PythonCodeStructuredTool

The PythonCodeStructuredTool component allows you to turn your own Python code into a reusable tool that can be called by other parts of the Nappai dashboard.
You simply paste your dataclass code, give the tool a name and a short description, and the component builds a tool object that the rest of your workflow can use.

How it Works

When you drop this component into a workflow, it reads the information you provide:

1. Tool Code – The actual Python dataclass that defines the tool’s logic.
2. Tool Name – A short identifier that will appear in the list of available tools.
3. Description – A brief explanation of what the tool does, shown to users when they select it.
4. Return Directly – A toggle that tells the system whether the tool should return the raw result of its function or wrap it in a standard response format.
5. Tool Function – A dropdown that lets you pick which function inside your code should be exposed as the tool’s main action.
6. Parent document vectorstore, Classes, Functions – Optional fields that let you link the tool to existing data stores or additional helper classes/functions.

The component then creates two outputs:

• Tool – An object that can be passed to other components that need to call a tool.
• Text – A plain text response that can be displayed directly in the dashboard.

No external APIs are called; everything runs locally inside the Nappai environment.

Inputs

• Tool Code: Enter the dataclass code that implements the tool’s logic.
• Tool Name: Provide a short, unique name for the tool.
• Description: Write a brief description of what the tool does.
• Return Directly: Check this if you want the tool to return the function’s output straight away, without any extra formatting.
• Tool Function: Choose the function from your code that should be exposed as the tool’s main action.
• Parent document vectorstore: (Optional) Connect a vectorstore that the tool can use for document retrieval.
• Classes: (Optional) List any additional classes that the tool needs.
• Functions: (Optional) List any helper functions that the tool should have access to.

Outputs

• Tool: The constructed tool object that can be used by other components.
• Text: A plain‑text response that can be shown in the dashboard or sent to a user.

Usage Example

Suppose you want to create a simple “Add Numbers” tool:

1. Tool Code

   from dataclasses import dataclass
   from typing import Any
   
   @dataclass
   class AddNumbers:
       def add(self, a: float, b: float) -> float:
           return a + b

2. Tool Name: AddNumbersTool

3. Description: Adds two numbers together.

4. Return Directly: Checked

5. Tool Function: add

Drop the component into your workflow, connect the Tool output to a “Run Tool” component, and you can now call AddNumbersTool from anywhere in your automation.

Related Components

• CustomPythonCodeStructured – The base component that this tool extends.
• Run Tool – Executes a tool object and returns its result.
• Text Response – Formats plain text for display in the dashboard.

Tips and Best Practices

• Keep your dataclass code simple and well‑documented; this makes it easier to maintain.
• Test your tool locally before adding it to a workflow to catch syntax errors early.
• Use the Return Directly option only if you need the raw output; otherwise, let Nappai handle formatting.
• If your tool needs to access external data, use the Parent document vectorstore field to link it to the appropriate store.

Security Considerations

• The code you paste runs inside Nappai’s sandboxed environment, but it still has access to the data you provide.
• Avoid including sensitive credentials directly in the code; use environment variables or secure secrets instead.
• Review any external libraries you import to ensure they are trusted and up to date.