/waste_wise

Smart waste management system

Primary LanguagePython

Waste Wise - Smart Waste Management System

Waste Wise is a smart waste management system designed to streamline and optimize waste collection. It allows users to request pickups via a mobile app, scan their waste for reuse tips, and ensures efficient clustering and assignment of pickup requests to drivers based on location.

System Overview

System Diagram

Key Components

  1. Customer and Picker (Driver)

    • Customers can request waste pickups via the mobile app.
    • Pickers (drivers) receive notifications of assigned pickups, perform the collection, and mark it as completed.

  2. Mobile App

    • Interfaces with customers and pickers.
    • Allows customers to scan waste for reuse tips using an Object Detection Model.
    • Sends pickup requests to the primary server via REST APIs.

  3. Admin WebApp

    • Admins can monitor and manage pickup requests.
    • Interacts with the primary server using REST APIs.

  4. Primary Server

    • Central hub for processing requests and managing data.
    • Interfaces with the Primary Database and various microservices.
    • Use Django and Django Rest Framework
    • Facilitates REST and gRPC communications.

  5. Clustering Microservice

    • Node.js server that clusters pickup requests for optimized collection routes.
    • Communicates with the primary server via gRPC.
    • Stores cluster data in the Cluster Database.

  6. Prompt Microservice

    • Generates reuse tips for scanned waste using the Llama LLM (Language Learning Model).
    • Utilizes the Object Detection Model to analyze waste images.

  7. FastAPI Server

    • Fetches data from the Llama model and stores it in the database for quicker access to similar queries.
    • Ensures efficient retrieval and caching of frequently requested data.

Technologies Used

  • Mobile App: Flutter
  • Admin WebApp: ReactJs
  • Primary Server: Django
  • Clustering Microservice: Node.js server with gRPC communication
  • Primary Database: PostgreSQL
  • Cluster Database: MongoDB
  • Object Detection Model: Machine Learning model for waste identification
  • Llama LLM: Natural Language Processing model for generating reuse tips
  • FastAPI Server: FastAPI for efficient data retrieval and caching

Workflow

  1. Customer Interaction:

    • The customer scans the waste using the mobile app.
    • The Object Detection Model identifies the waste and suggests reuse tips via the Prompt Microservice.
    • The customer requests a pickup through the app.

  2. Pickup Request Handling:

    • The pickup request is sent to the primary server and recorded in the Primary Database.
    • Admins can view and manage requests through the Admin WebApp.

  3. Clustering and Assignment:

    • The Clustering Microservice groups pickup requests based on location.
    • Drivers are automatically assigned to clusters using location data.

  4. Driver Notification and Pickup:

    • The assigned driver receives a notification of the pickup request.
    • The driver completes the pickup and marks it as completed in the system.

  5. Reuse Tips Generation and Caching:

    • The Prompt Microservice interacts with the Llama LLM to generate reuse tips.
    • The FastAPI Server fetches these tips and stores them in the database for quicker access to similar queries.

Installation and Setup

  1. Clone the repository: 
    git clone https://github.com/libreprojectss/waste_wise.git
cd waste-wise
  2. Setup main django server 
    cd main_server/ecycle
python manage.py migrate
python manage.py runserver
  3. Setup admin server (clustering microservice) 
    cd admin_server/
yarn install
yarn dev
  4. Setup prompt server (prompt microservice) 
    cd admin_server/services
python main.py
  5. Run the react admin dashboard 
    cd admin_web
yarn install
yarn dev

The setup can also be done using docker.

docker-compose build
docker-compose up