Skip to content

Xploit-Ghost/Petpooja-AI-Copilot

Repository files navigation

Petpooja AI Ecosystem 🍽️🤖

A complete, dual-sided intelligent restaurant management solution built for the hackathon. This project features two main components working in tandem:

  1. Part 1: Owner Intelligence Dashboard (Streamlit) - A backend analytics suite for pricing, combo generation, and ML demand forecasting.
  2. Part 2: AI Revenue Copilot (React) - A 100% offline, WebGPU-powered Voice/Text Point-of-Sale (POS) interface for the restaurant staff.

Part 1: Owner Intelligence Dashboard (Backend Analytics)

A powerful analytics suite built in Python that ingests POS data to optimize menu pricing, identify customer segments, and forecast future demand.

✨ Dashboard Features

  • Menu Engineering (BCG Matrix): Automatically categorizes menu items into Stars, Plowhorses, Puzzles, and Dogs based on popularity and contribution margin.
  • Smart Combo Generator: Analyzes historical co-purchase frequencies to build intelligent, high-margin combo offers (e.g., pairing Butter Chicken with Naan).
  • Price Simulator & ML Forecasting: Uses Scikit-Learn Linear Regression to predict how price changes will impact sales volume and overall profit.
  • Customer RFM Segmentation: Groups customers based on Recency, Frequency, and Monetary value to identify champions and at-risk patrons.
  • Dynamic Upsell Engine: Automatically generates and updates an upsell_config.json file based on real-time data filters.

💻 How to Run the Dashboard (Streamlit)

IMPORTANT

Prerequisites: Python 3.9+ installed on your machine.

  1. Open a terminal and navigate to the Streamlit project folder.
  2. Install the required Python libraries:
    pip install streamlit pandas plotly scikit-learn numpy
    

To Start the Streamlit Part: MODULE 1

  1. streamlit run app.py

💻 How to Run Locally (For Judges) MODULE 2

Prerequisites

  • Node.js installed on your machine.
  • A modern browser with WebGPU support (Google Chrome or Microsoft Edge recommended).
  • Important: Ensure "Use graphics acceleration when available" is turned ON in your browser settings to allow the AI to access your GPU.

Setup Instructions

  1. Clone or extract this repository and navigate to the project folder.
  2. Install the required dependencies:
    npm install
    

TO START THE SERVER AFTER ALL DEPENDENCIES ARE INSTALLED!

  1. npm run dev
  2. Ctrl + Left Click on the Localhost link(for example: http://localhost:5173)

IMPORTANT !

🛑 IMPORTANT TESTING NOTE FOR JUDGES: > On the very first run, navigate to the "Voice Copilot" tab and type a test order. The browser will download the compressed AI engine (~350MB) and cache it natively. After this initial load, completely disconnect your Wi-Fi and place another order to verify that the Natural Language Processing runs 100% offline!

VERY IMPORTANT!

RUNNING THE VOICE COPILOT PROCESSING FOR THE FIRST TIME WILL TAKE AROUND 5 MINUTES TO DOWNLOAD THE LOCAL LLM FROM HUGGING FACE WEBSITE AND WILL TAKE AROUND 360 MB OF SPACE AFTERWARDS EVERY PROCESSING WILL BE ALMOST INSTANT AND THE WEBSITE WILL BE READY TO USE

About

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors