What Is An AI Agent: O'Reilly Edition

Table of Contents

1. Introduction
2. Installation
3. Usage
   • Document Question-Answering
   • Chatbot
4. File Structure
5. Contributing
6. License
7. Acknowledgments

Introduction

This repository is created as part of an O'Reilly project to explore the fundamentals of AI agents. It features two main components:

1. Document Question-Answering (src/01_document_qa.py): Utilizes a conversational model to answer questions based on a given text dataset.
2. Chatbot (src/02_chatbot.py): A simple chatbot implemented using Hugging Face's hugchat library.

Installation

Requirements

• Python 3.6+
• poetry for dependency management

Clone the repository:

git clone https://github.com/yourusername/what_is_an_ai_agent_oreilly.git

Navigate to the project directory:

cd what_is_an_ai_agent_oreilly/

Install the dependencies:

poetry install

Usage

Document Question-Answering

1. Environment Setup: Create a .env file in the project root and set your OpenAI API key:

   OPENAI_API_KEY=your_api_key_here

2. Data Preparation: Place your .csv data files in the data/sample_book_dataset/ directory.

3. Run the Script: Execute the following command:

   python src/01_document_qa.py

4. Follow the Instructions: Upload your text file when prompted and interact with the agent.

Chatbot

1. Environment Setup: Add your Hugging Face credentials to Streamlit's secrets:

   st.secrets["HUGGINGFACE_EMAIL"] = your_email_here
   st.secrets["HUGGINGFACE_PASSWORD"] = your_password_here

2. Run the Script: Execute the following command:

   streamlit run src/02_chatbot.py

3. Interact: Open the Streamlit app and start chatting with the bot.

File Structure

.
├── data/
│   ├── sample_book_dataset/
├── Makefile
├── poetry.lock
├── pyproject.toml
├── README.md
└── src/
    ├── 01_document_qa.py
    └── 02_chatbot.py

Contributing

If you wish to contribute to this project, please read the CONTRIBUTING.md guide.

License

This project is licensed under the MIT License. See the LICENSE.md file for details.

Acknowledgments

• O'Reilly for the opportunity to explore AI agents.
• OpenAI for their conversational models.
• Hugging Face for their hugchat library.