ContextQMD
Back to Aws Doc Sdk Examples

Amazon Bedrock Tool Use Demo

python/example_code/bedrock-runtime/cross-model-scenarios/tool_use_demo/README.md

latest7.3 KB
Original Source

Amazon Bedrock Tool Use Demo

This project demonstrates a tool use scenario using Amazon Bedrock's Converse API and a weather tool. The script interacts with a foundation model on Amazon Bedrock to provide weather information based on user input. It uses the Open-Meteo API to retrieve current weather data for a given location.

⚠️ Important

  • Running this demo might result in charges to your AWS account.
  • We recommend that you grant your code least privilege. At most, grant only the minimum permissions required to perform the task. For more information, see Grant least privilege.
  • This code is not tested in every AWS Region. For more information, see AWS Regional Services.

Project overview

This directory contains the primary script, utility scripts, and the configuration required to demonstrate the weather assistant's functionality.

Primary script: tool_use_demo.py

This primary execution script manages a conversational application using the Amazon Bedrock Converse API integrated with a weather tool.

The application will interactively request weather information for locations specified by the user, demonstrating how to build a chatbot that provides real-time weather information based on user queries.

  • Usage: python tool_use_demo.py

Utility scripts

  • weather_tool.py: Provides functionality to fetch weather data from the Open-Meteo API based on geographical coordinates. This script defines the tool specification and implements the logic to retrieve weather data using latitude and longitude.
  • utils/tool_use_print_utils.py: Contains utility functions for printing headers, footers, and other formatted messages to the console, enhancing the readability of logs or messages displayed during the interaction.

Project configuration and requirements

  • requirements.txt: Lists all Python packages that need to be installed to run the project. This file is used with pip install -r requirements.txt to ensure all dependencies are met.
  • pyproject.toml: A configuration file that specifies project metadata and dependencies. It is compatible with various Python packaging tools such as pip, setuptools, pipenv, and poetry.

Prerequisites

For general prerequisites, see the README in the python folder.

Note: You must request access to an AI model on Amazon Bedrock before you can use it. If you try to use the model (with the API or console) before you have requested access to it, you will receive an error message. For more information, see Model access.

Install dependencies

Depending on your operating system and how you have Python installed, the commands to install and run might vary slightly. For example, you may have to use py or python3 in place of python.

To get started, you can choose one of the following options to set up the environment:

Using requirements.txt

1. Create a virtual environment:

shell
python -m venv .venv

A virtual environment is a self-contained directory that contains a Python installation and its associated packages, ensuring that your project's dependencies are isolated from other projects.

2. Activate the virtual environment:

  • On Linux/macOS

    shell
    source .venv/bin/activate

  • On Windows

    shell
    .\.venv\Scripts\activate

Activating the virtual environment ensures that the Python interpreter and packages used in your shell session are those from the virtual environment.

3. Install dependencies:

shell
pip install -r requirements.txt

This command installs all the packages listed in the requirements.txt file, ensuring your project has all the necessary dependencies.

Alternative: Using pyproject.toml

Alternatively, you can use the pyproject.toml file with various package managers like Poetry, PDM, or Hatch. It includes metadata about the project and a list of dependencies required for the project. Ensure you have the necessary tool installed and follow its documentation for setting up the project environment.

Run the demo

1. Run the primary script:

shell
python tool_use_demo.py

2. Follow the prompts:

  • Provide a location name or coordinates when prompted, e.g., "What's the weather like in New York?" or "Current weather for latitude 40.71, longitude -74.01".
  • The assistant will fetch and display the current weather information using the Open-Meteo API.

Note: You're not limited to single locations, or even to using English! Feel free to experiment with the app.

Step-by-step walkthrough of the process

This demo illustrates a typical interaction between a generative AI model, an application, and connected tools or APIs to solve a problem or achieve a specific goal. The flow follows these steps:

Step 1: 🛠️ Initialization

  • Set up the system prompt and tool configuration.
  • Specify the AI model to be used (e.g., Anthropic Claude 3 Sonnet).
  • Create a client to interact with Amazon Bedrock.

Step 2: 💬 User interaction

  • Prompt the user for an input related to the problem or goal, such as requesting weather information.
  • Send the user input including the conversation history to the AI model.

Step 3: 🤖 Request processing

  • The AI model processes the input and determines if a connected tool or API needs to be used.
  • If this is the case, the model returns a tool use request with specific parameters needed to invoke the tool, and a unique tool use ID to correlate tool responses to the request.

Step 4: 📡 Tool invocation

  • The application uses the tool use request to invoke the specified tool or API.
  • For example, fetching weather data from an external API using provided coordinates.
  • The tool's response, along with the related tool use ID, is then sent back to the AI model for further processing.

Step 5: 🤖 Response handling

  • The AI model processes the tool's response and generates a final response.
  • If the AI model requires additional information or further tool interactions, the process from steps 2 to 5 may be repeated.
  • This ensures that the AI model has all necessary data to generate accurate and complete responses.

Step 6. ✅ Response output

  • Once the AI model delivers the final response, the application prints the response in a clear and concise format.

This flow exemplifies how generative AI models can be integrated with applications and external tools or APIs to effectively solve user problems and achieve specific goals.

Additional Resources

Copyright Amazon.com, Inc. or its affiliates. All Rights Reserved.

SPDX-License-Identifier: Apache-2.0