This project contains a script for creating an interactive generative agent, leveraging the power of OpenAI's GPT-3.5-turbo model.

The script orange_agent.py comprises three main components:

1. ChatBot: A simple interface to OpenAI's GPT-3.5-turbo model. It takes a user's message as input and returns a generated response.

2. MemoryStore: Manages the memory of the generative agent, providing functionality to save, load, and search for relevant memories. Memories are saved as strings and searched using TF-IDF vectorization and cosine similarity.

3. GenerativeAgent: Represents a generative agent that uses both the ChatBot and MemoryStore classes. The agent generates responses to user input, learns from interactions by storing them in memory, and can provide a summary of its state.

The script ends with some sample interactions with an instance of the GenerativeAgent class. The agent is created, given some memories, asked to generate responses, aged by a few years, and finally asked to provide a summary of its state.

Run the script with python orange_agent.py and interact with the generative agent via the console.

The script requires the following Python packages:

• sqlite3
• sklearn
• os
• pickle
• openai

Install them with pip install -r requirements.txt.