Jessup Cellars Chatbot

This repository contains the code for the Jessup Cellars Chatbot, a sophisticated conversational agent designed for a wine business. The chatbot leverages advanced Natural Language Processing (NLP) and Deep Learning techniques to provide users with information about wines, visiting hours, events, and more.

Table of Contents

• Overview
• Features
• Setup
• Usage
• Project Structure
• Skills Utilized
• License

Overview

The Jessup Cellars Chatbot is built using Flask for the backend, SocketIO for real-time communication, and a fine-tuned GPT-2 model for generating responses. It also uses Spacy for NLP tasks and integrates TF-IDF vectorization and cosine similarity for handling user queries effectively.

Features

• Real-time Chat Interface: Users can interact with the chatbot through a web-based interface.
• Wine Information: Provides detailed information about different wines.
• Visiting Hours: Answers queries regarding the business's visiting hours and location.
• Events and Experiences: Information about events and experiences offered by the business.
• Membership Details: Details about the wine club membership.

Setup

1. Clone the repository:

   git clone https://github.com/your-username/jessup-cellars-chatbot.git
   cd jessup-cellars-chatbot

2. Create a virtual environment:

   python -m venv venv
   source venv/bin/activate   # On Windows: venv\Scripts\activate

3. Install the dependencies:

   pip install -r requirements.txt

4. Download the fine-tuned GPT-2 model:

   python load_model_test.py

5. Run the application:

   python app.py

6. Open your browser and navigate to http://127.0.0.1:5000.

Usage

• Interact with the chatbot: Type your queries into the chat interface and receive instant responses.
• Intents: Use predefined buttons to ask about wine information, visiting hours, events, and more.

Project Structure

jessup-cellars-chatbot/
│
├── models/
│   └── fine_tuned_gpt/
│       └── checkpoint-15/
│           ├── config.json
│           ├── pytorch_model.bin
│           └── tokenizer_config.json
│
├── data/
│   └── corpus.json
│
├── templates/
│   └── index.html
│
├── static/
│   ├── styles.css
│   └── script.js
│
├── app.py
├── load_model_test.py
├── requirements.txt
└── README.md

Skills Utilized

• Natural Language Processing (NLP): Using Spacy for text preprocessing and understanding.
• Deep Learning: Fine-tuned GPT-2 model for generating responses.
• Web Development: Flask and SocketIO for building the web-based chat interface.
• Machine Learning: TF-IDF vectorization and cosine similarity for query matching.
• Python Programming: Integrating various libraries and frameworks to create a seamless chatbot experience.

License

This project is licensed under the MIT License - see the LICENSE file for details.