nat.plugins.strands.tool_wrapper#

Attributes#

logger

Classes#

NATFunctionAgentTool

Concrete Strands AgentTool that wraps a NAT Function.

Functions#

`_json_schema_from_pydantic`(→ dict[str, Any])
`_to_tool_result`(→ dict[str, Any])
`_to_error_result`(→ dict[str, Any])
`strands_tool_wrapper`(→ NATFunctionAgentTool)	Create a Strands `AgentTool` wrapper for a NAT `Function`.

Module Contents#

logger#

_json_schema_from_pydantic( model: type[pydantic.BaseModel], ) → dict[str, Any]#

_to_tool_result(tool_use_id: str, value: Any) → dict[str, Any]#

_to_error_result(tool_use_id: str, err: Exception) → dict[str, Any]#

class NATFunctionAgentTool( name: str, description: str | None, input_schema: dict[str, Any], fn: nat.builder.function.Function, )#

Bases: strands.types.tools.AgentTool

Concrete Strands AgentTool that wraps a NAT Function.

_tool_name#

_tool_spec: strands.types.tools.ToolSpec#

_fn#

property tool_name: str#

property tool_spec: strands.types.tools.ToolSpec#

property tool_type: str#

async stream(

tool_use: strands.types.tools.ToolUse,

_invocation_state: dict[str, Any],

**_kwargs: Any,

) → collections.abc.AsyncGenerator[Any, None]#

Stream tool events and return the final result.

Parameters#

tool_useToolUse
The tool use request containing tool ID and parameters.

_invocation_statedict[str, Any]
Unused parameter for compatibility.

_kwargsAny
Unused parameter for compatibility.

Yields#

Tool events with the last being the tool result.

strands_tool_wrapper( name: str, fn: nat.builder.function.Function, _builder: nat.builder.builder.Builder, ) → NATFunctionAgentTool#: Create a Strands AgentTool wrapper for a NAT Function.