chore(api): add mypy coverage to meta_reference

Signed-off-by: Mustafa Elbehery <melbeher@redhat.com>

llama_stack/providers/inline/agents/meta_reference/openai_responses.py

@@ -91,7 +91,7 @@ async def _convert_response_content_to_chat_content(
     if isinstance(content, str):
         return content
 
-    converted_parts = []
+    converted_parts: list[OpenAIChatCompletionContentPartParam] = []
     for content_part in content:
         if isinstance(content_part, OpenAIResponseInputMessageContentText):
             converted_parts.append(OpenAIChatCompletionContentPartTextParam(text=content_part.text))
@@ -136,7 +136,7 @@ async def _convert_response_input_to_chat_messages(
                     ),
                 )
                 messages.append(OpenAIAssistantMessageParam(tool_calls=[tool_call]))
-            else:
+            elif isinstance(input_item, OpenAIResponseMessage):
                 content = await _convert_response_content_to_chat_content(input_item.content)
                 message_type = await _get_message_type_by_role(input_item.role)
                 if message_type is None:
@@ -144,6 +144,11 @@ async def _convert_response_input_to_chat_messages(
                         f"Llama Stack OpenAI Responses does not yet support message role '{input_item.role}' in this context"
                     )
                 messages.append(message_type(content=content))
+            else:
+                # Handle other tool call types that don't have content/role attributes
+                raise ValueError(
+                    f"Llama Stack OpenAI Responses does not yet support input item type '{type(input_item)}' in this context"
+                )
     else:
         messages.append(OpenAIUserMessageParam(content=input))
     return messages
@@ -175,13 +180,17 @@ async def _convert_response_text_to_chat_response_format(text: OpenAIResponseTex
     """
     Convert an OpenAI Response text parameter into an OpenAI Chat Completion response format.
     """
-    if not text.format or text.format["type"] == "text":
+    if not text.format or text.format.get("type") == "text":
         return OpenAIResponseFormatText(type="text")
-    if text.format["type"] == "json_object":
+    if text.format.get("type") == "json_object":
         return OpenAIResponseFormatJSONObject()
-    if text.format["type"] == "json_schema":
+    if text.format.get("type") == "json_schema":
+        name = text.format.get("name")
+        schema = text.format.get("schema")
+        if name is None or schema is None:
+            raise ValueError(f"json_schema format requires both name and schema fields")
         return OpenAIResponseFormatJSONSchema(
-            json_schema=OpenAIJSONSchema(name=text.format["name"], schema=text.format["schema"])
+            json_schema=OpenAIJSONSchema(name=name, schema=schema)
         )
     raise ValueError(f"Unsupported text format: {text.format}")
 
@@ -472,11 +481,12 @@ class OpenAIResponsesImpl:
                             response_tool_call = chat_response_tool_calls.get(tool_call.index, None)
                             if response_tool_call:
                                 # Don't attempt to concatenate arguments if we don't have any new argumentsAdd commentMore actions
-                                if tool_call.function.arguments:
+                                if tool_call.function and tool_call.function.arguments:
                                     # Guard against an initial None argument before we concatenate
-                                    response_tool_call.function.arguments = (
+                                    if response_tool_call.function:
-                                        response_tool_call.function.arguments or ""
+                                        response_tool_call.function.arguments = (
-                                    ) + tool_call.function.arguments
+                                            response_tool_call.function.arguments or ""
+                                        ) + tool_call.function.arguments
                             else:
                                 tool_call_dict: dict[str, Any] = tool_call.model_dump()
                                 tool_call_dict.pop("type", None)
@@ -530,15 +540,16 @@ class OpenAIResponsesImpl:
                     next_turn_messages.append(tool_response_message)
 
             for tool_call in function_tool_calls:
-                output_messages.append(
+                if tool_call.function:
-                    OpenAIResponseOutputMessageFunctionToolCall(
+                    output_messages.append(
-                        arguments=tool_call.function.arguments or "",
+                        OpenAIResponseOutputMessageFunctionToolCall(
-                        call_id=tool_call.id,
+                            arguments=tool_call.function.arguments or "",
-                        name=tool_call.function.name or "",
+                            call_id=tool_call.id,
-                        id=f"fc_{uuid.uuid4()}",
+                            name=tool_call.function.name or "",
-                        status="completed",
+                            id=f"fc_{uuid.uuid4()}",
+                            status="completed",
+                        )
                     )
-                )
 
             if not function_tool_calls and not non_function_tool_calls:
                 break
@@ -602,7 +613,7 @@ class OpenAIResponsesImpl:
                         required=param.required,
                         default=param.default,
                     )
-                    for param in tool.parameters
+                    for param in (tool.parameters or [])
                 },
             )
             return convert_tooldef_to_openai_tool(tool_def)

llama_stack/providers/inline/agents/meta_reference/persistence.py

@@ -11,7 +11,8 @@ from datetime import UTC, datetime
 
 from llama_stack.apis.agents import AgentConfig, Session, ToolExecutionStep, Turn
 from llama_stack.distribution.access_control.access_control import AccessDeniedError, is_action_allowed
-from llama_stack.distribution.access_control.datatypes import AccessRule
+from llama_stack.distribution.access_control.conditions import User as ProtectedResourceUser
+from llama_stack.distribution.access_control.datatypes import AccessRule, Action
 from llama_stack.distribution.datatypes import User
 from llama_stack.distribution.request_headers import get_authenticated_user
 from llama_stack.providers.utils.kvstore import KVStore
@@ -23,8 +24,8 @@ class AgentSessionInfo(Session):
     # TODO: is this used anywhere?
     vector_db_id: str | None = None
     started_at: datetime
-    owner: User | None = None
+    owner: ProtectedResourceUser
-    identifier: str | None = None
+    identifier: str
     type: str = "session"
 
 
@@ -43,17 +44,21 @@ class AgentPersistence:
 
         # Get current user's auth attributes for new sessions
         user = get_authenticated_user()
+        
+        # If no user is authenticated, create a default user for backward compatibility
+        if user is None:
+            user = User(principal="anonymous", attributes=None)
 
         session_info = AgentSessionInfo(
             session_id=session_id,
             session_name=name,
             started_at=datetime.now(UTC),
-            owner=user,
+            owner=user,  # User from datatypes is compatible with ProtectedResourceUser protocol
             turns=[],
             identifier=name,  # should this be qualified in any way?
         )
-        if not is_action_allowed(self.policy, "create", session_info, user):
+        if not is_action_allowed(self.policy, Action.CREATE, session_info, user):
-            raise AccessDeniedError("create", session_info, user)
+            raise AccessDeniedError(Action.CREATE, session_info, user)
 
         await self.kvstore.set(
             key=f"session:{self.agent_id}:{session_id}",
@@ -68,7 +73,17 @@ class AgentPersistence:
         if not value:
             return None
 
-        session_info = AgentSessionInfo(**json.loads(value))
+        session_data = json.loads(value)
+        
+        # Handle backward compatibility for sessions without owner field
+        if "owner" not in session_data or session_data["owner"] is None:
+            session_data["owner"] = User(principal="anonymous", attributes=None)
+        
+        # Handle backward compatibility for sessions without identifier field
+        if "identifier" not in session_data or session_data["identifier"] is None:
+            session_data["identifier"] = session_data.get("session_name", "unknown")
+
+        session_info = AgentSessionInfo(**session_data)
 
         # Check access to session
         if not self._check_session_access(session_info):
@@ -79,10 +94,10 @@ class AgentPersistence:
     def _check_session_access(self, session_info: AgentSessionInfo) -> bool:
         """Check if current user has access to the session."""
         # Handle backward compatibility for old sessions without access control
-        if not hasattr(session_info, "access_attributes") and not hasattr(session_info, "owner"):
+        if not hasattr(session_info, "access_attributes"):
             return True
 
-        return is_action_allowed(self.policy, "read", session_info, get_authenticated_user())
+        return is_action_allowed(self.policy, Action.READ, session_info, get_authenticated_user())
 
     async def get_session_if_accessible(self, session_id: str) -> AgentSessionInfo | None:
         """Get session info if the user has access to it. For internal use by sub-session methods."""

pyproject.toml

@@ -242,7 +242,7 @@ exclude = [
     "^llama_stack/models/llama/llama3/interface\\.py$",
     "^llama_stack/models/llama/llama3/tokenizer\\.py$",
     "^llama_stack/models/llama/llama3/tool_utils\\.py$",
-    "^llama_stack/providers/inline/agents/meta_reference/",
+    "^llama_stack/models/llama/llama3_3/prompts\\.py$",
     "^llama_stack/providers/inline/agents/meta_reference/agent_instance\\.py$",
     "^llama_stack/providers/inline/agents/meta_reference/agents\\.py$",
     "^llama_stack/providers/inline/datasetio/localfs/",