Munch

Cornell AppDev Hackathon – a platform to track meals, review foods, and split the bill with friends. Snap. Split. Rate. Dining made easy.

🍽️ A feature-rich backend for restaurant discovery, receipt parsing, and payment coordination is fully functional and deployed!

🎥 Demo video: https://www.youtube.com/watch?v=NXS4YVv5evk&ab_channel=FanhaoYu

🔨 Backend Developers: Jason Guo, Fanhao Yu
📱 Frontend Developers and Repository: Andrew Gao, Jimmy Chen

Frontend Repository Link

---

Table of Contents

• Application Description
• Tech Stack and Tools
• System Design Overview
• Database Schema
• Relationships
• Key Features
• API Endpoints Overview
• Additional

---

Application Description

Munch is a social food tracking application that allows users to:

• Log their favorite meals and restaurant items
• Review and rate food from local restaurants
• Automatically extract itemized data from receipts
• Split the bill and generate Venmo payment links
• Scrape online restaurant menus for rapid onboarding

It’s your all-in-one hub for food recommendations, receipts, and reviews.

---

Tech Stack and Tools

• Flask – Python-based server framework for all API logic.
• SQLAlchemy – ORM layer for database access and relationships.
• SQLite – Lightweight, persistent backend database for rapid development.
• OpenAI API – Used for:
  • Receipt parsing via gpt-4o
  • Fuzzy matching restaurant/food names from OCR data
• BeautifulSoup & Requests – Used in custom web scraper for importing online restaurant menus.
• Pydantic – Enforces JSON schema structure for parsed receipts.
• Docker (planned) – Will be used for future containerization and scalable deployment.

---

System Design Overview

The backend was built with modularity in mind. Here's a high-level overview of the architecture:

• Flask routes are organized by resource (Users, Food, Restaurants, etc.)
• SQLAlchemy models encapsulate all relationship logic
• Third-party APIs (OpenAI) are used in isolation via convert.py and receiptparser.py
• Tasks like scraping and parsing are cleanly separated for potential async or background handling

---

Database Schema

Our database consists of 3 core models:

• Users
• Foods
• Restaurants

And 4 join tables that support many-to-many or enriched relationships:

• favorites_table
• user_food_association_table
• user_food_reviews
• requests

Each model implements a .serialize() method to cleanly output JSON.

---

Relationships

• Users ↔️ Foods (Consumption): Many-to-many via user_food_association_table. A user can log multiple foods, and a food can be eaten by multiple users.
• Users ↔️ Foods (Favorites): Many-to-many via favorites_table. A user can favorite many foods and vice versa.
• Users ↔️ Foods (Reviews): Many-to-many via user_food_reviews, with additional fields rating and review. This is implemented as a full SQLAlchemy model.
• Restaurants → Foods: One-to-many. A restaurant can serve many foods. This is represented by a foreign key in the foods table.
• Users ↔️ Users (Payment Requests): Many-to-many via the requests table. Users can send each other payment requests, and this model includes amount and message fields for additional context.

---

Key Features

🍱 Food Logging

• Add and fetch foods
• Sort by category
• View aggregate ratings and reviews

🏬 Restaurant Management

• Add, retrieve, and delete restaurants
• View restaurant-specific menus
• Scrape restaurants (currently Pho Time) with one command

🧾 Receipt Parsing

• Upload receipt images

• Parse itemized data from resized standardized image using GPT-4o

• 🍽️ Menu Scraping

• Proof of concept automatic menu scraping for Pho Time using Beautiful Soup

• Scraping for multiple menus coming in future

🔍 Fuzzy Matching

• Convert scanned receipt items to best-match database entries using OpenAI

💸 Payment Coordination

• Generate Venmo deeplinks pre-filled with message and amount
• Request payments between users

---

API Endpoints Overview

All routes are described in the API specification. All responses are JSON-formatted and follow RESTful conventions.

Users

• GET /api/users/ – Get all users
• POST /api/users/ – Create new user
• GET /api/users/<id>/ – Fetch a specific user
• DELETE /api/users/<id>/ – Delete user by ID
• POST /api/users/<id>/food/ – Add review and rating for food
• POST /api/users/<id>/favorites/ – Favorite a food item
• GET /api/users/<id>/favorites/ – Get all favorited foods
• POST /api/payment/<id>/ – Generate Venmo link

Restaurants

• GET /api/restaurants/ – Get all restaurants
• POST /api/restaurants/ – Add a new restaurant
• GET /api/restaurants/<id>/ – Get a restaurant by ID
• DELETE /api/restaurants/<id>/ – Delete a restaurant
• GET /api/restaurants/<id>/menu/ – Get full menu
• GET /api/restaurants/<restaurant_name>/menu/ – Get restaurant ID by name

Food

• GET /api/food/ – Get all foods
• GET /api/food/categories/ – Get all categories
• GET /api/<category>/foods/ – Foods by category
• GET /api/food/<id>/ – Get food by ID
• DELETE /api/food/<id>/ – Delete food
• GET /api/food/<id>/name – Get name from food ID
• POST /api/restaurants/<id>/food/ – Add food to restaurant
• POST /api/food/<id>/image/ – Update food image
• POST /api/food/<id>/category/ – Update food category
• GET /api/food/<id>/reviews/ – All reviews for food
• GET /api/food/<category>/reviews/ – All reviews for category
• POST /api/food/reviews/ – Submit a new review

Receipt / Conversion

• POST /api/receipts/ – Upload receipt image (form-data: image)
• GET /api/convert/ – Match restaurant/item text to known DB entries
• POST /api/scrape/ – Run restaurant web scraper (currently targets Pho Time)

---

Additional

I. Authentication

Authentication is currently not enforced. Future versions may implement:

• Firebase Auth
• OAuth for social login
• Admin-level token permissions

Sample Responses:

I. Create a New User

POST /api/users/
Request:
{
  "username": "jjasonguo",
  "password": "password123",
  "email": "[email protected]",
  "phone": "7814676237",
  "profile_image": "http://example.com/image.jpg"
}

Response:
<HTTP STATUS CODE 201>
{
  "users": [
    {
      "id": <USER_ID>,
      "username": "jjasonguo",
      "email": "[email protected]",
      "phone": "7814676237",
      "venmo": "jasonguo1",
      "profile_image": "http://example.com/image.jpg",
      "foods": [<serialized food>],
      "received_requests": [<serialized requests>],
      "reviews": [
        {
          "food": [<simple serialized foods>],
          "rating": "4.5",
          "review": "Great meal"
        }
      ]
    }
  ]
}

II. Parse Receipt

Upload receipt image
POST /api/receipts/
 Form-data with "image" key.


Response
<HTTP STATUS CODE 200>
[
  {
    "assigned_friends": [],
    "items": [
        {
            "id": 1,
            "name": "Chicken Wings (6)",
            "price": 8.5
        },
        {
            "id": 2,
            "name": "P12. Chicken Pho",
            "price": 12.0
        },
        {
            "id": 3,
            "name": "P1. House Pho 1 - Pho Xe Lua",
            "price": 14.95
        },
        {
            "id": 4,
            "name": "P1. House Pho 1 - Pho Xe Lua",
            "price": 14.95
        },
        {
            "id": 5,
            "name": "P2. House Pho 2",
            "price": 14.95
        }
    ],
    "payment_total": 83.29,
    "store_name": "Asian chili spot and Photime",
    "tax": 5.23,
    "tips": 12.71,
    "total": 70.58
  }
]

III. Venmo Payment

POST /api/payment/1/
{
  "recipient_username": "jasonguo1",
  "payment_amount": "15",
  "message": "Lunch payment"
}

Response:
{
  "payment_link": "https://venmo.com/jasonguo1?txn=pay&amount=15&note=Lunch+payment"
}

IV. Get foods by category

Get all foods by category
GET /api/{category}/foods/
Response
<HTTP STATUS CODE 200>
[
  {
      "id": <id>,
"name": “Big Mac”,
"price": 5.99,
"category": “Burger”,
"image_url": “http://example.com/image.jpg”,
"avg_rating": 3.9,
"restaurant": [<serialized restaurant for food>],
  },
{
      "id": <id>,
"name": “Quarter Pounder”,
"price": 5.99,
"category": “Burger”,
"image_url": “http://example.com/image.jpg”,
"avg_rating": 3.9,
"restaurant": [<serialized restaurant for food>],
  },
	...
]

V. Get restaurant menu

Get a restaurant’s menu
GET /api/restaurants/{id}/menu/
Response
<HTTP STATUS CODE 200>
[
  {
    "id": 1,
    "name": "Beef Pho",
    “price”: “14.95”,
 	...
  },
{
    "id": 2,
    "name": "Chicken Pho",
    “price”: “13.95”,
 	...
  },
  ...
]

]