bedrock-tools

A small Python library that simplifies Amazon Bedrock Converse API function calling (i.e., tool use).

This library reduces the boilerplate code needed to integrate native Python functions with the Amazon Bedrock Converse API, making it easier to create powerful, tool-augmented conversational AI applications.

Usage Example

pip install bedrock-tools

from bedrock_tools import BedrockTools

# define native functions as tools (using type annotations)

def add_numbers(a: int, b: int) -> int:
    """Add two numbers together."""
    return a + b


def greet(name: str) -> str:
    """Greet a person by name."""
    return f"Hello, {name}!"


def get_weather(city: str, state: str) -> dict:
    """Get the weather for a location."""
    return {
        "city": city,
        "state": state,
        "temperature": "75°F",
        "condition": "Partly Cloudy",
    }

# setup
tools = BedrockTools()
tools.add_function(add_numbers)
tools.add_function(greet)
tools.add_function(get_weather)

# Use the config in your Bedrock Converse API call
response = bedrock.converse(
    modelId=model_id,
    toolConfig=tools.get_tool_config()
    messages=messages,
)

# When you receive a toolUse from the API, invoke the tool
if "toolUse" in content_block:
    tool_results.append(tools.invoke(content_block["toolUse"]))

message = {"role": "user", "content": tool_results}

Here's an example (from the Bedrock Converse API docs) with and without the library:

Notes

Currently supports function parameters of type:

• scalar
  • str -> string
  • int -> integegr
  • bool -> boolean
  • float -> number
• list
  • str -> string
  • int -> integer
  • bool -> boolean
  • float -> number
• dict -> object

Development

 Choose a make command to run

  init      run this once to initialize a new python project
  install   install project dependencies
  start     run local project
  test      run unit tests