Book Genre Predictor

A Flask-based web application that predicts the genre of a book based on its plot summary using a pre-trained CatBoost classifier and CountVectorizer. The app is built with a modern, visually appealing front-end using Bootstrap 5 and custom CSS, leveraging the CMU Book Summary Dataset for training. Deployed on an AWS EC2 instance for accessibility.

Table of Contents

• Features
• Technologies
• Prerequisites
• Setup Instructions
  • Local Setup
  • AWS EC2 Deployment
• Project Structure
• Usage
• Dataset
• Troubleshooting
• Contributing
• License

Features

• Predicts book genres (Fantasy, Science Fiction, Crime Fiction, Historical Novel, Horror, Thriller) from user-provided summaries.
• Modern, responsive UI with a cosmic-themed gradient background, glassmorphism effects, and animations.
• Input validation to handle empty summaries with user-friendly error messages.
• Deployed on AWS EC2 for scalable, cloud-based access.
• Uses a pre-trained CatBoost classifier and CountVectorizer for accurate genre prediction.

Technologies

• Backend: Python 3, Flask
• Machine Learning: CatBoost, scikit-learn, NLTK
• Frontend: HTML, Bootstrap 5, Custom CSS
• Deployment: AWS EC2, PuTTY, PuTTYgen, WinSCP
• Dependencies: Listed in requirements.txt

Prerequisites

• Local Development:
  • Python 3.8+
  • pip
  • Git
• AWS Deployment:
  • AWS account with an EC2 instance (Ubuntu recommended)
  • PuTTY and PuTTYgen for SSH access
  • WinSCP for file transfer
  • Security group configured to allow HTTP (port 80) and SSH (port 22)
• Files:
  • catboostclassifier.pkl (pre-trained model)
  • count_vectorizer.pkl (fitted CountVectorizer)
  • requirements.txt (dependency list)

Setup Instructions

Local Setup

1. Clone the Repository:

   git clone <repository-url>
   cd book-genre-predictor
   

2. Install Dependencies:

• Ensure Python 3 is installed:

python3 --version

3. Install required packages:

pip install -r requirements.txt

4. Download NLTK data:

import nltk
nltk.download('stopwords')
nltk.download('wordnet')

5. Verify Model and Vectorizer:

import pickle
from sklearn.feature_extraction.text import CountVectorizer

model_file = open('catboostclassifier.pkl', 'rb')
model = pickle.load(model_file)
model_file.close()

vectorizer_file = open('count_vectorizer.pkl', 'rb')
vectorizer = pickle.load(vectorizer_file)
vectorizer_file.close()

text = "A magical world where dragons rule the skies."
vector = vectorizer.transform([text])
prediction = model.predict(vector)
print(prediction)

6. Run the App

python app.py

AWS EC2 Deployment

• Launch an EC2 Instance:
• Create an Ubuntu EC2 instance (e.g., t2.micro).
• Download the .pem key file.
• Configure the security group to allow:
• SSH (port 22) for PuTTY access.
• HTTP (port 80) for web access.

Convert PEM to PPK:

• Use PuTTYgen to convert the .pem key to .ppk:
• Open PuTTYgen, load the .pem file, and save as .ppk.

Transfer Files with WinSCP:

• Connect to the EC2 instance using WinSCP:

• Hostname:

• Username: ubuntu

• Private key: Select the .ppk file.

• Upload the project files (app.py, templates/index.html, catboostclassifier.pkl, count_vectorizer.pkl, requirements.txt) to /home/ubuntu/book-genre-predictor.

SSH into EC2 with PuTTY:

• Open PuTTY, set:
• Hostname: ubuntu@
• Port: 8080
• Private key: Load the .ppk file under Connection > SSH > Auth.
• Connect to the instance.

Install Dependencies on EC2:

Update the package list and install Python:

sudo apt-get update
sudo apt-get install python3 python3-pip
sudo apt install python-is-python3

Navigate to the project directory:

cd /home/ubuntu/book-genre-predictor

Install Python packages:

pip install -r requirements.txt --break-system-packages

Install NLTK data:

import nltk
nltk.download('stopwords')
nltk.download('wordnet')

Run the Flask App:

python app.py

• Access the app at http://:8080.

• To keep the app running after closing PuTTY, use nohup:

nohup python app.py &

• To stop the app, find the process ID and kill it:

ps aux | grep python
kill <pid>

Project Structure

book-genre-predictor/
├── app.py                    # Flask application
├── templates/
│   └── index.html            # HTML template with Bootstrap 5 and custom CSS
├── catboostclassifier.pkl    # Pre-trained CatBoost model
├── count_vectorizer.pkl      # Fitted CountVectorizer
├── requirements.txt          # Python dependencies
└── README.md                 # This file

Usage

• Open the app in a browser (http://:8080 or http://127.0.0.1:8080 locally).

• Enter a book summary in the textarea (e.g., "A wizard embarks on a quest to defeat a dark sorcerer in a magical realm.").

• Click "Predict Genre" to see the predicted genre (e.g., Fantasy, Thriller).

• If the summary is empty, an error message will appear.

• The result displays the first 100 characters of the summary and the predicted genre in a styled alert.