nvidia-nat 1.3.0rc3__py3-none-any.whl → 1.4.0a20251010__py3-none-any.whl

nat/agent/react_agent/register.py

@@ -24,7 +24,6 @@ from nat.builder.function_info import FunctionInfo
 from nat.cli.register_workflow import register_function
 from nat.data_models.agent import AgentBaseConfig
 from nat.data_models.api_server import ChatRequest
-from nat.data_models.api_server import ChatRequestOrMessage
 from nat.data_models.api_server import ChatResponse
 from nat.data_models.api_server import Usage
 from nat.data_models.component_ref import FunctionGroupRef
@@ -71,6 +70,9 @@ class ReActAgentWorkflowConfig(AgentBaseConfig, OptimizableMixin, name="react_ag
         default=None,
         description="Provides the SYSTEM_PROMPT to use with the agent")  # defaults to SYSTEM_PROMPT in prompt.py
     max_history: int = Field(default=15, description="Maximum number of messages to keep in the conversation history.")
+    use_openai_api: bool = Field(default=False,
+                                 description=("Use OpenAI API for the input/output types to the function. "
+                                              "If False, strings will be used."))
     additional_instructions: str | None = OptimizableField(
         default=None,
         description="Additional instructions to provide to the agent in addition to the base prompt.",
@@ -116,23 +118,21 @@ async def react_agent_workflow(config: ReActAgentWorkflowConfig, builder: Builde
         pass_tool_call_errors_to_agent=config.pass_tool_call_errors_to_agent,
         normalize_tool_input_quotes=config.normalize_tool_input_quotes).build_graph()
 
-    async def _response_fn(chat_request_or_message: ChatRequestOrMessage) -> ChatResponse | str:
+    async def _response_fn(input_message: ChatRequest) -> ChatResponse:
         """
         Main workflow entry function for the ReAct Agent.
 
         This function invokes the ReAct Agent Graph and returns the response.
 
         Args:
-            chat_request_or_message (ChatRequestOrMessage): The input message to process
+            input_message (ChatRequest): The input message to process
 
         Returns:
-            ChatResponse | str: The response from the agent or error message
+            ChatResponse: The response from the agent or error message
         """
         try:
-            message = GlobalTypeConverter.get().convert(chat_request_or_message, to_type=ChatRequest)
-
             # initialize the starting state with the user query
-            messages: list[BaseMessage] = trim_messages(messages=[m.model_dump() for m in message.messages],
+            messages: list[BaseMessage] = trim_messages(messages=[m.model_dump() for m in input_message.messages],
                                                         max_tokens=config.max_history,
                                                         strategy="last",
                                                         token_counter=len,
@@ -153,16 +153,25 @@ async def react_agent_workflow(config: ReActAgentWorkflowConfig, builder: Builde
             content = str(output_message.content)
 
             # Create usage statistics for the response
-            prompt_tokens = sum(len(str(msg.content).split()) for msg in message.messages)
+            prompt_tokens = sum(len(str(msg.content).split()) for msg in input_message.messages)
             completion_tokens = len(content.split()) if content else 0
             total_tokens = prompt_tokens + completion_tokens
             usage = Usage(prompt_tokens=prompt_tokens, completion_tokens=completion_tokens, total_tokens=total_tokens)
-            response = ChatResponse.from_string(content, usage=usage)
-            if chat_request_or_message.is_string:
-                return GlobalTypeConverter.get().convert(response, to_type=str)
-            return response
+            return ChatResponse.from_string(content, usage=usage)
+
         except Exception as ex:
             logger.exception("%s ReAct Agent failed with exception: %s", AGENT_LOG_PREFIX, str(ex))
             raise RuntimeError
 
-    yield FunctionInfo.from_fn(_response_fn, description=config.description)
+    if (config.use_openai_api):
+        yield FunctionInfo.from_fn(_response_fn, description=config.description)
+    else:
+
+        async def _str_api_fn(input_message: str) -> str:
+            oai_input = GlobalTypeConverter.get().try_convert(input_message, to_type=ChatRequest)
+
+            oai_output = await _response_fn(oai_input)
+
+            return GlobalTypeConverter.get().try_convert(oai_output, to_type=str)
+
+        yield FunctionInfo.from_fn(_str_api_fn, description=config.description)

nat/agent/rewoo_agent/register.py

@@ -25,7 +25,6 @@ from nat.builder.function_info import FunctionInfo
 from nat.cli.register_workflow import register_function
 from nat.data_models.agent import AgentBaseConfig
 from nat.data_models.api_server import ChatRequest
-from nat.data_models.api_server import ChatRequestOrMessage
 from nat.data_models.api_server import ChatResponse
 from nat.data_models.api_server import Usage
 from nat.data_models.component_ref import FunctionGroupRef
@@ -55,6 +54,9 @@ class ReWOOAgentWorkflowConfig(AgentBaseConfig, name="rewoo_agent"):
                                                description="The number of retries before raising a tool call error.",
                                                ge=1)
     max_history: int = Field(default=15, description="Maximum number of messages to keep in the conversation history.")
+    use_openai_api: bool = Field(default=False,
+                                 description=("Use OpenAI API for the input/output types to the function. "
+                                              "If False, strings will be used."))
     additional_planner_instructions: str | None = Field(
         default=None,
         validation_alias=AliasChoices("additional_planner_instructions", "additional_instructions"),
@@ -123,23 +125,21 @@ async def rewoo_agent_workflow(config: ReWOOAgentWorkflowConfig, builder: Builde
         tool_call_max_retries=config.tool_call_max_retries,
         raise_tool_call_error=config.raise_tool_call_error).build_graph()
 
-    async def _response_fn(chat_request_or_message: ChatRequestOrMessage) -> ChatResponse | str:
+    async def _response_fn(input_message: ChatRequest) -> ChatResponse:
         """
         Main workflow entry function for the ReWOO Agent.
 
         This function invokes the ReWOO Agent Graph and returns the response.
 
         Args:
-            chat_request_or_message (ChatRequestOrMessage): The input message to process
+            input_message (ChatRequest): The input message to process
 
         Returns:
-            ChatResponse | str: The response from the agent or error message
+            ChatResponse: The response from the agent or error message
         """
         try:
-            message = GlobalTypeConverter.get().convert(chat_request_or_message, to_type=ChatRequest)
-
             # initialize the starting state with the user query
-            messages: list[BaseMessage] = trim_messages(messages=[m.model_dump() for m in message.messages],
+            messages: list[BaseMessage] = trim_messages(messages=[m.model_dump() for m in input_message.messages],
                                                         max_tokens=config.max_history,
                                                         strategy="last",
                                                         token_counter=len,
@@ -160,16 +160,25 @@ async def rewoo_agent_workflow(config: ReWOOAgentWorkflowConfig, builder: Builde
                 output_message = str(output_message)
 
             # Create usage statistics for the response
-            prompt_tokens = sum(len(str(msg.content).split()) for msg in message.messages)
+            prompt_tokens = sum(len(str(msg.content).split()) for msg in input_message.messages)
             completion_tokens = len(output_message.split()) if output_message else 0
             total_tokens = prompt_tokens + completion_tokens
             usage = Usage(prompt_tokens=prompt_tokens, completion_tokens=completion_tokens, total_tokens=total_tokens)
-            response = ChatResponse.from_string(output_message, usage=usage)
-            if chat_request_or_message.is_string:
-                return GlobalTypeConverter.get().convert(response, to_type=str)
-            return response
+            return ChatResponse.from_string(output_message, usage=usage)
+
         except Exception as ex:
             logger.exception("ReWOO Agent failed with exception: %s", ex)
             raise RuntimeError
 
-    yield FunctionInfo.from_fn(_response_fn, description=config.description)
+    if (config.use_openai_api):
+        yield FunctionInfo.from_fn(_response_fn, description=config.description)
+
+    else:
+
+        async def _str_api_fn(input_message: str) -> str:
+            oai_input = GlobalTypeConverter.get().try_convert(input_message, to_type=ChatRequest)
+            oai_output = await _response_fn(oai_input)
+
+            return GlobalTypeConverter.get().try_convert(oai_output, to_type=str)
+
+        yield FunctionInfo.from_fn(_str_api_fn, description=config.description)

nat/agent/tool_calling_agent/register.py

@@ -23,10 +23,8 @@ from nat.builder.function_info import FunctionInfo
 from nat.cli.register_workflow import register_function
 from nat.data_models.agent import AgentBaseConfig
 from nat.data_models.api_server import ChatRequest
-from nat.data_models.api_server import ChatRequestOrMessage
 from nat.data_models.component_ref import FunctionGroupRef
 from nat.data_models.component_ref import FunctionRef
-from nat.utils.type_converter import GlobalTypeConverter
 
 logger = logging.getLogger(__name__)
 
@@ -83,23 +81,21 @@ async def tool_calling_agent_workflow(config: ToolCallAgentWorkflowConfig, build
                                                          handle_tool_errors=config.handle_tool_errors,
                                                          return_direct=return_direct_tools).build_graph()
 
-    async def _response_fn(chat_request_or_message: ChatRequestOrMessage) -> str:
+    async def _response_fn(input_message: ChatRequest) -> str:
         """
         Main workflow entry function for the Tool Calling Agent.
 
         This function invokes the Tool Calling Agent Graph and returns the response.
 
         Args:
-            chat_request_or_message (ChatRequestOrMessage): The input message to process
+            input_message (ChatRequest): The input message to process
 
         Returns:
             str: The response from the agent or error message
         """
         try:
-            message = GlobalTypeConverter.get().convert(chat_request_or_message, to_type=ChatRequest)
-
             # initialize the starting state with the user query
-            messages: list[BaseMessage] = trim_messages(messages=[m.model_dump() for m in message.messages],
+            messages: list[BaseMessage] = trim_messages(messages=[m.model_dump() for m in input_message.messages],
                                                         max_tokens=config.max_history,
                                                         strategy="last",
                                                         token_counter=len,

nat/builder/function.py

@@ -416,9 +416,8 @@ class FunctionGroup:
         """
         if not name.strip():
             raise ValueError("Function name cannot be empty or blank")
-        if not re.match(r"^[a-zA-Z0-9_.-]+$", name):
-            raise ValueError(
-                f"Function name can only contain letters, numbers, underscores, periods, and hyphens: {name}")
+        if not re.match(r"^[a-zA-Z0-9_-]+$", name):
+            raise ValueError(f"Function name can only contain letters, numbers, underscores, and hyphens: {name}")
         if name in self._functions:
             raise ValueError(f"Function {name} already exists in function group {self._instance_name}")
 

nat/data_models/api_server.py

@@ -28,7 +28,6 @@ from pydantic import HttpUrl
 from pydantic import conlist
 from pydantic import field_serializer
 from pydantic import field_validator
-from pydantic import model_validator
 from pydantic_core.core_schema import ValidationInfo
 
 from nat.data_models.interactive import HumanPrompt
@@ -121,7 +120,15 @@ 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,
@@ -146,16 +153,6 @@ class ChatRequestOptionals(BaseModel):
     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": {
@@ -197,42 +194,6 @@ class ChatRequest(ChatRequestOptionals):
                            top_p=top_p)
 
 
-class ChatRequestOrMessage(ChatRequestOptionals):
-    """
-    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: typing.Annotated[list[Message] | None, conlist(Message, min_length=1)] = Field(
-        default=None, description="The conversation messages to process.")
-
-    input_string: str | None = Field(default=None, alias="input_message", description="The input message to process.")
-
-    @property
-    def is_string(self) -> bool:
-        return self.input_string 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)
-            if len(extra_fields) > 0:
-                raise ValueError("no extra fields are permitted when input_message/input_string is provided")
-        return self
-
-
 class ChoiceMessage(BaseModel):
     content: str | None = None
     role: UserMessageContentRoleType | None = None
@@ -700,36 +661,6 @@ def _string_to_nat_chat_request(data: str) -> ChatRequest:
 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"}))
-
-
-GlobalTypeConverter.register_converter(_chat_request_or_message_to_chat_request)
-
-
-def _chat_request_to_chat_request_or_message(data: ChatRequest) -> ChatRequestOrMessage:
-    return ChatRequestOrMessage(**data.model_dump(by_alias=True))
-
-
-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 ""
-
-
-GlobalTypeConverter.register_converter(_chat_request_or_message_to_string)
-
-
-def _string_to_chat_request_or_message(data: str) -> ChatRequestOrMessage:
-    return ChatRequestOrMessage(input_message=data)
-
-
-GlobalTypeConverter.register_converter(_string_to_chat_request_or_message)
-
-
 # ======== ChatResponse Converters ========
 def _nat_chat_response_to_string(data: ChatResponse) -> str:
     if data.choices and data.choices[0].message:

nat/utils/type_converter.py

@@ -93,14 +93,6 @@ class TypeConverter:
         if to_type is None or decomposed.is_instance(data):
             return data
 
-        # 2) If data is a union type, try to convert to each type in the union
-        if decomposed.is_union:
-            for union_type in decomposed.args:
-                result = self._convert(data, union_type)
-                if result is not None:
-                    return result
-            return None
-
         root = decomposed.root
 
         # 2) Attempt direct in *this* converter

{nvidia_nat-1.3.0rc3.dist-info → nvidia_nat-1.4.0a20251010.dist-info}/METADATA

@@ -1,10 +1,10 @@
 Metadata-Version: 2.4
 Name: nvidia-nat
-Version: 1.3.0rc3
+Version: 1.4.0a20251010
 Summary: NVIDIA NeMo Agent toolkit
 Author: NVIDIA Corporation
 Maintainer: NVIDIA Corporation
-License: Apache-2.0
+License-Expression: Apache-2.0
 Project-URL: documentation, https://docs.nvidia.com/nemo/agent-toolkit/latest/
 Project-URL: source, https://github.com/NVIDIA/NeMo-Agent-Toolkit
 Keywords: ai,rag,agents
@@ -14,8 +14,8 @@ Classifier: Programming Language :: Python :: 3.12
 Classifier: Programming Language :: Python :: 3.13
 Requires-Python: <3.14,>=3.11
 Description-Content-Type: text/markdown
-License-File: LICENSE-3rd-party.txt
 License-File: LICENSE.md
+License-File: LICENSE-3rd-party.txt
 Requires-Dist: aioboto3>=11.0.0
 Requires-Dist: authlib~=1.5
 Requires-Dist: click~=8.1

{nvidia_nat-1.3.0rc3.dist-info → nvidia_nat-1.4.0a20251010.dist-info}/RECORD

@@ -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=wAoPkly7dE8bb5x8XFf5-C1qJQausLKQwQcFCby_dwU,9307
 nat/agent/reasoning_agent/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 nat/agent/reasoning_agent/reasoning_agent.py,sha256=k_0wEDqACQn1Rn1MAKxoXyqOKsthHCQ1gt990YYUqHU,9575
 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=GfJRQgpFWl-LQ-pPaG7EUeBH5u7pDZZNVP5cSweZJdM,9599
 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/register.py,sha256=ijiRfgDVtt2p7_q1YbIQZmUVV8-jf3yT18HwtKyReUI,6822
 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=RrfKSCt9WunPhwn5fk8X7wuvb9A21iO8T-IySHUi3KM,27763
 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=V8y1v9-5p4kmaQmmDU2N6m_V_CFJeozDzJEoIHOSV8w,26177
 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
@@ -448,7 +448,7 @@ nat/utils/metadata_utils.py,sha256=BSsiB6jIWd8oEuEynJi55qCG762UuTYFaiUH0OT9HdY,2
 nat/utils/optional_imports.py,sha256=jQSVBc2fBSRw-2d6r8cEwvh5-di2EUUPakuuo9QbbwA,4039
 nat/utils/producer_consumer_queue.py,sha256=AcSYkAMBxLx06A5Xdy960PP3AJ7YaSPGJ7rbN_hJsjI,6599
 nat/utils/string_utils.py,sha256=71HuIzGx7rF8ocTmeoUBpnCi1Qf1yynYlNLLIKP4BVs,1415
-nat/utils/type_converter.py,sha256=vDZzrZ9ycWgZJdkWB1sHB2ivZX-E8fPfkrB-vAAxroI,10968
+nat/utils/type_converter.py,sha256=-2PwMsEm7tlmrniZzO7x2DnRxhOEeJGVAIJc3c5n2g4,10655
 nat/utils/type_utils.py,sha256=SMo5hM4dKf2G3U_0J0wvdFX6-lzMVSh8vd-W34Oixow,14836
 nat/utils/url_utils.py,sha256=UzDP_xaS6brWTu7vAws0B4jZyrITIK9Si3U6pZBZqDE,1028
 nat/utils/data_models/__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-1.4.0a20251010.dist-info/licenses/LICENSE-3rd-party.txt,sha256=fOk5jMmCX9YoKWyYzTtfgl-SUy477audFC5hNY4oP7Q,284609
+nvidia_nat-1.4.0a20251010.dist-info/licenses/LICENSE.md,sha256=QwcOLU5TJoTeUhuIXzhdCEEDDvorGiC6-3YTOl4TecE,11356
+nvidia_nat-1.4.0a20251010.dist-info/METADATA,sha256=00VTeVFR2xV-WeSbD9_vuaSgV-cmgEgGTxnDpualLPM,10239
+nvidia_nat-1.4.0a20251010.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
+nvidia_nat-1.4.0a20251010.dist-info/entry_points.txt,sha256=4jCqjyETMpyoWbCBf4GalZU8I_wbstpzwQNezdAVbbo,698
+nvidia_nat-1.4.0a20251010.dist-info/top_level.txt,sha256=lgJWLkigiVZuZ_O1nxVnD_ziYBwgpE2OStdaCduMEGc,8
+nvidia_nat-1.4.0a20251010.dist-info/RECORD,,