nvidia-nat 1.3.0rc3__py3-none-any.whl → 1.3.0rc5__py3-none-any.whl

nat/agent/base.py

@@ -102,11 +102,11 @@ class BaseAgent(ABC):
         AIMessage
             The LLM response
         """
-        output_message = ""
+        output_message = []
         async for event in runnable.astream(inputs, config=config):
-            output_message += event.content
+            output_message.append(event.content)
 
-        return AIMessage(content=output_message)
+        return AIMessage(content="".join(output_message))
 
     async def _call_llm(self, llm: Runnable, inputs: dict[str, Any], config: RunnableConfig | None = None) -> AIMessage:
         """

nat/agent/react_agent/register.py

@@ -162,7 +162,7 @@ async def react_agent_workflow(config: ReActAgentWorkflowConfig, builder: Builde
                 return GlobalTypeConverter.get().convert(response, to_type=str)
             return response
         except Exception as ex:
-            logger.exception("%s ReAct Agent failed with exception: %s", AGENT_LOG_PREFIX, str(ex))
-            raise RuntimeError
+            logger.error("%s ReAct Agent failed with exception: %s", AGENT_LOG_PREFIX, str(ex))
+            raise
 
     yield FunctionInfo.from_fn(_response_fn, description=config.description)

nat/agent/reasoning_agent/reasoning_agent.py

@@ -157,12 +157,12 @@ async def build_reasoning_function(config: ReasoningFunctionConfig, builder: Bui
             prompt = prompt.to_string()
 
             # Get the reasoning output from the LLM
-            reasoning_output = ""
+            reasoning_output = []
 
             async for chunk in llm.astream(prompt):
-                reasoning_output += chunk.content
+                reasoning_output.append(chunk.content)
 
-            reasoning_output = remove_r1_think_tags(reasoning_output)
+            reasoning_output = remove_r1_think_tags("".join(reasoning_output))
 
             output = await downstream_template.ainvoke(input={
                 "input_text": input_text, "reasoning_output": reasoning_output
@@ -200,12 +200,12 @@ async def build_reasoning_function(config: ReasoningFunctionConfig, builder: Bui
             prompt = prompt.to_string()
 
             # Get the reasoning output from the LLM
-            reasoning_output = ""
+            reasoning_output = []
 
             async for chunk in llm.astream(prompt):
-                reasoning_output += chunk.content
+                reasoning_output.append(chunk.content)
 
-            reasoning_output = remove_r1_think_tags(reasoning_output)
+            reasoning_output = remove_r1_think_tags("".join(reasoning_output))
 
             output = await downstream_template.ainvoke(input={
                 "input_text": input_text, "reasoning_output": reasoning_output

nat/agent/rewoo_agent/register.py

@@ -169,7 +169,7 @@ async def rewoo_agent_workflow(config: ReWOOAgentWorkflowConfig, builder: Builde
                 return GlobalTypeConverter.get().convert(response, to_type=str)
             return response
         except Exception as ex:
-            logger.exception("ReWOO Agent failed with exception: %s", ex)
-            raise RuntimeError
+            logger.error("ReWOO Agent failed with exception: %s", ex)
+            raise
 
     yield FunctionInfo.from_fn(_response_fn, description=config.description)

nat/agent/tool_calling_agent/agent.py

@@ -233,14 +233,10 @@ def create_tool_calling_agent_prompt(config: "ToolCallAgentWorkflowConfig") -> s
     """
     # the Tool Calling Agent prompt can be customized via config option system_prompt and additional_instructions.
 
-    if config.system_prompt:
-        prompt_str = config.system_prompt
-    else:
-        prompt_str = ""
-
-    if config.additional_instructions:
-        prompt_str += f" {config.additional_instructions}"
-
-    if len(prompt_str) > 0:
-        return prompt_str
+    prompt_strs = []
+    for msg in [config.system_prompt, config.additional_instructions]:
+        if msg is not None:
+            prompt_strs.append(msg)
+    if prompt_strs:
+        return " ".join(prompt_strs)
     return None

nat/agent/tool_calling_agent/register.py

@@ -118,8 +118,8 @@ async def tool_calling_agent_workflow(config: ToolCallAgentWorkflowConfig, build
             output_message = state.messages[-1]
             return str(output_message.content)
         except Exception as ex:
-            logger.exception("%s Tool Calling Agent failed with exception: %s", AGENT_LOG_PREFIX, ex)
-            raise RuntimeError
+            logger.error("%s Tool Calling Agent failed with exception: %s", AGENT_LOG_PREFIX, ex)
+            raise
 
     try:
         yield FunctionInfo.from_fn(_response_fn, description=config.description)

nat/builder/function.py

@@ -159,8 +159,7 @@ class Function(FunctionBase[InputT, StreamingOutputT, SingleOutputT], ABC):
 
                 return result
             except Exception as e:
-                err_msg = f"Error: {e}" if str(e).strip() else ""
-                logger.error("Error with ainvoke in function with input: %s. %s", value, err_msg)
+                logger.error("Error with ainvoke in function with input: %s. Error: %s", value, e)
                 raise
 
     @typing.final

nat/data_models/api_server.py

@@ -121,7 +121,15 @@ class Message(BaseModel):
     role: UserMessageContentRoleType
 
 
-class ChatRequestOptionals(BaseModel):
+class ChatRequest(BaseModel):
+    """
+    ChatRequest is a data model that represents a request to the NAT chat API.
+    Fully compatible with OpenAI Chat Completions API specification.
+    """
+
+    # Required fields
+    messages: typing.Annotated[list[Message], conlist(Message, min_length=1)]
+
     # Optional fields (OpenAI Chat Completions API compatible)
     model: str | None = Field(default=None, description="name of the model to use")
     frequency_penalty: float | None = Field(default=0.0,
@@ -145,17 +153,6 @@ class ChatRequestOptionals(BaseModel):
     tool_choice: str | dict[str, typing.Any] | None = Field(default=None, description="Controls which tool is called")
     parallel_tool_calls: bool | None = Field(default=True, description="Whether to enable parallel function calling")
     user: str | None = Field(default=None, description="Unique identifier representing end-user")
-
-
-class ChatRequest(ChatRequestOptionals):
-    """
-    ChatRequest is a data model that represents a request to the NAT chat API.
-    Fully compatible with OpenAI Chat Completions API specification.
-    """
-
-    # Required fields
-    messages: typing.Annotated[list[Message], conlist(Message, min_length=1)]
-
     model_config = ConfigDict(extra="allow",
                               json_schema_extra={
                                   "example": {
@@ -197,39 +194,82 @@ class ChatRequest(ChatRequestOptionals):
                            top_p=top_p)
 
 
-class ChatRequestOrMessage(ChatRequestOptionals):
+class ChatRequestOrMessage(BaseModel):
     """
-    ChatRequestOrMessage is a data model that represents either a conversation or a string input.
+    `ChatRequestOrMessage` is a data model that represents either a conversation or a string input.
     This is useful for functions that can handle either type of input.
 
-    `messages` is compatible with the OpenAI Chat Completions API specification.
-
-    `input_string` is a string input that can be used for functions that do not require a conversation.
-    """
+    - `messages` is compatible with the OpenAI Chat Completions API specification.
+    - `input_message` is a string input that can be used for functions that do not require a conversation.
+
+    Note: When `messages` is provided, extra fields are allowed to enable lossless round-trip
+    conversion with ChatRequest. When `input_message` is provided, no extra fields are permitted.
+    """
+    model_config = ConfigDict(
+        extra="allow",
+        json_schema_extra={
+            "examples": [
+                {
+                    "input_message": "What can you do?"
+                },
+                {
+                    "messages": [{
+                        "role": "user", "content": "What can you do?"
+                    }],
+                    "model": "nvidia/nemotron",
+                    "temperature": 0.7
+                },
+            ],
+            "oneOf": [
+                {
+                    "required": ["input_message"],
+                    "properties": {
+                        "input_message": {
+                            "type": "string"
+                        },
+                    },
+                    "additionalProperties": {
+                        "not": True, "errorMessage": 'remove additional property ${0#}'
+                    },
+                },
+                {
+                    "required": ["messages"],
+                    "properties": {
+                        "messages": {
+                            "type": "array"
+                        },
+                    },
+                    "additionalProperties": True
+                },
+            ]
+        },
+    )
 
     messages: typing.Annotated[list[Message] | None, conlist(Message, min_length=1)] = Field(
-        default=None, description="The conversation messages to process.")
+        default=None, description="A non-empty conversation of messages to process.")
 
-    input_string: str | None = Field(default=None, alias="input_message", description="The input message to process.")
+    input_message: str | None = Field(
+        default=None,
+        description="A single input message to process. Useful for functions that do not require a conversation")
 
     @property
     def is_string(self) -> bool:
-        return self.input_string is not None
+        return self.input_message is not None
 
     @property
     def is_conversation(self) -> bool:
         return self.messages is not None
 
     @model_validator(mode="after")
-    def validate_messages_or_input_string(self):
-        if self.messages is not None and self.input_string is not None:
-            raise ValueError("Either messages or input_message/input_string must be provided, not both")
-        if self.messages is None and self.input_string is None:
-            raise ValueError("Either messages or input_message/input_string must be provided")
-        if self.input_string is not None:
-            extra_fields = self.model_dump(exclude={"input_string"}, exclude_none=True, exclude_unset=True)
+    def validate_model(self):
+        if self.messages is not None and self.input_message is not None:
+            raise ValueError("Either messages or input_message must be provided, not both")
+        if self.messages is None and self.input_message is None:
+            raise ValueError("Either messages or input_message must be provided")
+        if self.input_message is not None:
+            extra_fields = self.model_dump(exclude={"input_message"}, exclude_none=True, exclude_unset=True)
             if len(extra_fields) > 0:
-                raise ValueError("no extra fields are permitted when input_message/input_string is provided")
+                raise ValueError("no extra fields are permitted when input_message is provided")
         return self
 
 
@@ -701,9 +741,9 @@ GlobalTypeConverter.register_converter(_string_to_nat_chat_request)
 
 
 def _chat_request_or_message_to_chat_request(data: ChatRequestOrMessage) -> ChatRequest:
-    if data.input_string is not None:
-        return _string_to_nat_chat_request(data.input_string)
-    return ChatRequest(**data.model_dump(exclude={"input_string"}))
+    if data.input_message is not None:
+        return _string_to_nat_chat_request(data.input_message)
+    return ChatRequest(**data.model_dump(exclude={"input_message"}))
 
 
 GlobalTypeConverter.register_converter(_chat_request_or_message_to_chat_request)
@@ -717,7 +757,17 @@ GlobalTypeConverter.register_converter(_chat_request_to_chat_request_or_message)
 
 
 def _chat_request_or_message_to_string(data: ChatRequestOrMessage) -> str:
-    return data.input_string or ""
+    if data.input_message is not None:
+        return data.input_message
+    # Extract content from last message in conversation
+    if data.messages is None:
+        return ""
+    content = data.messages[-1].content
+    if content is None:
+        return ""
+    if isinstance(content, str):
+        return content
+    return str(content)
 
 
 GlobalTypeConverter.register_converter(_chat_request_or_message_to_string)

nat/data_models/thinking_mixin.py

@@ -51,7 +51,7 @@ class ThinkingMixin(
         Returns the system prompt to use for thinking.
         For NVIDIA Nemotron, returns "/think" if enabled, else "/no_think".
         For Llama Nemotron v1.5, returns "/think" if enabled, else "/no_think".
-        For Llama Nemotron v1.0, returns "detailed thinking on" if enabled, else "detailed thinking off".
+        For Llama Nemotron v1.0 or v1.1, returns "detailed thinking on" if enabled, else "detailed thinking off".
         If thinking is not supported on the model, returns None.
 
         Returns:
@@ -72,7 +72,7 @@ class ThinkingMixin(
                 return "/think" if self.thinking else "/no_think"
 
             if model.startswith("nvidia/llama"):
-                if "v1-0" in model or "v1-1" in model:
+                if "v1-0" in model or "v1-1" in model or model.endswith("v1"):
                     return f"detailed thinking {'on' if self.thinking else 'off'}"
 
                 if "v1-5" in model:

nat/front_ends/fastapi/fastapi_front_end_plugin_worker.py

@@ -1184,6 +1184,7 @@ class FastApiFrontEndPluginWorker(FastApiFrontEndPluginWorkerBase):
                             "server": client.server_name,
                             "transport": config.server.transport,
                             "session_healthy": session_healthy,
+                            "protected": True if config.server.auth_provider is not None else False,
                             "tools": tools_info,
                             "total_tools": len(configured_short_names),
                             "available_tools": available_count
@@ -1196,6 +1197,7 @@ class FastApiFrontEndPluginWorker(FastApiFrontEndPluginWorkerBase):
                             "server": "unknown",
                             "transport": config.server.transport if config.server else "unknown",
                             "session_healthy": False,
+                            "protected": False,
                             "error": str(e),
                             "tools": [],
                             "total_tools": 0,
@@ -1226,6 +1228,7 @@ class FastApiFrontEndPluginWorker(FastApiFrontEndPluginWorkerBase):
                                     "server": "streamable-http:http://localhost:9901/mcp",
                                     "transport": "streamable-http",
                                     "session_healthy": True,
+                                    "protected": False,
                                     "tools": [{
                                         "name": "tool_a",
                                         "description": "Tool A description",

nat/front_ends/fastapi/message_handler.py

@@ -25,6 +25,7 @@ from pydantic import ValidationError
 from starlette.websockets import WebSocketDisconnect
 
 from nat.authentication.interfaces import FlowHandlerBase
+from nat.data_models.api_server import ChatRequest
 from nat.data_models.api_server import ChatResponse
 from nat.data_models.api_server import ChatResponseChunk
 from nat.data_models.api_server import Error
@@ -33,6 +34,8 @@ from nat.data_models.api_server import ResponsePayloadOutput
 from nat.data_models.api_server import ResponseSerializable
 from nat.data_models.api_server import SystemResponseContent
 from nat.data_models.api_server import TextContent
+from nat.data_models.api_server import UserMessageContentRoleType
+from nat.data_models.api_server import UserMessages
 from nat.data_models.api_server import WebSocketMessageStatus
 from nat.data_models.api_server import WebSocketMessageType
 from nat.data_models.api_server import WebSocketSystemInteractionMessage
@@ -64,12 +67,12 @@ class WebSocketMessageHandler:
         self._running_workflow_task: asyncio.Task | None = None
         self._message_parent_id: str = "default_id"
         self._conversation_id: str | None = None
-        self._workflow_schema_type: str = None
-        self._user_interaction_response: asyncio.Future[HumanResponse] | None = None
+        self._workflow_schema_type: str | None = None
+        self._user_interaction_response: asyncio.Future[TextContent] | None = None
 
         self._flow_handler: FlowHandlerBase | None = None
 
-        self._schema_output_mapping: dict[str, type[BaseModel] | None] = {
+        self._schema_output_mapping: dict[str, type[BaseModel] | type[None]] = {
             WorkflowSchemaType.GENERATE: self._session_manager.workflow.single_output_schema,
             WorkflowSchemaType.CHAT: ChatResponse,
             WorkflowSchemaType.CHAT_STREAM: ChatResponseChunk,
@@ -114,36 +117,58 @@ class WebSocketMessageHandler:
                     pass
 
                 elif (isinstance(validated_message, WebSocketUserInteractionResponseMessage)):
-                    user_content = await self.process_user_message_content(validated_message)
+                    user_content = await self._process_websocket_user_interaction_response_message(validated_message)
+                    assert self._user_interaction_response is not None
                     self._user_interaction_response.set_result(user_content)
             except (asyncio.CancelledError, WebSocketDisconnect):
                 # TODO: Handle the disconnect
                 break
 
-    async def process_user_message_content(
-            self, user_content: WebSocketUserMessage | WebSocketUserInteractionResponseMessage) -> BaseModel | None:
+    def _extract_last_user_message_content(self, messages: list[UserMessages]) -> TextContent:
         """
-        Processes the contents of a user message.
+        Extracts the last user's TextContent from a list of messages.
 
-        :param user_content: Incoming content data model.
-        :return: A validated Pydantic user content model or None if not found.
-        """
+        Args:
+            messages: List of UserMessages.
 
-        for user_message in user_content.content.messages[::-1]:
-            if (user_message.role == "user"):
+        Returns:
+            TextContent object from the last user message.
 
+        Raises:
+            ValueError: If no user text content is found.
+        """
+        for user_message in messages[::-1]:
+            if user_message.role == UserMessageContentRoleType.USER:
                 for attachment in user_message.content:
-
                     if isinstance(attachment, TextContent):
                         return attachment
+        raise ValueError("No user text content found in messages.")
+
+    async def _process_websocket_user_interaction_response_message(
+            self, user_content: WebSocketUserInteractionResponseMessage) -> TextContent:
+        """
+        Processes a WebSocketUserInteractionResponseMessage.
+        """
+        return self._extract_last_user_message_content(user_content.content.messages)
 
-        return None
+    async def _process_websocket_user_message(self, user_content: WebSocketUserMessage) -> ChatRequest | str:
+        """
+        Processes a WebSocketUserMessage based on schema type.
+        """
+        if self._workflow_schema_type in [WorkflowSchemaType.CHAT, WorkflowSchemaType.CHAT_STREAM]:
+            return ChatRequest(**user_content.content.model_dump(include={"messages"}))
+
+        elif self._workflow_schema_type in [WorkflowSchemaType.GENERATE, WorkflowSchemaType.GENERATE_STREAM]:
+            return self._extract_last_user_message_content(user_content.content.messages).text
+
+        raise ValueError("Unsupported workflow schema type for WebSocketUserMessage")
 
     async def process_workflow_request(self, user_message_as_validated_type: WebSocketUserMessage) -> None:
         """
         Process user messages and routes them appropriately.
 
-        :param user_message_as_validated_type: A WebSocketUserMessage Data Model instance.
+        Args:
+            user_message_as_validated_type (WebSocketUserMessage): The validated user message to process.
         """
 
         try:
@@ -151,18 +176,15 @@ class WebSocketMessageHandler:
             self._workflow_schema_type = user_message_as_validated_type.schema_type
             self._conversation_id = user_message_as_validated_type.conversation_id
 
-            content: BaseModel | None = await self.process_user_message_content(user_message_as_validated_type)
-
-            if content is None:
-                raise ValueError(f"User message content could not be found: {user_message_as_validated_type}")
+            message_content: typing.Any = await self._process_websocket_user_message(user_message_as_validated_type)
 
-            if isinstance(content, TextContent) and (self._running_workflow_task is None):
+            if (self._running_workflow_task is None):
 
-                def _done_callback(task: asyncio.Task):
+                def _done_callback(_task: asyncio.Task):
                     self._running_workflow_task = None
 
                 self._running_workflow_task = asyncio.create_task(
-                    self._run_workflow(payload=content.text,
+                    self._run_workflow(payload=message_content,
                                        user_message_id=self._message_parent_id,
                                        conversation_id=self._conversation_id,
                                        result_type=self._schema_output_mapping[self._workflow_schema_type],
@@ -180,13 +202,14 @@ class WebSocketMessageHandler:
     async def create_websocket_message(self,
                                        data_model: BaseModel,
                                        message_type: str | None = None,
-                                       status: str = WebSocketMessageStatus.IN_PROGRESS) -> None:
+                                       status: WebSocketMessageStatus = WebSocketMessageStatus.IN_PROGRESS) -> None:
         """
         Creates a websocket message that will be ready for routing based on message type or data model.
 
-        :param data_model: Message content model.
-        :param message_type: Message content model.
-        :param status: Message content model.
+        Args:
+            data_model (BaseModel): Message content model.
+            message_type (str | None): Message content model.
+            status (WebSocketMessageStatus): Message content model.
         """
         try:
             message: BaseModel | None = None
@@ -196,8 +219,8 @@ class WebSocketMessageHandler:
 
             message_schema: type[BaseModel] = await self._message_validator.get_message_schema_by_type(message_type)
 
-            if 'id' in data_model.model_fields:
-                message_id: str = data_model.id
+            if hasattr(data_model, 'id'):
+                message_id: str = str(getattr(data_model, 'id'))
             else:
                 message_id = str(uuid.uuid4())
 
@@ -253,12 +276,15 @@ class WebSocketMessageHandler:
         Registered human interaction callback that processes human interactions and returns
         responses from websocket connection.
 
-        :param prompt: Incoming interaction content data model.
-        :return: A Text Content Base Pydantic model.
+        Args:
+            prompt: Incoming interaction content data model.
+
+        Returns:
+            A Text Content Base Pydantic model.
         """
 
         # First create a future from the loop for the human response
-        human_response_future: asyncio.Future[HumanResponse] = asyncio.get_running_loop().create_future()
+        human_response_future: asyncio.Future[TextContent] = asyncio.get_running_loop().create_future()
 
         # Then add the future to the outstanding human prompts dictionary
         self._user_interaction_response = human_response_future
@@ -274,10 +300,10 @@ class WebSocketMessageHandler:
                 return HumanResponseNotification()
 
             # Wait for the human response future to complete
-            interaction_response: HumanResponse = await human_response_future
+            text_content: TextContent = await human_response_future
 
             interaction_response: HumanResponse = await self._message_validator.convert_text_content_to_human_response(
-                interaction_response, prompt.content)
+                text_content, prompt.content)
 
             return interaction_response
 
@@ -293,13 +319,12 @@ class WebSocketMessageHandler:
                             output_type: type | None = None) -> None:
 
         try:
-            async with self._session_manager.session(
-                    user_message_id=user_message_id,
-                    conversation_id=conversation_id,
-                    http_connection=self._socket,
-                    user_input_callback=self.human_interaction_callback,
-                    user_authentication_callback=(self._flow_handler.authenticate
-                                                  if self._flow_handler else None)) as session:
+            auth_callback = self._flow_handler.authenticate if self._flow_handler else None
+            async with self._session_manager.session(user_message_id=user_message_id,
+                                                     conversation_id=conversation_id,
+                                                     http_connection=self._socket,
+                                                     user_input_callback=self.human_interaction_callback,
+                                                     user_authentication_callback=auth_callback) as session:
 
                 async for value in generate_streaming_response(payload,
                                                                session_manager=session,

nat/front_ends/fastapi/message_validator.py

@@ -240,8 +240,7 @@ class MessageValidator:
         thread_id: str = "default",
         parent_id: str = "default",
         conversation_id: str | None = None,
-        content: SystemResponseContent
-        | Error = SystemResponseContent(),
+        content: SystemResponseContent | Error = SystemResponseContent(),
         status: WebSocketMessageStatus = WebSocketMessageStatus.IN_PROGRESS,
         timestamp: str = str(datetime.datetime.now(datetime.UTC))
     ) -> WebSocketSystemResponseTokenMessage | None:

nat/front_ends/mcp/mcp_front_end_config.py

@@ -13,13 +13,17 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
+import logging
 from typing import Literal
 
 from pydantic import Field
+from pydantic import model_validator
 
 from nat.authentication.oauth2.oauth2_resource_server_config import OAuth2ResourceServerConfig
 from nat.data_models.front_end import FrontEndBaseConfig
 
+logger = logging.getLogger(__name__)
+
 
 class MCPFrontEndConfig(FrontEndBaseConfig, name="mcp"):
     """MCP front end configuration.
@@ -43,3 +47,31 @@ class MCPFrontEndConfig(FrontEndBaseConfig, name="mcp"):
 
     server_auth: OAuth2ResourceServerConfig | None = Field(
         default=None, description=("OAuth 2.0 Resource Server configuration for token verification."))
+
+    @model_validator(mode="after")
+    def validate_security_configuration(self):
+        """Validate security configuration to prevent accidental misconfigurations."""
+        # Check if server is bound to a non-localhost interface without authentication
+        localhost_hosts = {"localhost", "127.0.0.1", "::1"}
+        if self.host not in localhost_hosts and self.server_auth is None:
+            logger.warning(
+                "MCP server is configured to bind to '%s' without authentication. "
+                "This may expose your server to unauthorized access. "
+                "Consider either: (1) binding to localhost for local-only access, "
+                "or (2) configuring server_auth for production deployments on public interfaces.",
+                self.host)
+
+        # Check if SSE transport is used (which doesn't support authentication)
+        if self.transport == "sse":
+            if self.server_auth is not None:
+                logger.warning("SSE transport does not support authentication. "
+                               "The configured server_auth will be ignored. "
+                               "For production use with authentication, use 'streamable-http' transport instead.")
+            elif self.host not in localhost_hosts:
+                logger.warning(
+                    "SSE transport does not support authentication and is bound to '%s'. "
+                    "This configuration is not recommended for production use. "
+                    "For production deployments, use 'streamable-http' transport with server_auth configured.",
+                    self.host)
+
+        return self

nat/front_ends/mcp/mcp_front_end_plugin.py

@@ -105,9 +105,12 @@ class MCPFrontEndPlugin(FrontEndBase[MCPFrontEndConfig]):
 
             # Start the MCP server with configurable transport
             # streamable-http is the default, but users can choose sse if preferred
-            if self.front_end_config.transport == "sse":
-                logger.info("Starting MCP server with SSE endpoint at /sse")
-                await mcp.run_sse_async()
-            else:  # streamable-http
-                logger.info("Starting MCP server with streamable-http endpoint at /mcp/")
-                await mcp.run_streamable_http_async()
+            try:
+                if self.front_end_config.transport == "sse":
+                    logger.info("Starting MCP server with SSE endpoint at /sse")
+                    await mcp.run_sse_async()
+                else:  # streamable-http
+                    logger.info("Starting MCP server with streamable-http endpoint at /mcp/")
+                    await mcp.run_streamable_http_async()
+            except KeyboardInterrupt:
+                logger.info("MCP server shutdown requested (Ctrl+C). Shutting down gracefully.")

nat/runtime/runner.py

@@ -196,8 +196,7 @@ class Runner:
 
             return result
         except Exception as e:
-            err_msg = f": {e}" if str(e).strip() else "."
-            logger.error("Error running workflow%s", err_msg)
+            logger.error("Error running workflow: %s", e)
             event_stream = self._context_state.event_stream.get()
             if event_stream:
                 event_stream.on_complete()

{nvidia_nat-1.3.0rc3.dist-info → nvidia_nat-1.3.0rc5.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: nvidia-nat
-Version: 1.3.0rc3
+Version: 1.3.0rc5
 Summary: NVIDIA NeMo Agent toolkit
 Author: NVIDIA Corporation
 Maintainer: NVIDIA Corporation
@@ -17,7 +17,7 @@ Description-Content-Type: text/markdown
 License-File: LICENSE-3rd-party.txt
 License-File: LICENSE.md
 Requires-Dist: aioboto3>=11.0.0
-Requires-Dist: authlib~=1.5
+Requires-Dist: authlib<2.0.0,>=1.6.5
 Requires-Dist: click~=8.1
 Requires-Dist: colorama~=0.4.6
 Requires-Dist: datasets~=4.0
@@ -103,7 +103,6 @@ Requires-Dist: nat_multi_frameworks; extra == "examples"
 Requires-Dist: nat_plot_charts; extra == "examples"
 Requires-Dist: nat_por_to_jiratickets; extra == "examples"
 Requires-Dist: nat_profiler_agent; extra == "examples"
-Requires-Dist: nat_redact_pii; extra == "examples"
 Requires-Dist: nat_router_agent; extra == "examples"
 Requires-Dist: nat_semantic_kernel_demo; extra == "examples"
 Requires-Dist: nat_sequential_executor; extra == "examples"

{nvidia_nat-1.3.0rc3.dist-info → nvidia_nat-1.3.0rc5.dist-info}/RECORD

@@ -1,6 +1,6 @@
 aiq/__init__.py,sha256=qte-NlwgM990yEeyaRUxA4IBQq3PaEkQUCK3i95iwPw,2341
 nat/agent/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-nat/agent/base.py,sha256=8cXz9Py8sFs7rTF4V9d3uzZAYFVquuWL0RQSkVMdnb0,10155
+nat/agent/base.py,sha256=Q6byRPl4mrfVFD3boKt5yftnnJ3ucH9l39FTt3QOQbg,10169
 nat/agent/dual_node.py,sha256=pfvXa1iLKtrNBHsh-tM5RWRmVe7QkyYhQNanOfWdJJs,2569
 nat/agent/register.py,sha256=rPhHDyqRBIKyon3HqhOmpAjqPP18Or6wDQb0wRUBIjQ,1032
 nat/agent/prompt_optimizer/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
@@ -10,16 +10,16 @@ nat/agent/react_agent/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3h
 nat/agent/react_agent/agent.py,sha256=sWrg9WrglTKQQyG3EcjNm2JTEchCPEo9li-Po7TJKss,21294
 nat/agent/react_agent/output_parser.py,sha256=m7K6wRwtckBBpAHqOf3BZ9mqZLwrP13Kxz5fvNxbyZE,4219
 nat/agent/react_agent/prompt.py,sha256=N47JJrT6xwYQCv1jedHhlul2AE7EfKsSYfAbgJwWRew,1758
-nat/agent/react_agent/register.py,sha256=lpiso1tKq70ZYKbV9zXZegtXPLJNBaBrnG25R9hyA9Q,9008
+nat/agent/react_agent/register.py,sha256=qkPaK6AvXjolL-q_Z3waVobXDz24GMfuqGqCn-2un2Q,8991
 nat/agent/reasoning_agent/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-nat/agent/reasoning_agent/reasoning_agent.py,sha256=k_0wEDqACQn1Rn1MAKxoXyqOKsthHCQ1gt990YYUqHU,9575
+nat/agent/reasoning_agent/reasoning_agent.py,sha256=fFQtzvaBWtmr_B6S9KSkqAfyl1BdcOc9xkhnbO4O8Pk,9603
 nat/agent/rewoo_agent/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 nat/agent/rewoo_agent/agent.py,sha256=XXgVXY9xwkyxnr093KXUtfgyNxAQbyGAecoGqN5mMLY,26199
 nat/agent/rewoo_agent/prompt.py,sha256=B0JeL1xDX4VKcShlkkviEcAsOKAwzSlX8NcAQdmUUPw,3645
-nat/agent/rewoo_agent/register.py,sha256=s6D9W4x5jIkda8l67gj3A46aefk6KQPuZ4H-ZJkVAtY,9300
+nat/agent/rewoo_agent/register.py,sha256=XArlOR37QOBtAvsdKJUjRok5qTmx39S2mJHSteOwU58,9283
 nat/agent/tool_calling_agent/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-nat/agent/tool_calling_agent/agent.py,sha256=4SIp29I56oznPRQu7B3HCoX53Ri3_o3BRRYNJjeBkF8,11006
-nat/agent/tool_calling_agent/register.py,sha256=h1Xfr1KPvQkslPg-NqdOMQAmx1PNFAIvvOC5bAIJtbE,7074
+nat/agent/tool_calling_agent/agent.py,sha256=9CRQbFlcJ02WvuRojaWcRS8ISl38JlS18BUflApuoAw,10960
+nat/agent/tool_calling_agent/register.py,sha256=OucceyELA2xZL3KdANWK9w12fnVP75eVbZgzOnmXHys,7057
 nat/authentication/__init__.py,sha256=Xs1JQ16L9btwreh4pdGKwskffAw1YFO48jKrU4ib_7c,685
 nat/authentication/interfaces.py,sha256=1J2CWEJ_n6CLA3_HD3XV28CSbyfxrPAHzr7Q4kKDFdc,3511
 nat/authentication/register.py,sha256=lFhswYUk9iZ53mq33fClR9UfjJPdjGIivGGNHQeWiYo,915
@@ -48,7 +48,7 @@ nat/builder/eval_builder.py,sha256=I-ScvupmorClYoVBIs_PhSsB7Xf9e2nGWe0rCZp3txo,6
 nat/builder/evaluator.py,sha256=xWHMND2vcAUkdFP7FU3jnVki1rUHeTa0-9saFh2hWKs,1162
 nat/builder/framework_enum.py,sha256=n7IaTQBxhFozIQqRMcX9kXntw28JhFzCj82jJ0C5tNU,901
 nat/builder/front_end.py,sha256=FCJ87NSshVVuTg8zZrq3YAr_u0RaYVZVcibnqlRFy-M,2173
-nat/builder/function.py,sha256=3h51TA0D6EQGWjHDsoxa_8ooQcZpk_-yAndk4oc5dGo,27790
+nat/builder/function.py,sha256=eZZWLwhphgQTnPvbga8sGleX7HCP46usZPIegE7zFzs,27725
 nat/builder/function_base.py,sha256=0Eg8RtjWhEU3Yme0CVxcRutobA0Qo8-YHZLI6L2qAgM,13116
 nat/builder/function_info.py,sha256=7Rmrn-gOFrT2TIJklJwA_O-ycx_oimwZ0-qMYpbuZrU,25161
 nat/builder/intermediate_step_manager.py,sha256=iOuMLWTaES0J0XzaLxhTUqFvuoCAChJu3V69T43K0k0,7599
@@ -112,7 +112,7 @@ nat/control_flow/router_agent/prompt.py,sha256=fIAiNsAs1zXRAatButR76zSpHJNxSkXXK
 nat/control_flow/router_agent/register.py,sha256=4RGmS9sy-QtIMmvh8mfMcR1VqxFPLpG4RckWCIExh40,4144
 nat/data_models/__init__.py,sha256=Xs1JQ16L9btwreh4pdGKwskffAw1YFO48jKrU4ib_7c,685
 nat/data_models/agent.py,sha256=IwDyb9Zc3R4Zd5rFeqt7q0EQswczAl5focxV9KozIzs,1625
-nat/data_models/api_server.py,sha256=NWT1ChN2qaakD2DgyYCy_7MhfzvEBQX15qnUXnpCQmk,28883
+nat/data_models/api_server.py,sha256=IowzLwxJqnSkUehCbK0WJp98hBZFXUQDd1cq8lr9PVs,30582
 nat/data_models/authentication.py,sha256=XPu9W8nh4XRSuxPv3HxO-FMQ_JtTEoK6Y02JwnzDwTg,8457
 nat/data_models/common.py,sha256=nXXfGrjpxebzBUa55mLdmzePLt7VFHvTAc6Znj3yEv0,5875
 nat/data_models/component.py,sha256=b_hXOA8Gm5UNvlFkAhsR6kEvf33ST50MKtr5kWf75Ao,1894
@@ -146,7 +146,7 @@ nat/data_models/streaming.py,sha256=sSqJqLqb70qyw69_4R9QC2RMbRw7UjTLPdo3FYBUGxE,
 nat/data_models/swe_bench_model.py,sha256=uZs-hLFuT1B5CiPFwFg1PHinDW8PHne8TBzu7tHFv_k,1718
 nat/data_models/telemetry_exporter.py,sha256=P7kqxIQnFVuvo_UFpH9QSB8fACy_0U2Uzkw_IfWXagE,998
 nat/data_models/temperature_mixin.py,sha256=LlpfWrWtDrPJfSKfNx5E0P3p5SNGZli7ACRRpmO0QqA,1628
-nat/data_models/thinking_mixin.py,sha256=lzAnUk5vyv1nTYG9ho4BD3U2NTVZ50gBysN62iGj2KM,3303
+nat/data_models/thinking_mixin.py,sha256=VRDUJZ8XP_Vv0gW2FRZUf8O9-kVgNEdZCEZ8oEmHyMk,3335
 nat/data_models/top_p_mixin.py,sha256=mu0DLnCAiwNzpSFR8FOW4kQBUpodSrvUR4MsLrNtbgA,1599
 nat/data_models/ttc_strategy.py,sha256=tAkKWcyEBmBOOYtHMtQTgeCbHxFTk5SEkmFunNVnfyE,1114
 nat/embedder/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
@@ -242,12 +242,12 @@ nat/front_ends/fastapi/dask_client_mixin.py,sha256=N_tw4yxA7EKIFTKp5_C2ZksIZucWx
 nat/front_ends/fastapi/fastapi_front_end_config.py,sha256=BcuzrVlA5b7yYyQKNvQgEanDBtKEHdpC8TAd-O7lfF0,11992
 nat/front_ends/fastapi/fastapi_front_end_controller.py,sha256=ei-34KCMpyaeAgeAN4gVvSGFjewjjRhHZPN0FqAfhDY,2548
 nat/front_ends/fastapi/fastapi_front_end_plugin.py,sha256=e33YkMcLzvm4OUG34bhl-WYiBTqkR-_wJYKG4GODkGM,11169
-nat/front_ends/fastapi/fastapi_front_end_plugin_worker.py,sha256=yrUSjbo9ge7yZi4fcFOsVFhLL5zxSh8ftZtHAExfm_s,60342
+nat/front_ends/fastapi/fastapi_front_end_plugin_worker.py,sha256=T6uslFdkHl_r0U54_7cRRKLnWYP2tTMcD7snx9Gv1xs,60547
 nat/front_ends/fastapi/intermediate_steps_subscriber.py,sha256=kbyWlBVpyvyQQjeUnFG9nsR4RaqqNkx567ZSVwwl2RU,3104
 nat/front_ends/fastapi/job_store.py,sha256=cWIBnIgRdkGL7qbBunEKzTYzdPp3l3QCDHMP-qTZJpc,22743
 nat/front_ends/fastapi/main.py,sha256=s8gXCy61rJjK1aywMRpgPvzlkMGsCS-kI_0EIy4JjBM,2445
-nat/front_ends/fastapi/message_handler.py,sha256=8pdA3K8hLCcR-ohHXYtLUgX1U2sYFzqgeslIszlQbRo,15181
-nat/front_ends/fastapi/message_validator.py,sha256=Opx9ZjaNUfS3MS6w25bq_h_XASY_i2prmQRlY_sn5xM,17614
+nat/front_ends/fastapi/message_handler.py,sha256=9CH42D-Ti5cczvRTLVuvlfCV9LaTqFGJCny46zrrEnk,16584
+nat/front_ends/fastapi/message_validator.py,sha256=9n-DvUNfqlTtPgcBMC684quD619XYSzjgPfe6f23LPQ,17606
 nat/front_ends/fastapi/register.py,sha256=rA12NPFgV9ZNHOEIgB7_SB6NytjRxgBTLo7fJ-73_HM,1153
 nat/front_ends/fastapi/response_helpers.py,sha256=MGE9E73sQSCYjsR5YXRga2qbl44hrTAPW2N5Ui3vXX0,9028
 nat/front_ends/fastapi/step_adaptor.py,sha256=J6UtoXL9De8bgAg93nE0ASLUHZbidWOfRiuFo-tyZgY,12412
@@ -259,8 +259,8 @@ nat/front_ends/fastapi/html_snippets/__init__.py,sha256=GUJrgGtpvyMUCjUBvR3faAdv
 nat/front_ends/fastapi/html_snippets/auth_code_grant_success.py,sha256=BNpWwzmA58UM0GK4kZXG4PHJy_5K9ihaVHu8SgCs5JA,1131
 nat/front_ends/mcp/__init__.py,sha256=Xs1JQ16L9btwreh4pdGKwskffAw1YFO48jKrU4ib_7c,685
 nat/front_ends/mcp/introspection_token_verifier.py,sha256=s7Q4Q6rWZJ0ZVujSxxpvVI6Bnhkg1LJQ3RLkvhzFIGE,2836
-nat/front_ends/mcp/mcp_front_end_config.py,sha256=fFbg800FDJfwv-nXc0jEPaNVtW4ApESmVRexr-werks,2313
-nat/front_ends/mcp/mcp_front_end_plugin.py,sha256=NiIIgApk1X2yAEwtG9tHaY6SexQMbZrd6Drs7uIJix8,5055
+nat/front_ends/mcp/mcp_front_end_config.py,sha256=m6z5qSz8YGnFnfu8hRID69suvO1YT_L6sxy1Ki64Ufw,4042
+nat/front_ends/mcp/mcp_front_end_plugin.py,sha256=4u_kpen_T-_Uh62V5M7dfW9KyzbqXI7tGBG4AxJXWm0,5231
 nat/front_ends/mcp/mcp_front_end_plugin_worker.py,sha256=jMclC0qEd910oTGCqd1kQ8WjP3WPdQKTl854-2bU_KI,10200
 nat/front_ends/mcp/register.py,sha256=3aJtgG5VaiqujoeU1-Eq7Hl5pWslIlIwGFU2ASLTXgM,1173
 nat/front_ends/mcp/tool_converter.py,sha256=jyH6tFKUDXSfRBKkv8WjvJsQt05zk3FJBTCwnIuUh5M,11547
@@ -406,7 +406,7 @@ nat/retriever/nemo_retriever/register.py,sha256=3XdrvEJzX2Zc8wpdm__4YYlEWBW-FK3t
 nat/retriever/nemo_retriever/retriever.py,sha256=gi3_qJFqE-iqRh3of_cmJg-SwzaQ3z24zA9LwY_MSLY,6930
 nat/runtime/__init__.py,sha256=Xs1JQ16L9btwreh4pdGKwskffAw1YFO48jKrU4ib_7c,685
 nat/runtime/loader.py,sha256=obUdAgZVYCPGC0R8u3wcoKFJzzSPQgJvrbU4OWygtog,7953
-nat/runtime/runner.py,sha256=sUF-zJMgqcFq4xRx8y5bxct2EzgiKbmFkvWkYxlDsQg,11798
+nat/runtime/runner.py,sha256=qa_AqtmB8TUHX6nVJ0TLEYCKUsm2L99kq5O72AuL3yc,11736
 nat/runtime/session.py,sha256=E8RTbnAhPbY5KCoSfiHzOJksmBh7xWjsoX0BC7Rn1ck,9101
 nat/runtime/user_metadata.py,sha256=ce37NRYJWnMOWk6A7VAQ1GQztjMmkhMOq-uYf2gNCwo,3692
 nat/settings/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
@@ -470,10 +470,10 @@ nat/utils/reactive/base/observer_base.py,sha256=6BiQfx26EMumotJ3KoVcdmFBYR_fnAss
 nat/utils/reactive/base/subject_base.py,sha256=UQOxlkZTIeeyYmG5qLtDpNf_63Y7p-doEeUA08_R8ME,2521
 nat/utils/settings/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 nat/utils/settings/global_settings.py,sha256=9JaO6pxKT_Pjw6rxJRsRlFCXdVKCl_xUKU2QHZQWWNM,7294
-nvidia_nat-1.3.0rc3.dist-info/licenses/LICENSE-3rd-party.txt,sha256=fOk5jMmCX9YoKWyYzTtfgl-SUy477audFC5hNY4oP7Q,284609
-nvidia_nat-1.3.0rc3.dist-info/licenses/LICENSE.md,sha256=QwcOLU5TJoTeUhuIXzhdCEEDDvorGiC6-3YTOl4TecE,11356
-nvidia_nat-1.3.0rc3.dist-info/METADATA,sha256=qZ0sBY6rZTYae27D_pu8g_fy-S9T4lOeReHqaNXKNOE,10222
-nvidia_nat-1.3.0rc3.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
-nvidia_nat-1.3.0rc3.dist-info/entry_points.txt,sha256=4jCqjyETMpyoWbCBf4GalZU8I_wbstpzwQNezdAVbbo,698
-nvidia_nat-1.3.0rc3.dist-info/top_level.txt,sha256=lgJWLkigiVZuZ_O1nxVnD_ziYBwgpE2OStdaCduMEGc,8
-nvidia_nat-1.3.0rc3.dist-info/RECORD,,
+nvidia_nat-1.3.0rc5.dist-info/licenses/LICENSE-3rd-party.txt,sha256=fOk5jMmCX9YoKWyYzTtfgl-SUy477audFC5hNY4oP7Q,284609
+nvidia_nat-1.3.0rc5.dist-info/licenses/LICENSE.md,sha256=QwcOLU5TJoTeUhuIXzhdCEEDDvorGiC6-3YTOl4TecE,11356
+nvidia_nat-1.3.0rc5.dist-info/METADATA,sha256=nm0UvCzWa259-7OnT21duAI9yl9EqwKs3lXaxup59OA,10180
+nvidia_nat-1.3.0rc5.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
+nvidia_nat-1.3.0rc5.dist-info/entry_points.txt,sha256=4jCqjyETMpyoWbCBf4GalZU8I_wbstpzwQNezdAVbbo,698
+nvidia_nat-1.3.0rc5.dist-info/top_level.txt,sha256=lgJWLkigiVZuZ_O1nxVnD_ziYBwgpE2OStdaCduMEGc,8
+nvidia_nat-1.3.0rc5.dist-info/RECORD,,