||
model_region=getattr(user_api_key_dict, "allowed_model_region", ""),
|
||||
)
|
||||
)
|
||||
|
||||
return response
|
||||
except Exception as e:
|
||||
data["litellm_status"] = "fail" # used for alerting
|
||||
await proxy_logging_obj.post_call_failure_hook(
|
||||
user_api_key_dict=user_api_key_dict, original_exception=e, request_data=data
|
||||
)
|
||||
traceback.print_exc()
|
||||
if isinstance(e, HTTPException):
|
||||
raise ProxyException(
|
||||
message=getattr(e, "message", str(e.detail)),
|
||||
type=getattr(e, "type", "None"),
|
||||
param=getattr(e, "param", "None"),
|
||||
code=getattr(e, "status_code", status.HTTP_400_BAD_REQUEST),
|
||||
)
|
||||
else:
|
||||
error_traceback = traceback.format_exc()
|
||||
error_msg = f"{str(e)}"
|
||||
raise ProxyException(
|
||||
message=getattr(e, "message", error_msg),
|
||||
type=getattr(e, "type", "None"),
|
||||
param=getattr(e, "param", "None"),
|
||||
code=getattr(e, "status_code", 500),
|
||||
)
|
||||
|
||||
|
||||
######################################################################
|
||||
|
||||
# /v1/batches Endpoints
|
||||
|
|
|
|||
Loading…
Add table
Add a link
Reference in a new issue