chore(lint): bring ANN202 and BLE001 back under the strict-rule budget

Annotate the return types of dispatch_async and transform_then_dispatch in llm_http_handler and _send_batch in azure_sentinel, and mark four legitimate broad catches with the repo's noqa convention, so the promote PR's lint job passes the strict gate again. Supersedes #40328.

litellm/integrations/azure_sentinel/azure_sentinel.py

@@ -21,6 +21,8 @@ from types import MappingProxyType
 from typing import Final, TypeVar
 from urllib.parse import urlparse
 
+import httpx
+
 from litellm._logging import verbose_logger
 from litellm.integrations.batch_utils import (
     BatchSendCancelled,
@@ -418,7 +420,7 @@ class AzureSentinelLogger(CustomBatchLogger):
             "Content-Type": "application/json",
         }
 
-        async def _send_batch(batch: Sequence[_QueuedPayload]):
+        async def _send_batch(batch: Sequence[_QueuedPayload]) -> httpx.Response:
             body: Final = safe_dumps(batch)
             return await self.async_httpx_client.post(
                 url=api_endpoint,

litellm/litellm_core_utils/get_model_cost_map.py

@@ -585,7 +585,7 @@ def _retry_remote_fetch_in_background(
         _cost_map_source_info.fallback_reason = None
         _cost_map_source_info.loaded_at = datetime.now(timezone.utc)
         adopt_model_cost_map(finalized)
-    except Exception as e:
+    except Exception as e:  # noqa: BLE001  # a failed background retry must not kill the task; the backup stays
         verbose_logger.warning("LiteLLM: Background model cost map retry failed: %s", e)
 
 

litellm/litellm_core_utils/llm_cost_calc/utils.py

@@ -1444,7 +1444,7 @@ def get_billed_token_rates(
         return _custom_pricing_rates(custom_cost_per_token)
     try:
         model_info: Final = get_model_info(model=model, custom_llm_provider=custom_llm_provider)
-    except Exception:
+    except Exception:  # noqa: BLE001  # get_model_info raises a bare Exception for an unmapped model: no rates
         return None
     return _cost_map_billed_rates(
         model_info=model_info,

litellm/llms/custom_httpx/llm_http_handler.py

@@ -579,7 +579,7 @@ class BaseLLMHTTPHandler:
             data: dict[str, object],  # mutable-ok: async_completion takes dict
             signed_headers: dict[str, object],  # mutable-ok: async_completion takes dict
             signed_json_body: bytes | None,
-        ):
+        ) -> Coroutine[object, object, ModelResponse | CustomStreamWrapper]:
             async_client: Final = client if isinstance(client, AsyncHTTPHandler) else None
             if stream is True:
                 return self.acompletion_stream_function(
@@ -626,7 +626,7 @@ class BaseLLMHTTPHandler:
 
         if acompletion is True and provider_config.uses_async_transform_request:
 
-            async def transform_then_dispatch():
+            async def transform_then_dispatch() -> ModelResponse | CustomStreamWrapper:
                 transformed: Final = cast(  # cast-ok: async_transform_request is declared as a bare dict
                     "dict[str, object]",
                     await provider_config.async_transform_request(

litellm/proxy/_experimental/mcp_server/server.py

@@ -1053,7 +1053,7 @@ if MCP_AVAILABLE:
                                 route="/mcp/call_tool",
                                 traceback_str=failure_traceback,
                             )
-                    except Exception:
+                    except Exception:  # noqa: BLE001  # a failing failure hook must not mask the tool call's own error
                         verbose_logger.exception("Error logging failed MCP proxy tool call")
                 raise
             if proxy_logging_obj is not None:

litellm/proxy/management_endpoints/cost_tracking_settings.py

@@ -643,7 +643,7 @@ async def estimate_cost(
                 custom_cost_per_token=resolved.custom_cost_per_token,
                 litellm_logging_obj=litellm_logging_obj,
             )
-        except Exception as e:
+        except Exception as e:  # noqa: BLE001  # completion_cost raises a bare Exception for an unpriceable model
             raise HTTPException(
                 status_code=404,
                 detail={