MetaGPT

English

中文

English

中文

Appearance

Create and Use Tools ​

Creating tools in MetaGPT is a straightforward process that involves creating your own functions or classes within the metagpt/tools/libs directory. This tutorial provides a step-by-step guide to creating tools and utilizing them within the DataInterpreter.

Steps for creating tools ​

  1. Create Pre-provided Functions or Classes:

    Craft functions or classes tailored to enable specific interactions with the external environment and place them in the metagpt/tools/libs directory.

  2. Employ Google Style Docstring:

    Accompany each function or class with a Google-style Docstring. This serves as a concise yet comprehensive reference, detailing the purpose, input parameters, and expected output.

  3. Apply @register_tool Decorator:

    Utilize the @register_tool decorator to ensure accurate registration within the tool registry. This decorator simplifies the integration of functions or classes with the DataInterpreter.

Customizing tools case ​

To illustrate the process, consider the following examples of customizing tools from function and class.

Customizing the calculate factorial tool from a function ​

  1. Create and place your own function in metagpt/tools/libs, supposing it is calculate_factorial.py, and add the decorator @register_tool for it to register as a tool.

    python
    # metagpt/tools/libs/calculate_factorial.py
import math
from metagpt.tools.tool_registry import register_tool

# Register tool with the decorator
@register_tool()
def calculate_factorial(n):
    """
    Calculate the factorial of a non-negative integer.
    """
    if n < 0:
        raise ValueError("Input must be a non-negative integer")
    return math.factorial(n)
    # metagpt/tools/libs/calculate_factorial.py
import math
from metagpt.tools.tool_registry import register_tool

# Register tool with the decorator
@register_tool()
def calculate_factorial(n):
    """
    Calculate the factorial of a non-negative integer.
    """
    if n < 0:
        raise ValueError("Input must be a non-negative integer")
    return math.factorial(n)

  2. Use DataInterpreter with your custom tool, supposing in the main.py file.

    python
    # main.py
import asyncio
from metagpt.roles.di.data_interpreter import DataInterpreter
from metagpt.tools.libs import calculate_factorial

async def main(requirement: str):
   role = DataInterpreter(tools=["calculate_factorial"])    # integrate the tool
   await role.run(requirement)

if __name__ == "__main__":
   requirement = "Please calculate the factorial of 5."
   asyncio.run(main(requirement))
    # main.py
import asyncio
from metagpt.roles.di.data_interpreter import DataInterpreter
from metagpt.tools.libs import calculate_factorial

async def main(requirement: str):
   role = DataInterpreter(tools=["calculate_factorial"])    # integrate the tool
   await role.run(requirement)

if __name__ == "__main__":
   requirement = "Please calculate the factorial of 5."
   asyncio.run(main(requirement))

Note:

  1. Don't forget to write the docstring for your function that will help to select the tool and understand how it works for the DataInterpreter.
  2. The tool name is the function name when registering the tool.
  3. Remember to import your calculate_factorial from metagpt.tools.libs to make sure the tool is registered before running DataInterpreter

Customizing the calculator tool from a class ​

  1. Create and place your own class in metagpt/tools/libs, supposing it is calculator.py, and add the decorator @register_tool for it to register as a tool.

    python
    # metagpt/tools/libs/calculator.py
import math
from metagpt.tools.tool_registry import register_tool

# Register tool with the decorator.
# The tag "math" is used to categorize the tool and the include_functions list specifies the functions to include, which makes `DataInterpreter` select and understand the tool.
@register_tool(tags=["math"], include_functions=["__init__", "add", "subtract", "multiply", "divide", "factorial"])
class Calculator:
   """
   A simple calculator tool that performs basic arithmetic operations and calculates factorials.
   """

   @staticmethod
   def add(a, b):
       """
       Calculate the sum of two numbers.
       """
       return a + b

   @staticmethod
   def subtract(a, b):
       """
       Calculate the difference of two numbers.
       """
       return a - b

   @staticmethod
   def multiply(a, b):
       """
       Calculate the product of two numbers.
       """
       return a * b

   @staticmethod
   def divide(a, b):
       """
       Calculate the quotient of two numbers.
       """
       if b == 0:
           return "Error: Division by zero"
       else:
           return a / b

   @staticmethod
   def factorial(n):
       """
       Calculate the factorial of a non-negative integer.
       """
       if n < 0:
           raise ValueError("Input must be a non-negative integer")
       return math.factorial(n)
    # metagpt/tools/libs/calculator.py
import math
from metagpt.tools.tool_registry import register_tool

# Register tool with the decorator.
# The tag "math" is used to categorize the tool and the include_functions list specifies the functions to include, which makes `DataInterpreter` select and understand the tool.
@register_tool(tags=["math"], include_functions=["__init__", "add", "subtract", "multiply", "divide", "factorial"])
class Calculator:
   """
   A simple calculator tool that performs basic arithmetic operations and calculates factorials.
   """

   @staticmethod
   def add(a, b):
       """
       Calculate the sum of two numbers.
       """
       return a + b

   @staticmethod
   def subtract(a, b):
       """
       Calculate the difference of two numbers.
       """
       return a - b

   @staticmethod
   def multiply(a, b):
       """
       Calculate the product of two numbers.
       """
       return a * b

   @staticmethod
   def divide(a, b):
       """
       Calculate the quotient of two numbers.
       """
       if b == 0:
           return "Error: Division by zero"
       else:
           return a / b

   @staticmethod
   def factorial(n):
       """
       Calculate the factorial of a non-negative integer.
       """
       if n < 0:
           raise ValueError("Input must be a non-negative integer")
       return math.factorial(n)

  2. Use DataInterpreter with your custom tool, supposing in the main.py file.

    python
    # main.py
import asyncio
from metagpt.roles.di.data_interpreter import DataInterpreter
from metagpt.tools.libs import calculator

async def main(requirement: str):
    role = DataInterpreter(tools=["Calculator"]) # integrate the tool
    await role.run(requirement)

if __name__ == "__main__":
    requirement = "Please calculate 5 plus 3 and then calculate the factorial of 5."
    asyncio.run(main(requirement))
    # main.py
import asyncio
from metagpt.roles.di.data_interpreter import DataInterpreter
from metagpt.tools.libs import calculator

async def main(requirement: str):
    role = DataInterpreter(tools=["Calculator"]) # integrate the tool
    await role.run(requirement)

if __name__ == "__main__":
    requirement = "Please calculate 5 plus 3 and then calculate the factorial of 5."
    asyncio.run(main(requirement))

Note:

  1. Don't forget to write the docstring for your class and its functions that will help to select the tool and understand how it works for the DataInterpreter.
  2. The tool name is the class name when registering the tool.
  3. Remember to import your calculator from metagpt.tools.libs to make sure the tool is registered before running DataInterpreter

By following these steps, users can seamlessly create and integrate tools within the Tool framework of MetaGPT. This empowers the DataInterpreter to effectively interact with the external environment.

Released under the MIT License.

Copyright © 2023-present alexanderwu