mirror of
https://github.com/BerriAI/litellm.git
synced 2026-09-13 23:11:40 +00:00
chore(typing): clear basedpyright Any errors in proxy endpoints and http handlers
Replace Any-typed seams with concrete types across the proxy's highest Any-count production files: typed FastAPI/aiohttp/DB helper signatures in proxy_server.py (TypedDict for aiohttp connector kwargs, concrete route and pydantic model types instead of Dict[str, Any]/kwargs-spread constructors), llm_http_handler.py (logging_obj, encoding, provider config parameters), proxy/utils.py and the key/team management endpoints (CustomLogger/CustomGuardrail-typed callbacks, concrete Prisma repository return types in place of raw .table access), and streaming_handler.py (model_validate over dict-kwargs spreads). Whole-tree basedpyright, same environment: reportExplicitAny 7283 -> 7228 (-55), reportAny 22710 -> 22217 (-493), with knock-on drops across reportUnknownArgumentType, reportUnknownMemberType, reportPrivateUsage, reportArgumentType, and reportUnnecessaryComparison. Zero regressions in any rule. All mapped tests for the six touched files pass. Also satisfies the ruff-strict and type-discipline gates (Optional[X] modernized to X | None, read-only DB-query results and route lists typed Sequence/Mapping instead of List/Dict, and the handful of genuinely in-place-mutated values marked # mutable-ok with a reason), and ratchets down the ruff-strict, type-discipline, and basedpyright lint budgets via make lint-budget-update to match.
This commit is contained in:
mateo-berri
parent
71b825a7f0
commit
68fb66c84e
9 changed files with 222 additions and 144 deletions
|
|
@ -1,15 +1,15 @@
|
|||
{
|
||||
"reportAny": {
|
||||
"limit": 33171
|
||||
"limit": 32185
|
||||
},
|
||||
"reportArgumentType": {
|
||||
"limit": 2645
|
||||
"limit": 2635
|
||||
},
|
||||
"reportAssignmentType": {
|
||||
"limit": 329
|
||||
},
|
||||
"reportAttributeAccessIssue": {
|
||||
"limit": 516
|
||||
"limit": 514
|
||||
},
|
||||
"reportCallIssue": {
|
||||
"limit": 123
|
||||
|
|
@ -24,7 +24,7 @@
|
|||
"limit": 42
|
||||
},
|
||||
"reportExplicitAny": {
|
||||
"limit": 10228
|
||||
"limit": 10118
|
||||
},
|
||||
"reportFunctionMemberAccess": {
|
||||
"limit": 11
|
||||
|
|
@ -54,10 +54,10 @@
|
|||
"limit": 0
|
||||
},
|
||||
"reportMissingParameterType": {
|
||||
"limit": 5869
|
||||
"limit": 5867
|
||||
},
|
||||
"reportMissingTypeArgument": {
|
||||
"limit": 15864
|
||||
"limit": 15863
|
||||
},
|
||||
"reportMissingTypeStubs": {
|
||||
"limit": 41
|
||||
|
|
@ -84,7 +84,7 @@
|
|||
"limit": 77
|
||||
},
|
||||
"reportPrivateUsage": {
|
||||
"limit": 2437
|
||||
"limit": 2415
|
||||
},
|
||||
"reportRedeclaration": {
|
||||
"limit": 12
|
||||
|
|
@ -99,25 +99,25 @@
|
|||
"limit": 0
|
||||
},
|
||||
"reportUnknownArgumentType": {
|
||||
"limit": 45522
|
||||
"limit": 45254
|
||||
},
|
||||
"reportUnknownLambdaType": {
|
||||
"limit": 113
|
||||
},
|
||||
"reportUnknownMemberType": {
|
||||
"limit": 40479
|
||||
"limit": 40440
|
||||
},
|
||||
"reportUnknownParameterType": {
|
||||
"limit": 20341
|
||||
"limit": 20338
|
||||
},
|
||||
"reportUnknownVariableType": {
|
||||
"limit": 32052
|
||||
"limit": 32048
|
||||
},
|
||||
"reportUnnecessaryCast": {
|
||||
"limit": 177
|
||||
},
|
||||
"reportUnnecessaryComparison": {
|
||||
"limit": 1023
|
||||
"limit": 1015
|
||||
},
|
||||
"reportUnnecessaryContains": {
|
||||
"limit": 7
|
||||
|
|
|
|||
|
|
@ -30,7 +30,6 @@ from litellm._uuid import uuid
|
|||
from litellm.litellm_core_utils.model_response_utils import (
|
||||
is_model_response_stream_empty,
|
||||
)
|
||||
from litellm.litellm_core_utils.redact_messages import LiteLLMLoggingObject
|
||||
from litellm.litellm_core_utils.thread_pool_executor import executor
|
||||
from litellm.types.llms.openai import OpenAIChatCompletionChunk
|
||||
from litellm.types.router import GenericLiteLLMParams
|
||||
|
|
@ -125,14 +124,17 @@ class CustomStreamWrapper:
|
|||
self.model = model
|
||||
self.make_call = make_call
|
||||
self.custom_llm_provider = custom_llm_provider
|
||||
self.logging_obj: LiteLLMLoggingObject = logging_obj
|
||||
self.logging_obj: Any = logging_obj
|
||||
self.completion_stream = completion_stream
|
||||
self.sent_first_chunk = False
|
||||
self.sent_last_chunk = False
|
||||
self._stream_created_time: float = time.time()
|
||||
|
||||
litellm_params: GenericLiteLLMParams = GenericLiteLLMParams(
|
||||
**self.logging_obj.model_call_details.get("litellm_params", {})
|
||||
_litellm_params_kwargs = self.logging_obj.model_call_details.get("litellm_params", {})
|
||||
litellm_params: GenericLiteLLMParams = (
|
||||
GenericLiteLLMParams.model_validate(_litellm_params_kwargs)
|
||||
if isinstance(_litellm_params_kwargs, dict)
|
||||
else GenericLiteLLMParams()
|
||||
)
|
||||
self.merge_reasoning_content_in_choices: bool = litellm_params.merge_reasoning_content_in_choices or False
|
||||
self.sent_first_thinking_block = False
|
||||
|
|
@ -677,7 +679,7 @@ class CustomStreamWrapper:
|
|||
if chunk:
|
||||
args.update({k: v for k, v in chunk.items() if k != "stream"})
|
||||
|
||||
model_response = ModelResponseStream(**args)
|
||||
model_response = ModelResponseStream.model_validate(args)
|
||||
if self.response_id is not None:
|
||||
model_response.id = self.response_id
|
||||
if self.system_fingerprint is not None:
|
||||
|
|
@ -817,7 +819,7 @@ class CustomStreamWrapper:
|
|||
_initial_delta = model_response.choices[0].delta.model_dump()
|
||||
|
||||
_initial_delta.pop("role", None)
|
||||
model_response.choices[0].delta = Delta(**_initial_delta)
|
||||
model_response.choices[0].delta = Delta.model_validate(_initial_delta)
|
||||
return model_response
|
||||
|
||||
def _has_special_delta_content(self, model_response: ModelResponseStream) -> bool:
|
||||
|
|
@ -915,7 +917,7 @@ class CustomStreamWrapper:
|
|||
choice_json.pop(
|
||||
"finish_reason", None
|
||||
) # for mistral etc. which return a value in their last chunk (not-openai compatible).
|
||||
choices.append(StreamingChoices(**choice_json))
|
||||
choices.append(StreamingChoices.model_validate(choice_json))
|
||||
except Exception:
|
||||
choices.append(StreamingChoices())
|
||||
setattr(model_response, "choices", choices)
|
||||
|
|
@ -946,7 +948,7 @@ class CustomStreamWrapper:
|
|||
self.sent_first_chunk = True
|
||||
if response_obj.get("provider_specific_fields") is not None:
|
||||
completion_obj["provider_specific_fields"] = response_obj["provider_specific_fields"]
|
||||
model_response.choices[0].delta = Delta(**completion_obj)
|
||||
model_response.choices[0].delta = Delta.model_validate(completion_obj)
|
||||
_index: Optional[int] = completion_obj.get("index")
|
||||
if _index is not None:
|
||||
model_response.choices[0].index = _index
|
||||
|
|
|
|||
|
|
@ -160,8 +160,10 @@ def _rust_responses_websocket_enabled(
|
|||
from .http_handler import get_shared_realtime_ssl_context
|
||||
|
||||
if TYPE_CHECKING:
|
||||
import tiktoken
|
||||
from aiohttp import ClientSession
|
||||
|
||||
from litellm.integrations.custom_logger import CustomLogger
|
||||
from litellm.litellm_core_utils.litellm_logging import Logging as _LiteLLMLoggingObj
|
||||
from litellm.llms.anthropic.experimental_pass_through.messages.streaming_iterator import (
|
||||
AnthropicMessagesStreamingResponse,
|
||||
|
|
@ -210,7 +212,7 @@ def _responses_api_optional_request_param_names() -> frozenset[str]:
|
|||
return frozenset(get_type_hints(ResponsesAPIOptionalRequestParams).keys())
|
||||
|
||||
|
||||
def _custom_logger_callbacks(logging_obj: Any) -> list[Any]:
|
||||
def _custom_logger_callbacks(logging_obj: LiteLLMLoggingObj) -> tuple["CustomLogger", ...]:
|
||||
from litellm.integrations.custom_logger import CustomLogger
|
||||
from litellm.litellm_core_utils.litellm_logging import (
|
||||
get_custom_logger_compatible_class,
|
||||
|
|
@ -221,7 +223,7 @@ def _custom_logger_callbacks(logging_obj: Any) -> list[Any]:
|
|||
if isinstance(dynamic_success_callbacks, (list, tuple)):
|
||||
callbacks.extend(dynamic_success_callbacks)
|
||||
|
||||
custom_loggers: list[Any] = []
|
||||
custom_loggers = []
|
||||
for cb in callbacks:
|
||||
if isinstance(cb, str):
|
||||
resolved = get_custom_logger_compatible_class(cb) # type: ignore[arg-type]
|
||||
|
|
@ -230,10 +232,10 @@ def _custom_logger_callbacks(logging_obj: Any) -> list[Any]:
|
|||
cb = resolved
|
||||
if isinstance(cb, CustomLogger):
|
||||
custom_loggers.append(cb)
|
||||
return custom_loggers
|
||||
return tuple(custom_loggers)
|
||||
|
||||
|
||||
def _has_pre_call_deployment_hook(logging_obj: Any) -> bool:
|
||||
def _has_pre_call_deployment_hook(logging_obj: LiteLLMLoggingObj) -> bool:
|
||||
from litellm.integrations.custom_logger import CustomLogger
|
||||
|
||||
base_func = CustomLogger.async_pre_call_deployment_hook
|
||||
|
|
@ -359,7 +361,7 @@ class BaseLLMHTTPHandler:
|
|||
messages: list,
|
||||
optional_params: dict,
|
||||
litellm_params: dict,
|
||||
encoding: Any,
|
||||
encoding: Optional["tiktoken.Encoding"],
|
||||
api_key: Optional[str] = None,
|
||||
client: Optional[AsyncHTTPHandler] = None,
|
||||
json_mode: bool = False,
|
||||
|
|
@ -425,7 +427,7 @@ class BaseLLMHTTPHandler:
|
|||
api_base: Optional[str],
|
||||
custom_llm_provider: str,
|
||||
model_response: ModelResponse,
|
||||
encoding,
|
||||
encoding: Optional["tiktoken.Encoding"],
|
||||
logging_obj: LiteLLMLoggingObj,
|
||||
optional_params: dict,
|
||||
timeout: Union[float, httpx.Timeout],
|
||||
|
|
@ -5008,7 +5010,7 @@ class BaseLLMHTTPHandler:
|
|||
return depth, max(max_loops, 1), fingerprints
|
||||
|
||||
@staticmethod
|
||||
def _has_agentic_completion_hook(logging_obj: Any) -> bool:
|
||||
def _has_agentic_completion_hook(logging_obj: LiteLLMLoggingObj) -> bool:
|
||||
"""
|
||||
True if any registered callback actually overrides
|
||||
``async_should_run_agentic_loop`` (the gate every agentic hook goes
|
||||
|
|
@ -5248,7 +5250,7 @@ class BaseLLMHTTPHandler:
|
|||
self,
|
||||
result: Any,
|
||||
model: str,
|
||||
responses_api_provider_config: Any,
|
||||
responses_api_provider_config: BaseResponsesAPIConfig,
|
||||
logging_obj: "LiteLLMLoggingObj",
|
||||
custom_llm_provider: str,
|
||||
) -> Any:
|
||||
|
|
@ -5769,7 +5771,7 @@ class BaseLLMHTTPHandler:
|
|||
websockets_module: Any,
|
||||
url: str,
|
||||
headers: dict,
|
||||
ssl_context: Any,
|
||||
ssl_context: bool | str | ssl.SSLContext | None,
|
||||
*,
|
||||
open_timeout: float = 8.0,
|
||||
max_attempts: int = 3,
|
||||
|
|
|
|||
|
|
@ -490,7 +490,7 @@ _NON_ADMIN_SAFE_ALLOWED_ROUTES_PRESETS = frozenset({"llm_api_routes", "info_rout
|
|||
|
||||
def _validate_caller_can_change_key_ownership(
|
||||
data: Optional[BaseModel],
|
||||
existing_key_row: Any,
|
||||
existing_key_row: LiteLLM_VerificationToken,
|
||||
user_api_key_dict: UserAPIKeyAuth,
|
||||
) -> None:
|
||||
"""
|
||||
|
|
@ -523,7 +523,7 @@ def _validate_caller_can_change_key_ownership(
|
|||
status_code=403,
|
||||
detail="Non-admin users cannot remove the user_id from a key.",
|
||||
)
|
||||
existing_user_id = getattr(existing_key_row, "user_id", None)
|
||||
existing_user_id = existing_key_row.user_id
|
||||
if incoming_user_id != existing_user_id:
|
||||
raise HTTPException(
|
||||
status_code=403,
|
||||
|
|
@ -2112,7 +2112,7 @@ async def _process_single_key_update(
|
|||
litellm_changed_by: Optional[str],
|
||||
prisma_client: Optional[PrismaClient],
|
||||
user_api_key_cache: UserApiKeyCache,
|
||||
proxy_logging_obj: Any,
|
||||
proxy_logging_obj: ProxyLogging,
|
||||
llm_router: Optional[Router],
|
||||
user_custom_key_update: Optional[Callable] = None,
|
||||
existing_key_row: Optional[LiteLLM_VerificationToken] = None,
|
||||
|
|
@ -2265,9 +2265,9 @@ async def _process_single_key_update(
|
|||
async def _validate_mcp_servers_for_key_update(
|
||||
data: "UpdateKeyRequest",
|
||||
team_obj: Optional["LiteLLM_TeamTableCachedObj"],
|
||||
existing_key_row: Any,
|
||||
prisma_client: Any,
|
||||
user_api_key_cache: Any,
|
||||
existing_key_row: LiteLLM_VerificationToken,
|
||||
prisma_client: PrismaClient | None,
|
||||
user_api_key_cache: UserApiKeyCache,
|
||||
is_proxy_admin: bool,
|
||||
) -> Optional[ObjectPermissionDict]:
|
||||
"""Validate MCP servers in object_permission against the effective team."""
|
||||
|
|
@ -2302,14 +2302,20 @@ async def _validate_mcp_servers_for_key_update(
|
|||
|
||||
async def _validate_update_key_data(
|
||||
data: UpdateKeyRequest,
|
||||
existing_key_row: Any,
|
||||
existing_key_row: LiteLLM_VerificationToken,
|
||||
user_api_key_dict: UserAPIKeyAuth,
|
||||
llm_router: Any,
|
||||
llm_router: Router | None,
|
||||
premium_user: bool,
|
||||
prisma_client: Any,
|
||||
user_api_key_cache: Any,
|
||||
prisma_client: PrismaClient | None,
|
||||
user_api_key_cache: UserApiKeyCache,
|
||||
) -> None:
|
||||
"""Validate permissions and constraints for key update."""
|
||||
if prisma_client is None:
|
||||
raise HTTPException(
|
||||
status_code=500,
|
||||
detail={"error": "Database not connected"},
|
||||
)
|
||||
|
||||
# Reject NaN/±inf spend before it can reach the DB / spend counter.
|
||||
validate_finite_spend(data.spend)
|
||||
|
||||
|
|
@ -2414,11 +2420,12 @@ async def _validate_update_key_data(
|
|||
# _check_key_admin_access that would otherwise require team/org admin status.
|
||||
_key_is_team_key = getattr(existing_key_row, "team_id", None) is not None
|
||||
can_skip_admin_check = (caller_is_creator or _key_is_team_key) and not _is_budget_change
|
||||
if (not _is_proxy_admin) and prisma_client is not None and not can_skip_admin_check:
|
||||
hashed_key = existing_key_row.token
|
||||
if (not _is_proxy_admin) and not can_skip_admin_check:
|
||||
if existing_key_row.token is None:
|
||||
raise HTTPException(status_code=500, detail={"error": "Existing key is missing its token"})
|
||||
await _check_key_admin_access(
|
||||
user_api_key_dict=user_api_key_dict,
|
||||
hashed_token=hashed_key,
|
||||
hashed_token=existing_key_row.token,
|
||||
prisma_client=prisma_client,
|
||||
user_api_key_cache=user_api_key_cache,
|
||||
route=("/key/update (max_budget/spend)" if _is_budget_change else "/key/update"),
|
||||
|
|
@ -5421,7 +5428,7 @@ async def list_keys(
|
|||
|
||||
async def _apply_non_admin_alias_scope(
|
||||
user_api_key_dict: UserAPIKeyAuth,
|
||||
prisma_client: Any,
|
||||
prisma_client: PrismaClient | None,
|
||||
query_params: List[Any],
|
||||
where_parts: List[str],
|
||||
) -> None:
|
||||
|
|
@ -5852,9 +5859,7 @@ async def _list_key_helper(
|
|||
where=where # type: ignore
|
||||
)
|
||||
else:
|
||||
total_count = await VerificationTokenRepository(prisma_client).table.count(
|
||||
where=where # type: ignore
|
||||
)
|
||||
total_count = await VerificationTokenRepository(prisma_client).count(where=where)
|
||||
|
||||
verbose_proxy_logger.debug(f"Total count of keys: {total_count}")
|
||||
|
||||
|
|
@ -5932,7 +5937,7 @@ def _get_condition_to_filter_out_ui_session_tokens() -> Dict[str, Any]:
|
|||
async def _check_key_admin_access(
|
||||
user_api_key_dict: UserAPIKeyAuth,
|
||||
hashed_token: str,
|
||||
prisma_client: Any,
|
||||
prisma_client: PrismaClient | None,
|
||||
user_api_key_cache: UserApiKeyCache,
|
||||
route: str,
|
||||
) -> None:
|
||||
|
|
@ -6443,7 +6448,7 @@ def _validate_key_alias_format(key_alias: Optional[str]) -> None:
|
|||
|
||||
async def _enforce_unique_key_alias(
|
||||
key_alias: Optional[str],
|
||||
prisma_client: Any,
|
||||
prisma_client: PrismaClient | None,
|
||||
existing_key_token: Optional[str] = None,
|
||||
) -> None:
|
||||
"""
|
||||
|
|
@ -6451,7 +6456,7 @@ async def _enforce_unique_key_alias(
|
|||
|
||||
Args:
|
||||
key_alias (Optional[str]): The key alias to check
|
||||
prisma_client (Any): Prisma client instance
|
||||
prisma_client (Optional[PrismaClient]): Prisma client instance
|
||||
existing_key_token (Optional[str]): ID of existing key being updated, to exclude from uniqueness check
|
||||
(The Admin UI passes key_alias, in all Edit key requests. So we need to be sure that if we find a key with the same alias, it's not the same key we're updating)
|
||||
|
||||
|
|
|
|||
|
|
@ -14,7 +14,7 @@ import json
|
|||
import math
|
||||
import traceback
|
||||
from datetime import datetime, timezone
|
||||
from typing import Annotated, Any, Dict, List, Mapping, Optional, Tuple, Union, cast
|
||||
from typing import Annotated, Any, Dict, List, Mapping, Optional, Sequence, Tuple, Union, cast
|
||||
|
||||
import fastapi
|
||||
from fastapi import APIRouter, Depends, Header, HTTPException, Request, status
|
||||
|
|
@ -31,6 +31,7 @@ from litellm.proxy._types import (
|
|||
CommonProxyErrors,
|
||||
DeleteTeamRequest,
|
||||
LiteLLM_AuditLogs,
|
||||
LiteLLM_BudgetTable,
|
||||
LiteLLM_DeletedTeamTable,
|
||||
LiteLLM_ManagementEndpoint_MetadataFields,
|
||||
LiteLLM_ManagementEndpoint_MetadataFields_Premium,
|
||||
|
|
@ -78,6 +79,7 @@ from litellm.proxy.auth.auth_checks import (
|
|||
from litellm.proxy.auth.user_api_key_auth import user_api_key_auth
|
||||
from litellm.proxy.common_utils.callback_utils import encrypt_callback_vars
|
||||
from litellm.proxy.common_utils.json_merge_patch import apply_json_merge_patch
|
||||
from litellm.proxy.common_utils.user_api_key_cache import UserApiKeyCache
|
||||
from litellm.proxy.management_endpoints.common_utils import (
|
||||
_check_passthrough_routes_caller_permission,
|
||||
_is_user_org_admin_for_team,
|
||||
|
|
@ -106,7 +108,7 @@ from litellm.proxy.management_helpers.utils import (
|
|||
add_new_member,
|
||||
management_endpoint_wrapper,
|
||||
)
|
||||
from litellm.proxy.utils import PrismaClient, handle_exception_on_proxy
|
||||
from litellm.proxy.utils import PrismaClient, ProxyLogging, handle_exception_on_proxy
|
||||
from litellm.repositories.budget_repository import BudgetRepository
|
||||
from litellm.repositories.organization_repository import OrganizationRepository
|
||||
from litellm.repositories.table_repositories import (
|
||||
|
|
@ -151,9 +153,9 @@ def _sanitize_for_log(value: object) -> str:
|
|||
|
||||
|
||||
async def _refresh_cached_team(
|
||||
team_row: Any,
|
||||
user_api_key_cache: Any,
|
||||
proxy_logging_obj: Any,
|
||||
team_row: LiteLLM_TeamTable,
|
||||
user_api_key_cache: UserApiKeyCache,
|
||||
proxy_logging_obj: ProxyLogging,
|
||||
) -> None:
|
||||
"""
|
||||
Refresh the in-memory cached team object after a DB write.
|
||||
|
|
@ -274,7 +276,7 @@ class TeamMemberBudgetHandler:
|
|||
if team_member_budget_duration is not None:
|
||||
budget_request.budget_duration = team_member_budget_duration
|
||||
|
||||
team_member_budget_table = await new_budget(
|
||||
team_member_budget_table: LiteLLM_BudgetTable = await new_budget(
|
||||
budget_obj=budget_request,
|
||||
user_api_key_dict=user_api_key_dict,
|
||||
)
|
||||
|
|
@ -322,7 +324,7 @@ class TeamMemberBudgetHandler:
|
|||
if team_member_budget_duration is not None:
|
||||
budget_request.budget_duration = team_member_budget_duration
|
||||
|
||||
budget_row = await update_budget(
|
||||
budget_row: LiteLLM_BudgetTable = await update_budget(
|
||||
budget_obj=budget_request,
|
||||
user_api_key_dict=user_api_key_dict,
|
||||
)
|
||||
|
|
@ -395,7 +397,7 @@ class TeamMemberBudgetHandler:
|
|||
@staticmethod
|
||||
async def backfill_team_member_budget_entries(
|
||||
team_id: str,
|
||||
members_with_roles: List[Union[Member, dict]],
|
||||
members_with_roles: Sequence[Union[Member, dict]],
|
||||
team_member_budget_id: str,
|
||||
prisma_client: PrismaClient,
|
||||
) -> None:
|
||||
|
|
@ -414,7 +416,9 @@ class TeamMemberBudgetHandler:
|
|||
return
|
||||
|
||||
# Batch-fetch existing memberships for this team (avoids N+1 queries)
|
||||
existing_memberships = await TeamMembershipRepository(prisma_client).table.find_many(where={"team_id": team_id})
|
||||
existing_memberships: Sequence[LiteLLM_TeamMembership] = await TeamMembershipRepository(
|
||||
prisma_client
|
||||
).table.find_many(where={"team_id": team_id})
|
||||
existing_user_ids = {m.user_id for m in existing_memberships}
|
||||
|
||||
# Identify members with no existing membership row.
|
||||
|
|
@ -447,7 +451,7 @@ class TeamMemberBudgetHandler:
|
|||
# Heal existing membership rows that predate the team_member_budget
|
||||
# configuration: populate budget_id where it is currently NULL.
|
||||
# Rows with an explicit budget_id (per-member override) are left alone.
|
||||
updated = await TeamMembershipRepository(prisma_client).table.update_many(
|
||||
updated: int = await TeamMembershipRepository(prisma_client).table.update_many(
|
||||
where={"team_id": team_id, "budget_id": None},
|
||||
data={"budget_id": team_member_budget_id},
|
||||
)
|
||||
|
|
@ -503,7 +507,7 @@ async def get_all_team_memberships(
|
|||
# else:
|
||||
# where_obj = {"user_id": str(user_id), "team_id": {"in": team_id}}
|
||||
|
||||
team_memberships = await TeamMembershipRepository(prisma_client).table.find_many(
|
||||
team_memberships: Sequence[LiteLLM_TeamMembership] = await TeamMembershipRepository(prisma_client).table.find_many(
|
||||
where=where_obj,
|
||||
include={"litellm_budget_table": True},
|
||||
)
|
||||
|
|
@ -765,7 +769,7 @@ async def _check_org_team_limits(
|
|||
# calculate allocated tpm/rpm limit
|
||||
# check if specified tpm/rpm limit is greater than allocated tpm/rpm limit
|
||||
|
||||
teams = await TeamRepository(prisma_client).table.find_many(
|
||||
teams: Sequence[LiteLLM_TeamTable] = await TeamRepository(prisma_client).table.find_many(
|
||||
where={"organization_id": org_table.organization_id},
|
||||
)
|
||||
|
||||
|
|
@ -790,7 +794,7 @@ async def _check_user_team_limits(
|
|||
data: Union[NewTeamRequest, UpdateTeamRequest],
|
||||
user_api_key_dict: UserAPIKeyAuth,
|
||||
prisma_client: PrismaClient,
|
||||
user_api_key_cache: Any,
|
||||
user_api_key_cache: UserApiKeyCache,
|
||||
) -> None:
|
||||
"""
|
||||
Enforce the caller's personal limits when CREATING a standalone team.
|
||||
|
|
@ -1394,7 +1398,7 @@ async def _update_model_table(
|
|||
async def _auto_add_team_members_to_organization(
|
||||
team: LiteLLM_TeamTable,
|
||||
organization: LiteLLM_OrganizationTableWithMembers,
|
||||
prisma_client: Any,
|
||||
prisma_client: PrismaClient,
|
||||
) -> None:
|
||||
"""
|
||||
When moving a team to an org, ensure all team members are also org members.
|
||||
|
|
@ -1432,11 +1436,11 @@ async def _auto_add_team_members_to_organization(
|
|||
|
||||
async def fetch_and_validate_organization(
|
||||
organization_id: str,
|
||||
existing_team_row: Any,
|
||||
existing_team_row: LiteLLM_TeamTable,
|
||||
llm_router: Optional[Router],
|
||||
prisma_client: Any,
|
||||
prisma_client: PrismaClient,
|
||||
user_api_key_dict: Optional[UserAPIKeyAuth] = None,
|
||||
) -> Any:
|
||||
) -> LiteLLM_OrganizationTable | None:
|
||||
"""
|
||||
Fetch and validate an organization for team update operations.
|
||||
|
||||
|
|
@ -1455,7 +1459,7 @@ async def fetch_and_validate_organization(
|
|||
if llm_router is None:
|
||||
raise HTTPException(status_code=500, detail={"error": CommonProxyErrors.no_llm_router.value})
|
||||
|
||||
organization_row = await OrganizationRepository(prisma_client).table.find_unique(
|
||||
organization_row: LiteLLM_OrganizationTable | None = await OrganizationRepository(prisma_client).table.find_unique(
|
||||
where={"organization_id": organization_id},
|
||||
include={"litellm_budget_table": True, "members": True, "teams": True},
|
||||
)
|
||||
|
|
@ -1704,7 +1708,9 @@ async def update_team(
|
|||
detail={"error": f"soft_budget must be a non-negative finite number. Received: {data.soft_budget}"},
|
||||
)
|
||||
|
||||
existing_team_row = await TeamRepository(prisma_client).table.find_unique(where={"team_id": data.team_id})
|
||||
existing_team_row: LiteLLM_TeamTable | None = await TeamRepository(prisma_client).table.find_unique(
|
||||
where={"team_id": data.team_id}
|
||||
)
|
||||
|
||||
if existing_team_row is None:
|
||||
raise HTTPException(
|
||||
|
|
@ -1792,7 +1798,9 @@ async def update_team(
|
|||
|
||||
# check org team limits - if updating team that belongs to an org
|
||||
org_id_to_check = (
|
||||
data.organization_id if data.organization_id is not None else existing_team_row.organization_id
|
||||
data.organization_id
|
||||
if data.organization_id is not None
|
||||
else getattr(existing_team_row, "organization_id", None)
|
||||
)
|
||||
if org_id_to_check is not None and isinstance(org_id_to_check, str) and prisma_client is not None:
|
||||
org_table = await get_org_object(
|
||||
|
|
@ -1895,7 +1903,7 @@ async def update_team(
|
|||
updated_kv.pop("model_aliases")
|
||||
_model_id = await _update_model_table(
|
||||
data=data,
|
||||
model_id=existing_team_row.model_id,
|
||||
model_id=getattr(existing_team_row, "model_id", None),
|
||||
prisma_client=prisma_client,
|
||||
user_api_key_dict=user_api_key_dict,
|
||||
litellm_proxy_admin_name=litellm_proxy_admin_name,
|
||||
|
|
@ -2004,13 +2012,16 @@ async def patch_team(
|
|||
patch_fields = data.model_dump(exclude_unset=True, exclude={"team_id"})
|
||||
|
||||
if "metadata" in patch_fields:
|
||||
existing_team_row = await TeamRepository(prisma_client).table.find_unique(where={"team_id": team_id})
|
||||
existing_team_row: LiteLLM_TeamTable | None = await TeamRepository(prisma_client).table.find_unique(
|
||||
where={"team_id": team_id}
|
||||
)
|
||||
if existing_team_row is None:
|
||||
raise HTTPException(
|
||||
status_code=404,
|
||||
detail={"error": f"Team not found, passed team_id={team_id}"},
|
||||
)
|
||||
existing_metadata = existing_team_row.metadata if isinstance(existing_team_row.metadata, dict) else {}
|
||||
_existing_metadata = getattr(existing_team_row, "metadata", None)
|
||||
existing_metadata = _existing_metadata if isinstance(_existing_metadata, dict) else {}
|
||||
patch_fields["metadata"] = apply_json_merge_patch(existing_metadata, patch_fields["metadata"])
|
||||
|
||||
update_request = UpdateTeamRequest.model_validate({"team_id": team_id, **patch_fields})
|
||||
|
|
@ -2706,7 +2717,9 @@ async def team_member_delete(
|
|||
detail={"error": "Either user_id or user_email needs to be passed in"},
|
||||
)
|
||||
|
||||
_existing_team_row = await TeamRepository(prisma_client).table.find_unique(where={"team_id": data.team_id})
|
||||
_existing_team_row: LiteLLM_TeamTable | None = await TeamRepository(prisma_client).table.find_unique(
|
||||
where={"team_id": data.team_id}
|
||||
)
|
||||
|
||||
if _existing_team_row is None:
|
||||
raise HTTPException(
|
||||
|
|
@ -2910,7 +2923,9 @@ async def team_member_update(
|
|||
|
||||
_validate_budget_duration(data.budget_duration)
|
||||
|
||||
_existing_team_row = await TeamRepository(prisma_client).table.find_unique(where={"team_id": data.team_id})
|
||||
_existing_team_row: LiteLLM_TeamTable | None = await TeamRepository(prisma_client).table.find_unique(
|
||||
where={"team_id": data.team_id}
|
||||
)
|
||||
|
||||
if _existing_team_row is None:
|
||||
raise HTTPException(
|
||||
|
|
@ -3818,7 +3833,9 @@ async def block_team(
|
|||
if prisma_client is None:
|
||||
raise Exception("No DB Connected.")
|
||||
|
||||
existing_team = await TeamRepository(prisma_client).table.find_unique(where={"team_id": data.team_id})
|
||||
existing_team: LiteLLM_TeamTable | None = await TeamRepository(prisma_client).table.find_unique(
|
||||
where={"team_id": data.team_id}
|
||||
)
|
||||
if existing_team is None:
|
||||
raise HTTPException(
|
||||
status_code=404,
|
||||
|
|
@ -3831,7 +3848,7 @@ async def block_team(
|
|||
user_api_key_dict=user_api_key_dict,
|
||||
)
|
||||
|
||||
record = await TeamRepository(prisma_client).table.update(
|
||||
record: LiteLLM_TeamTable | None = await TeamRepository(prisma_client).table.update(
|
||||
where={"team_id": data.team_id},
|
||||
data={"blocked": True}, # type: ignore
|
||||
)
|
||||
|
|
@ -3867,7 +3884,9 @@ async def unblock_team(
|
|||
if prisma_client is None:
|
||||
raise Exception("No DB Connected.")
|
||||
|
||||
existing_team = await TeamRepository(prisma_client).table.find_unique(where={"team_id": data.team_id})
|
||||
existing_team: LiteLLM_TeamTable | None = await TeamRepository(prisma_client).table.find_unique(
|
||||
where={"team_id": data.team_id}
|
||||
)
|
||||
if existing_team is None:
|
||||
raise HTTPException(
|
||||
status_code=404,
|
||||
|
|
@ -3880,7 +3899,7 @@ async def unblock_team(
|
|||
user_api_key_dict=user_api_key_dict,
|
||||
)
|
||||
|
||||
record = await TeamRepository(prisma_client).table.update(
|
||||
record: LiteLLM_TeamTable | None = await TeamRepository(prisma_client).table.update(
|
||||
where={"team_id": data.team_id},
|
||||
data={"blocked": False}, # type: ignore
|
||||
)
|
||||
|
|
@ -3914,7 +3933,9 @@ async def list_available_teams(
|
|||
return []
|
||||
|
||||
# filter out teams that the user is already a member of
|
||||
user_info = await UserRepository(prisma_client).table.find_unique(where={"user_id": user_api_key_dict.user_id})
|
||||
user_info: LiteLLM_UserTable | None = await UserRepository(prisma_client).table.find_unique(
|
||||
where={"user_id": user_api_key_dict.user_id}
|
||||
)
|
||||
if user_info is None:
|
||||
raise HTTPException(
|
||||
status_code=404,
|
||||
|
|
@ -3924,7 +3945,9 @@ async def list_available_teams(
|
|||
|
||||
available_teams = [team for team in available_teams if team not in user_info_correct_type.teams]
|
||||
|
||||
available_teams_db = await TeamRepository(prisma_client).table.find_many(where={"team_id": {"in": available_teams}})
|
||||
available_teams_db: Sequence[LiteLLM_TeamTable] = await TeamRepository(prisma_client).table.find_many(
|
||||
where={"team_id": {"in": available_teams}}
|
||||
)
|
||||
|
||||
available_teams_correct_type = [LiteLLM_TeamTable.model_validate(team.model_dump()) for team in available_teams_db]
|
||||
|
||||
|
|
@ -3933,9 +3956,9 @@ async def list_available_teams(
|
|||
|
||||
async def _get_org_admin_org_ids(
|
||||
user_id: str,
|
||||
prisma_client: Any,
|
||||
user_api_key_cache: Any,
|
||||
proxy_logging_obj: Any,
|
||||
prisma_client: PrismaClient,
|
||||
user_api_key_cache: UserApiKeyCache,
|
||||
proxy_logging_obj: ProxyLogging,
|
||||
) -> Optional[List[str]]:
|
||||
"""
|
||||
Return the list of organization IDs where the user is an org admin.
|
||||
|
|
@ -4113,7 +4136,7 @@ def _convert_teams_to_response_models(
|
|||
|
||||
|
||||
async def _get_keys_count_by_team(
|
||||
prisma_client: Any,
|
||||
prisma_client: PrismaClient,
|
||||
teams: list,
|
||||
) -> Dict[str, int]:
|
||||
"""Aggregate virtual-key counts per team for the given page of teams.
|
||||
|
|
@ -4138,9 +4161,9 @@ async def _enforce_list_team_v2_access(
|
|||
user_api_key_dict: UserAPIKeyAuth,
|
||||
user_id: Optional[str],
|
||||
organization_id: Optional[str],
|
||||
prisma_client: Any,
|
||||
user_api_key_cache: Any,
|
||||
proxy_logging_obj: Any,
|
||||
prisma_client: PrismaClient,
|
||||
user_api_key_cache: UserApiKeyCache,
|
||||
proxy_logging_obj: ProxyLogging,
|
||||
) -> Tuple[Optional[str], Optional[List[str]]]:
|
||||
"""Enforce access control for list_team_v2.
|
||||
|
||||
|
|
@ -4391,9 +4414,9 @@ async def list_team_v2(
|
|||
async def _authorize_and_filter_teams(
|
||||
user_api_key_dict: UserAPIKeyAuth,
|
||||
user_id: Optional[str],
|
||||
prisma_client: Any,
|
||||
user_api_key_cache: Any,
|
||||
proxy_logging_obj: Any,
|
||||
prisma_client: PrismaClient,
|
||||
user_api_key_cache: UserApiKeyCache,
|
||||
proxy_logging_obj: ProxyLogging,
|
||||
) -> list:
|
||||
"""
|
||||
Authorize the /team/list request and return filtered teams.
|
||||
|
|
@ -4568,7 +4591,9 @@ async def get_paginated_teams(
|
|||
total_count = await TeamRepository(prisma_client).table.count()
|
||||
|
||||
# Get paginated teams
|
||||
teams = await TeamRepository(prisma_client).table.find_many(
|
||||
teams: list[LiteLLM_TeamTable] = await TeamRepository( # mutable-ok: matches this function's list return type
|
||||
prisma_client
|
||||
).table.find_many(
|
||||
skip=skip,
|
||||
take=page_size,
|
||||
order={"team_alias": "asc"}, # Sort by team_alias
|
||||
|
|
@ -4633,7 +4658,7 @@ async def ui_view_teams(
|
|||
}
|
||||
|
||||
# Query users with pagination and filters
|
||||
teams = await TeamRepository(prisma_client).table.find_many(
|
||||
teams: Sequence[LiteLLM_TeamTable] = await TeamRepository(prisma_client).table.find_many(
|
||||
where=where_conditions,
|
||||
skip=skip,
|
||||
take=page_size,
|
||||
|
|
@ -4701,7 +4726,9 @@ async def team_model_add(
|
|||
raise HTTPException(status_code=500, detail={"error": "No db connected"})
|
||||
|
||||
# Get existing team
|
||||
team_row = await TeamRepository(prisma_client).table.find_unique(where={"team_id": data.team_id})
|
||||
team_row: LiteLLM_TeamTable | None = await TeamRepository(prisma_client).table.find_unique(
|
||||
where={"team_id": data.team_id}
|
||||
)
|
||||
|
||||
if team_row is None:
|
||||
raise HTTPException(
|
||||
|
|
@ -4747,11 +4774,16 @@ async def team_model_add(
|
|||
# the writer and lets Prisma bump updated_at.
|
||||
# `include` mirrors the relations the auth path consumes off the cached
|
||||
# team object so that `_refresh_cached_team` doesn't null them out.
|
||||
updated_team = await TeamRepository(prisma_client).table.update(
|
||||
updated_team: LiteLLM_TeamTable | None = await TeamRepository(prisma_client).table.update(
|
||||
where={"team_id": data.team_id},
|
||||
data={"updated_at": datetime.now(timezone.utc)},
|
||||
include={"object_permission": True}, # type: ignore
|
||||
)
|
||||
if updated_team is None:
|
||||
raise HTTPException(
|
||||
status_code=404,
|
||||
detail={"error": f"Team not found, passed team_id={data.team_id}"},
|
||||
)
|
||||
|
||||
await _refresh_cached_team(
|
||||
team_row=updated_team,
|
||||
|
|
@ -4801,7 +4833,9 @@ async def team_model_delete(
|
|||
raise HTTPException(status_code=500, detail={"error": "No db connected"})
|
||||
|
||||
# Get existing team
|
||||
team_row = await TeamRepository(prisma_client).table.find_unique(where={"team_id": data.team_id})
|
||||
team_row: LiteLLM_TeamTable | None = await TeamRepository(prisma_client).table.find_unique(
|
||||
where={"team_id": data.team_id}
|
||||
)
|
||||
|
||||
if team_row is None:
|
||||
raise HTTPException(
|
||||
|
|
@ -4829,11 +4863,16 @@ async def team_model_delete(
|
|||
updated_models = [m for m in current_models if m not in data.models]
|
||||
|
||||
# Update team. See team_model_add for the rationale on `include`.
|
||||
updated_team = await TeamRepository(prisma_client).table.update(
|
||||
updated_team: LiteLLM_TeamTable | None = await TeamRepository(prisma_client).table.update(
|
||||
where={"team_id": data.team_id},
|
||||
data={"models": updated_models},
|
||||
include={"object_permission": True}, # type: ignore
|
||||
)
|
||||
if updated_team is None:
|
||||
raise HTTPException(
|
||||
status_code=404,
|
||||
detail={"error": f"Team not found, passed team_id={data.team_id}"},
|
||||
)
|
||||
|
||||
await _refresh_cached_team(
|
||||
team_row=updated_team,
|
||||
|
|
@ -4964,10 +5003,15 @@ async def update_team_member_permissions(
|
|||
},
|
||||
)
|
||||
# Update the team member permissions
|
||||
updated_team = await TeamRepository(prisma_client).table.update(
|
||||
updated_team: LiteLLM_TeamTable | None = await TeamRepository(prisma_client).table.update(
|
||||
where={"team_id": data.team_id},
|
||||
data={"team_member_permissions": data.team_member_permissions},
|
||||
)
|
||||
if updated_team is None:
|
||||
raise HTTPException(
|
||||
status_code=404,
|
||||
detail={"error": f"Team not found, passed team_id={data.team_id}"},
|
||||
)
|
||||
|
||||
return updated_team
|
||||
|
||||
|
|
@ -5058,7 +5102,7 @@ async def _compute_and_batch_updates(prisma_client, teams, permissions_to_add: s
|
|||
|
||||
async def _append_permissions_to_specific_teams(prisma_client, team_ids: List[str], permissions_to_add: set) -> int:
|
||||
"""Fetch specific teams by ID and append permissions."""
|
||||
teams = await TeamRepository(prisma_client).table.find_many(
|
||||
teams: Sequence[LiteLLM_TeamTable] = await TeamRepository(prisma_client).table.find_many(
|
||||
where={"team_id": {"in": team_ids}},
|
||||
)
|
||||
|
||||
|
|
@ -5088,7 +5132,7 @@ async def _append_permissions_to_all_teams(prisma_client, permissions_to_add: se
|
|||
find_args["cursor"] = {"team_id": cursor}
|
||||
find_args["skip"] = 1
|
||||
|
||||
teams = await TeamRepository(prisma_client).table.find_many(**find_args)
|
||||
teams: Sequence[LiteLLM_TeamTable] = await TeamRepository(prisma_client).table.find_many(**find_args)
|
||||
|
||||
if not teams:
|
||||
break
|
||||
|
|
@ -5188,8 +5232,12 @@ async def get_team_daily_activity(
|
|||
where_condition = {}
|
||||
if team_ids_list:
|
||||
where_condition["team_id"] = {"in": list(team_ids_list)}
|
||||
team_aliases = await TeamRepository(prisma_client).table.find_many(where=where_condition)
|
||||
team_alias_metadata = {t.team_id: {"team_alias": t.team_alias} for t in team_aliases}
|
||||
team_aliases: Sequence[LiteLLM_TeamTable] = await TeamRepository(prisma_client).table.find_many(
|
||||
where=where_condition
|
||||
)
|
||||
team_alias_metadata: Mapping[str, dict[str, object]] = {
|
||||
t.team_id: {"team_alias": t.team_alias} for t in team_aliases
|
||||
}
|
||||
|
||||
# Check if user is team admin or has /team/daily/activity permission
|
||||
# If not, filter by user's API keys.
|
||||
|
|
@ -5219,9 +5267,9 @@ async def get_team_daily_activity(
|
|||
# If user does not have full team view, filter by their API keys
|
||||
if not has_full_team_view:
|
||||
# Get all API keys for this user
|
||||
user_keys = await VerificationTokenRepository(prisma_client).table.find_many(
|
||||
where={"user_id": user_api_key_dict.user_id}
|
||||
)
|
||||
user_keys: Sequence[LiteLLM_VerificationToken] = await VerificationTokenRepository(
|
||||
prisma_client
|
||||
).table.find_many(where={"user_id": user_api_key_dict.user_id})
|
||||
user_api_keys = [key.token for key in user_keys if key.token]
|
||||
# If user has no API keys, return empty result
|
||||
if not user_api_keys:
|
||||
|
|
|
|||
|
|
@ -26,6 +26,7 @@ from typing import (
|
|||
List,
|
||||
Literal,
|
||||
Optional,
|
||||
Sequence,
|
||||
Set,
|
||||
Tuple,
|
||||
TypedDict,
|
||||
|
|
@ -128,7 +129,7 @@ from litellm.utils import (
|
|||
)
|
||||
|
||||
if TYPE_CHECKING:
|
||||
from aiohttp import ClientSession
|
||||
from aiohttp import ClientSession, SocketFactoryType
|
||||
from opentelemetry.trace import Span as _Span
|
||||
|
||||
from litellm.integrations.opentelemetry import OpenTelemetry
|
||||
|
|
@ -648,7 +649,7 @@ from fastapi.responses import (
|
|||
RedirectResponse,
|
||||
StreamingResponse,
|
||||
)
|
||||
from fastapi.routing import APIRouter
|
||||
from fastapi.routing import APIRoute, APIRouter
|
||||
from fastapi.security import OAuth2PasswordBearer
|
||||
from fastapi.security.api_key import APIKeyHeader
|
||||
from fastapi.staticfiles import StaticFiles
|
||||
|
|
@ -830,6 +831,15 @@ async def proxy_shutdown_event():
|
|||
cleanup_router_config_variables()
|
||||
|
||||
|
||||
class _AiohttpConnectorKwargs(TypedDict, total=False):
|
||||
keepalive_timeout: int
|
||||
ttl_dns_cache: int
|
||||
enable_cleanup_closed: bool
|
||||
limit: int
|
||||
limit_per_host: int
|
||||
socket_factory: "SocketFactoryType"
|
||||
|
||||
|
||||
async def _initialize_shared_aiohttp_session():
|
||||
"""Initialize shared aiohttp session for connection reuse with connection limits."""
|
||||
try:
|
||||
|
|
@ -839,7 +849,7 @@ async def _initialize_shared_aiohttp_session():
|
|||
_build_aiohttp_keepalive_socket_factory,
|
||||
)
|
||||
|
||||
connector_kwargs: Dict[str, Any] = {
|
||||
connector_kwargs: _AiohttpConnectorKwargs = {
|
||||
"keepalive_timeout": AIOHTTP_KEEPALIVE_TIMEOUT,
|
||||
"ttl_dns_cache": AIOHTTP_TTL_DNS_CACHE,
|
||||
}
|
||||
|
|
@ -1153,7 +1163,7 @@ async def proxy_startup_event(app: FastAPI):
|
|||
await proxy_shutdown_event() # type: ignore[reportGeneralTypeIssues]
|
||||
|
||||
|
||||
def _generate_stable_operation_id(route: Any) -> str:
|
||||
def _generate_stable_operation_id(route: APIRoute) -> str:
|
||||
operation_id = re.sub(r"\W", "_", f"{route.name}{route.path_format}")
|
||||
route_methods = sorted(route.methods or [])
|
||||
if len(route_methods) == 1:
|
||||
|
|
@ -1267,7 +1277,10 @@ vertex_live_passthrough_vertex_base = VertexBase()
|
|||
from fastapi.routing import APIWebSocketRoute
|
||||
|
||||
|
||||
def _inject_websocket_stubs_into_openapi_schema(openapi_schema: dict, websocket_routes: list) -> dict:
|
||||
def _inject_websocket_stubs_into_openapi_schema(
|
||||
openapi_schema: dict, # mutable-ok: mutated in place and returned to the caller
|
||||
websocket_routes: Sequence[APIWebSocketRoute],
|
||||
) -> dict: # mutable-ok: same object as the openapi_schema argument, mutated in place
|
||||
"""
|
||||
Add a synthetic GET stub for each WebSocket route so it appears in Swagger UI.
|
||||
|
||||
|
|
@ -1281,7 +1294,7 @@ def _inject_websocket_stubs_into_openapi_schema(openapi_schema: dict, websocket_
|
|||
|
||||
parameters = []
|
||||
try:
|
||||
if hasattr(route, "dependant") and route.dependant is not None:
|
||||
if hasattr(route, "dependant"):
|
||||
# Handle both FastAPI <0.120 and >=0.120
|
||||
query_params = getattr(route.dependant, "query_params", [])
|
||||
if query_params:
|
||||
|
|
@ -3518,7 +3531,7 @@ _DB_OVERLAY_REMOTE_MODULE_LIST_FIELDS: Dict[str, Tuple[str, ...]] = {
|
|||
}
|
||||
|
||||
|
||||
def _is_remote_module_url(value: Any) -> bool:
|
||||
def _is_remote_module_url(value: object) -> bool:
|
||||
return isinstance(value, str) and (value.startswith("s3://") or value.startswith("gcs://"))
|
||||
|
||||
|
||||
|
|
@ -5398,7 +5411,7 @@ class ProxyConfig:
|
|||
# decrypt values
|
||||
for k, v in _litellm_params.items():
|
||||
_litellm_params[k] = self._resolve_db_litellm_param(key=k, value=v)
|
||||
_litellm_params = LiteLLM_Params(**_litellm_params)
|
||||
_litellm_params = LiteLLM_Params.model_validate(_litellm_params)
|
||||
|
||||
else:
|
||||
verbose_proxy_logger.error(
|
||||
|
|
@ -5429,7 +5442,7 @@ class ProxyConfig:
|
|||
# decrypt values
|
||||
for k, v in _litellm_params.items():
|
||||
_litellm_params[k] = self._resolve_db_litellm_param(key=k, value=v)
|
||||
_litellm_params = LiteLLM_Params(**_litellm_params)
|
||||
_litellm_params = LiteLLM_Params.model_validate(_litellm_params)
|
||||
else:
|
||||
verbose_proxy_logger.error(
|
||||
f"Invalid model added to proxy db. Invalid litellm params. litellm_params={_litellm_params}"
|
||||
|
|
@ -6897,7 +6910,7 @@ class ProxyConfig:
|
|||
if isinstance(credential, dict):
|
||||
credential_object = CredentialItem(**credential)
|
||||
elif isinstance(credential, BaseModel):
|
||||
credential_object = CredentialItem(**credential.model_dump())
|
||||
credential_object = CredentialItem.model_validate(credential.model_dump())
|
||||
|
||||
decrypted_credential_values = {}
|
||||
for k, v in credential_object.credential_values.items():
|
||||
|
|
@ -13063,7 +13076,7 @@ def _get_model_group_info(
|
|||
_model_group_info = llm_router.get_model_group_info(model_group=model)
|
||||
|
||||
if _model_group_info is not None:
|
||||
model_groups.append(ModelGroupInfoProxy(**_model_group_info.model_dump()))
|
||||
model_groups.append(ModelGroupInfoProxy.model_validate(_model_group_info.model_dump()))
|
||||
else:
|
||||
model_group_info = ModelGroupInfoProxy(
|
||||
model_group=model,
|
||||
|
|
@ -14782,7 +14795,7 @@ async def update_config_general_settings(
|
|||
)
|
||||
|
||||
try:
|
||||
ConfigGeneralSettings(**{data.field_name: data.field_value})
|
||||
ConfigGeneralSettings.model_validate({data.field_name: data.field_value})
|
||||
except Exception:
|
||||
raise HTTPException(
|
||||
status_code=400,
|
||||
|
|
|
|||
|
|
@ -345,7 +345,7 @@ def _accepts_litellm_call_info(cb: CustomLogger) -> bool:
|
|||
return _CALLBACK_ACCEPTS_CALL_INFO[key]
|
||||
|
||||
|
||||
def _enrich_http_exception_with_guardrail_context(exc: BaseException, callback: Any) -> None:
|
||||
def _enrich_http_exception_with_guardrail_context(exc: BaseException, callback: CustomLogger) -> None:
|
||||
"""
|
||||
If `exc` is an HTTPException with a dict `detail`, mutate it in place to
|
||||
add `guardrail_name` and `guardrail_mode` taken from the callback instance.
|
||||
|
|
@ -394,11 +394,11 @@ class _CallbackCapabilities:
|
|||
# Tuple[(resolved_callback, "override" | "apply_guardrail"), ...]
|
||||
# Ordered the same as ``litellm.callbacks``; used to build the streaming
|
||||
# iterator chain without re-scanning per request.
|
||||
iterator_overrides: Tuple[Tuple[Any, str], ...] = field(default_factory=tuple)
|
||||
iterator_overrides: tuple[tuple[CustomLogger, str], ...] = field(default_factory=tuple)
|
||||
# Resolved CustomLogger callbacks in original order. Pre-resolving once
|
||||
# avoids the per-request ``get_custom_logger_compatible_class`` walk for
|
||||
# every string entry in ``litellm.callbacks``.
|
||||
resolved_callbacks: Tuple[Any, ...] = field(default_factory=tuple)
|
||||
resolved_callbacks: tuple[CustomLogger, ...] = field(default_factory=tuple)
|
||||
|
||||
|
||||
class ProxyLogging:
|
||||
|
|
@ -676,7 +676,7 @@ class ProxyLogging:
|
|||
|
||||
return synthetic_data
|
||||
|
||||
def _convert_llm_result_to_mcp_response(self, llm_result, request_obj) -> Optional[Any]:
|
||||
def _convert_llm_result_to_mcp_response(self, llm_result, request_obj) -> MCPPreCallResponseObject | None:
|
||||
"""
|
||||
Convert LLM guardrail result back to MCP response format.
|
||||
"""
|
||||
|
|
@ -802,7 +802,9 @@ class ProxyLogging:
|
|||
verbose_proxy_logger.error(f"Error in manual argument parsing: {e}")
|
||||
return None
|
||||
|
||||
def _convert_llm_result_to_mcp_during_response(self, llm_result, request_obj) -> Optional[Any]:
|
||||
def _convert_llm_result_to_mcp_during_response(
|
||||
self, llm_result, request_obj
|
||||
) -> MCPDuringCallResponseObject | None:
|
||||
"""
|
||||
Convert LLM guardrail result back to MCP during call response format.
|
||||
"""
|
||||
|
|
@ -1142,8 +1144,8 @@ class ProxyLogging:
|
|||
self,
|
||||
data: dict,
|
||||
litellm_logging_obj: Any,
|
||||
prompt_id: Any,
|
||||
prompt_version: Any,
|
||||
prompt_id: str,
|
||||
prompt_version: int | None,
|
||||
call_type: CallTypesLiteral,
|
||||
) -> None:
|
||||
"""Process prompt template if applicable."""
|
||||
|
|
@ -1438,9 +1440,7 @@ class ProxyLogging:
|
|||
data = result
|
||||
|
||||
elif (
|
||||
_callback is not None
|
||||
and isinstance(_callback, CustomLogger)
|
||||
and "async_pre_call_hook" in vars(_callback.__class__)
|
||||
"async_pre_call_hook" in vars(_callback.__class__)
|
||||
and _callback.__class__.async_pre_call_hook != CustomLogger.async_pre_call_hook
|
||||
):
|
||||
if call_type == "call_mcp_tool" and user_api_key_dict is None:
|
||||
|
|
@ -1614,7 +1614,7 @@ class ProxyLogging:
|
|||
break
|
||||
|
||||
@staticmethod
|
||||
async def _run_guardrail_with_metrics(callback: Any, coro: Awaitable[Any], hook_type: str) -> Any:
|
||||
async def _run_guardrail_with_metrics(callback: CustomLogger, coro: Awaitable[Any], hook_type: str) -> Any:
|
||||
"""
|
||||
Await `coro`, recording its latency and status to the
|
||||
`litellm_guardrail_latency_seconds` metric under `hook_type`, and
|
||||
|
|
@ -1646,7 +1646,7 @@ class ProxyLogging:
|
|||
|
||||
@staticmethod
|
||||
async def _wrap_streaming_iterator_with_enrichment(
|
||||
callback: Any, gen: AsyncGenerator[Any, None]
|
||||
callback: CustomLogger, gen: AsyncGenerator[Any, None]
|
||||
) -> AsyncGenerator[Any, None]:
|
||||
"""
|
||||
Yield from `gen`; if iteration raises an HTTPException with dict detail,
|
||||
|
|
@ -1691,8 +1691,11 @@ class ProxyLogging:
|
|||
has_streaming_chunk_override = False
|
||||
has_guardrail = False
|
||||
has_pre_call_override = False
|
||||
iterator_overrides: List[Tuple[Any, str]] = [] # (callback, kind)
|
||||
resolved_callbacks: List[Any] = []
|
||||
# (callback, kind) pairs; built alongside the 4 flags above and the
|
||||
# resolved_callbacks list below in one pass over callbacks, since a
|
||||
# functional per-field rewrite would re-walk this perf-sensitive scan.
|
||||
iterator_overrides: list[tuple[CustomLogger, str]] = [] # mutable-ok: see comment above
|
||||
resolved_callbacks: list[CustomLogger] = [] # mutable-ok: see comment above
|
||||
|
||||
for callback in callbacks:
|
||||
if isinstance(callback, str):
|
||||
|
|
@ -2795,7 +2798,7 @@ _DEPRECATED_KEY_CACHE_TTL_SECONDS = 60
|
|||
|
||||
|
||||
async def _lookup_deprecated_key(
|
||||
db: Any,
|
||||
db: Union[PrismaWrapper, RoutingPrismaWrapper],
|
||||
hashed_token: str,
|
||||
) -> Optional[str]:
|
||||
"""
|
||||
|
|
@ -2875,7 +2878,7 @@ def _unpack_config_row(cached: Any) -> Optional[_ConfigRow]:
|
|||
return None
|
||||
|
||||
|
||||
async def get_config_param(prisma_client: Any, param_name: str) -> Optional[Any]:
|
||||
async def get_config_param(prisma_client: "PrismaClient", param_name: str) -> Any | None:
|
||||
"""Cached read of a LiteLLM_Config row; returns row, _ConfigRow shim, or None."""
|
||||
cache_key = _config_cache_key(param_name)
|
||||
cached = await litellm_config_cache.async_get_cache(cache_key)
|
||||
|
|
@ -2893,7 +2896,7 @@ async def invalidate_config_param(param_name: str) -> None:
|
|||
await litellm_config_cache.async_delete_cache(_config_cache_key(param_name))
|
||||
|
||||
|
||||
async def prefetch_config_params(prisma_client: Any, param_names: List[str]) -> None:
|
||||
async def prefetch_config_params(prisma_client: "PrismaClient", param_names: Sequence[str]) -> None:
|
||||
"""Batch-load LiteLLM_Config rows into the cache with one find_many."""
|
||||
if not param_names:
|
||||
return
|
||||
|
|
@ -5932,7 +5935,10 @@ def _check_and_merge_model_level_guardrails(
|
|||
return _merge_guardrails_with_existing(data, model_level_guardrails)
|
||||
|
||||
|
||||
def _merge_guardrails_with_existing(data: dict, model_level_guardrails: Any) -> dict:
|
||||
def _merge_guardrails_with_existing(
|
||||
data: dict, # mutable-ok: request payload dict, needs .copy()/.setdefault()
|
||||
model_level_guardrails: Sequence[str] | str | None,
|
||||
) -> dict: # mutable-ok: returns a merged copy of the request dict, not a read-only view
|
||||
"""
|
||||
Merge model-level guardrails with any existing guardrails in the request data.
|
||||
|
||||
|
|
@ -5952,11 +5958,13 @@ def _merge_guardrails_with_existing(data: dict, model_level_guardrails: Any) ->
|
|||
existing_guardrails = [existing_guardrails] if existing_guardrails else []
|
||||
|
||||
# Ensure model_level_guardrails is a list
|
||||
if not isinstance(model_level_guardrails, list):
|
||||
model_level_guardrails = [model_level_guardrails] if model_level_guardrails else []
|
||||
if isinstance(model_level_guardrails, list):
|
||||
normalized_model_level_guardrails = model_level_guardrails
|
||||
else:
|
||||
normalized_model_level_guardrails = [model_level_guardrails] if model_level_guardrails else []
|
||||
|
||||
# Combine existing and model-level guardrails
|
||||
metadata["guardrails"] = list(set(existing_guardrails + model_level_guardrails))
|
||||
metadata["guardrails"] = list(set(existing_guardrails + normalized_model_level_guardrails))
|
||||
return modified_data
|
||||
|
||||
|
||||
|
|
|
|||
|
|
@ -1,6 +1,6 @@
|
|||
{
|
||||
"ANN001": {
|
||||
"limit": 3118
|
||||
"limit": 3116
|
||||
},
|
||||
"ANN002": {
|
||||
"limit": 69
|
||||
|
|
@ -24,7 +24,7 @@
|
|||
"limit": 130
|
||||
},
|
||||
"ANN401": {
|
||||
"limit": 2013
|
||||
"limit": 1915
|
||||
},
|
||||
"ASYNC230": {
|
||||
"limit": 14
|
||||
|
|
@ -222,7 +222,7 @@
|
|||
"limit": 38
|
||||
},
|
||||
"RET504": {
|
||||
"limit": 716
|
||||
"limit": 710
|
||||
},
|
||||
"RUF010": {
|
||||
"limit": 874
|
||||
|
|
@ -324,7 +324,7 @@
|
|||
"limit": 883
|
||||
},
|
||||
"UP006": {
|
||||
"limit": 12146
|
||||
"limit": 12137
|
||||
},
|
||||
"UP007": {
|
||||
"limit": 2526
|
||||
|
|
@ -363,6 +363,6 @@
|
|||
"limit": 105
|
||||
},
|
||||
"UP045": {
|
||||
"limit": 17806
|
||||
"limit": 17803
|
||||
}
|
||||
}
|
||||
|
|
|
|||
|
|
@ -1,6 +1,6 @@
|
|||
{
|
||||
"LIT001": {
|
||||
"limit": 23267
|
||||
"limit": 23256
|
||||
},
|
||||
"LIT002": {
|
||||
"limit": 27434
|
||||
|
|
@ -24,6 +24,6 @@
|
|||
"limit": 1004
|
||||
},
|
||||
"LIT009": {
|
||||
"limit": 2474
|
||||
"limit": 2472
|
||||
}
|
||||
}
|
||||
|
|
|
|||
Loading…
Add table
Reference in a new issue