sp.Tool

​

Overview

The Tool class in the Outspeed package is designed to facilitate the creation of tools that can be used within Outspeed applications. These tools are typically used to perform specific tasks, such as querying an API or processing data, and are integrated into the application workflow.

​

Creating a Tool

To create a tool, define a class that inherits from sp.Tool and implement the required methods. Here’s a minimal example:

import outspeed as sp
from pydantic import BaseModel

class Query(BaseModel):
    query: str

class SearchResult(BaseModel):
    result: str

class SearchTool(sp.Tool):
    name = "search"
    description = "Search the web for information"
    parameters_type = Query
    response_type = SearchResult

    async def run(self, query: Query) -> SearchResult:
        # Implement the tool's logic here
        return SearchResult(result="Example result")

• The name attribute is a string that identifies the tool.
• The description attribute provides a brief explanation of what the tool does.
• The parameters_type and response_type attributes are Pydantic models that define the input and output data structures, respectively.
• The run method contains the main logic of the tool and must be implemented by the subclass.

​

Integrating the Tool into an Application

Once defined, the tool can be integrated into an Outspeed application. Here’s how the SearchTool is utilized within the VoiceBot:

import outspeed as sp

@sp.App()
class VoiceBot:
    async def setup(self) -> None:
        # ... existing code ...

        self.llm_node = sp.GroqLLM(
            tool_choice="required",
            tools=[
                SearchTool(),
            ],
            model="llama-3.2-90b-vision-preview",
        )

        # ... existing code ...

​

Tool Lifecycle

The Tool class manages the tool lifecycle:

1. It validates the input parameters using the parameters_type model.
2. It executes the run method asynchronously.
3. It validates the response using the response_type model.

​

Best Practices

• Use Pydantic models for parameters_type and response_type to ensure data validation.
• Implement error handling within the run method to manage exceptions gracefully.
• Test your tool thoroughly to ensure it behaves as expected within the application.

​

Error Handling

The Tool class includes basic error handling:

• If an exception occurs during the run method execution, it should be logged and handled appropriately.
• Ensure that the tool’s input and output types match the defined Pydantic models to avoid runtime errors.

On more information on how to use Pydantic models with Open AI tools, see:

1. OpenAI Tools documentation.
2. OpenAI Pydantic documentation.