fix(mypy): fix scattered 1-off type errors across 14 files

- utils.py: Add explicit return None at end of get_provider_chat_config
- main.py: Add type: ignore for tools arg in token_counter call
- __init__.py: Add type: ignore[no-redef] for get_model_info stub
- vertex batch_embed: Annotate mode as Literal type, request_data as Dict
- vertex llama3: Add type: ignore for finish_reason = None
- types/utils.py: Add type: ignore for StreamingChoices finish_reason = None
- anthropic/files: Cast headers to httpx.Headers for AnthropicError
- key_management: Add None guard before model_dump() on object_permission
- mcp_server/db.py: Add type: ignore for TypedDict dynamic key access
- realtime_endpoints: Raise HTTPException instead of returning Response
- completion_transformation: Annotate new_tcs as list
- google_genai/main.py: Add type: ignore[valid-type] for TYPE_CHECKING classes
- brave/search: Add type: ignore[import-untyped] for dateutil

Co-authored-by: yuneng-jiang <yuneng-jiang@users.noreply.github.com>

litellm/__init__.py

@@ -1897,7 +1897,7 @@ if TYPE_CHECKING:
     supports_reasoning: Callable[..., bool]
     acreate: Callable[..., Any]
     get_max_tokens: Callable[..., int]
-    get_model_info: Callable[..., _ModelInfoType]
+    get_model_info: Callable[..., _ModelInfoType]  # type: ignore[no-redef]
     register_prompt_template: Callable[..., None]
     validate_environment: Callable[..., dict]
     check_valid_key: Callable[..., bool]

litellm/llms/anthropic/files/transformation.py

@@ -14,7 +14,7 @@ Anthropic Files API endpoints:
 
 import calendar
 import time
-from typing import Any, Dict, List, Optional, Union
+from typing import Any, Dict, List, Optional, Union, cast
 
 import httpx
 from openai.types.file_deleted import FileDeleted
@@ -79,7 +79,7 @@ class AnthropicFilesConfig(BaseFilesConfig):
         return AnthropicError(
             status_code=status_code,
             message=error_message,
-            headers=headers,
+            headers=cast(httpx.Headers, headers) if isinstance(headers, dict) else headers,
         )
 
     def validate_environment(

litellm/llms/brave/search/transformation.py

@@ -5,7 +5,7 @@ Documentation: https://api-dashboard.search.brave.com/app/documentation/web-sear
 
 from __future__ import annotations
 from datetime import datetime, timezone
-from dateutil import parser
+from dateutil import parser  # type: ignore[import-untyped]
 from typing import Dict, List, Literal, Optional, TypedDict, Union
 import httpx
 import re

litellm/llms/vertex_ai/gemini_embeddings/batch_embed_content_handler.py

@@ -153,6 +153,7 @@ class GoogleBatchEmbeddings(VertexLLM):
 
         is_multimodal = _is_multimodal_input(input)
         use_embed_content = is_multimodal or (custom_llm_provider == "vertex_ai")
+        mode: Literal["embedding", "batch_embedding"]
         if use_embed_content:
             mode = "embedding"
         else:
@@ -200,6 +201,7 @@ class GoogleBatchEmbeddings(VertexLLM):
             )
 
         ### TRANSFORMATION (sync path) ###
+        request_data: Dict[str, Any]
         if use_embed_content:
             resolved_files = {}
             if api_key:

litellm/llms/vertex_ai/vertex_ai_partner_models/llama3/transformation.py

@@ -190,7 +190,7 @@ class VertexAILlama3StreamingHandler(OpenAIChatCompletionStreamingHandler):
                     ],
                 )
                 # Modify current chunk to be the first chunk with role but no finish_reason
-                result.choices[0].finish_reason = None
+                result.choices[0].finish_reason = None  # type: ignore[assignment]
                 delta.role = "assistant"
                 # Ensure content is empty string for first chunk, not None
                 if delta.content is None:

litellm/main.py

@@ -7700,7 +7700,7 @@ async def acount_tokens(
     local_count = litellm.token_counter(
         model=model,
         messages=fallback_messages,
-        tools=tools,
+        tools=tools,  # type: ignore[arg-type]
     )
 
     return TokenCountResponse(

litellm/proxy/_experimental/mcp_server/db.py

@@ -142,9 +142,9 @@ def decrypt_credentials(
         "aws_session_token",
     ]
     for field in secret_fields:
-        value = credentials.get(field)
-        if value is not None:
-            credentials[field] = decrypt_value_helper(
+        value = credentials.get(field)  # type: ignore[literal-required]
+        if value is not None and isinstance(value, str):
+            credentials[field] = decrypt_value_helper(  # type: ignore[literal-required]
                 value=value,
                 key=field,
                 exception_type="debug",

litellm/proxy/management_endpoints/key_management_endpoints.py

@@ -1778,7 +1778,7 @@ async def _validate_mcp_servers_for_key_update(
         )
     object_permission_dict = (
         data.object_permission.model_dump()
-        if hasattr(data.object_permission, "model_dump")
+        if data.object_permission is not None and hasattr(data.object_permission, "model_dump")
         else data.object_permission
     )
     await validate_key_mcp_servers_against_team(

litellm/proxy/realtime_endpoints/endpoints.py

@@ -181,10 +181,9 @@ async def create_realtime_client_secret(
             upstream_resp.status_code,
             upstream_resp.text,
         )
-        return Response(
-            content=upstream_resp.content,
+        raise HTTPException(
             status_code=upstream_resp.status_code,
-            media_type="application/json",
+            detail=upstream_resp.text,
         )
 
     upstream_json: dict = upstream_resp.json()

litellm/responses/litellm_completion_transformation/transformation.py

@@ -410,7 +410,7 @@ class LiteLLMCompletionResponsesConfig:
                                     else getattr(new_msg, "role", None)
                                 )
                                 if new_role == "assistant":
-                                    new_tcs = (
+                                    new_tcs: list = (
                                         new_msg.get("tool_calls")
                                         if isinstance(new_msg, dict)
                                         else getattr(new_msg, "tool_calls", None)

litellm/types/google_genai/main.py

@@ -19,11 +19,11 @@ if TYPE_CHECKING:
     GenerateContentRequestParametersDict = _genai_types._GenerateContentParametersDict
     ToolConfigDict = _genai_types.ToolConfigDict
 
-    class GenerateContentRequestDict(GenerateContentRequestParametersDict):  # type: ignore[misc]
+    class GenerateContentRequestDict(GenerateContentRequestParametersDict):  # type: ignore[misc, valid-type]
         generationConfig: Optional[Any]
-        tools: Optional[ToolConfigDict]  # type: ignore[assignment]
+        tools: Optional[ToolConfigDict]  # type: ignore[assignment, valid-type]
 
-    class GenerateContentResponse(GoogleGenAIGenerateContentResponse, BaseLiteLLMOpenAIResponseObject):  # type: ignore[misc]
+    class GenerateContentResponse(GoogleGenAIGenerateContentResponse, BaseLiteLLMOpenAIResponseObject):  # type: ignore[misc, valid-type]
         _hidden_params: dict = {}
         pass
 

litellm/types/utils.py

@@ -1663,7 +1663,7 @@ class StreamingChoices(OpenAIObject):
         if finish_reason:
             self.finish_reason = map_finish_reason(finish_reason)
         else:
-            self.finish_reason = None
+            self.finish_reason = None  # type: ignore[assignment]
         self.index = index
         if delta is not None:
             if isinstance(delta, Delta):

litellm/utils.py

@@ -8141,6 +8141,8 @@ class ProviderConfigManager:
                 raise ValueError(f"Provider {provider.value} not found")
             return create_config_class(provider_config)()
 
+        return None
+
     @staticmethod
     def get_provider_embedding_config(
         model: str,