RoadWatch

AI-Powered Civic Accountability Platform for the BIMSTEC Region

RoadWatch empowers everyday citizens to seamlessly report road infrastructure defects using a conversational AI interface. By simply uploading a photo and a GPS location, our pipeline automatically categorizes the issue, calculates severity using Computer Vision, maps it to the precise civic road segment, and holds the assigned engineers accountable.

---

Features

• Conversational AI Agent: Report potholes and check road quality using natural, friendly language.
• Computer Vision (YOLOv8): Upload a photo of a road defect; the ML engine instantly categorizes the issue and calculates severity.
• Spatial Routing (Uber H3): Instantly matches your raw GPS coordinates to specific civic road segments and automatically flags the responsible engineer.
• Spending Dashboard: View localized budgets, contractor histories, and civic road maintenance statuses with a single click.

Tech Stack

• Frontend: React 18, Vite, TailwindCSS
• Backend API: FastAPI (Python 3.10+)
• ML Engine: Ultralytics YOLOv8 (Standalone FastAPI microservice)
• Database: Supabase (PostgreSQL with PostGIS)
• AI Brain: Groq (Llama 3.3) for intent classification
• DevOps: Dockerized with Docker Compose & Nginx

---

Environment Variables

Before running the application, you must create an apps/backend/.env file. You can use the provided .env.example as a template (if available) or create it manually with the following required keys:

# Supabase Configuration
SUPABASE_URL=https://<your-project>.supabase.co
SUPABASE_KEY=<your-anon-key>
SUPABASE_SERVICE_ROLE_KEY=<your-service-role-key>

# Groq LLM Configuration
GROQ_API_KEY=<your-groq-api-key>
GROQ_MODEL=llama-3.3-70b-versatile

(Note: If using Docker Compose, the CV_SERVICE_URL and HOST variables are injected automatically.)

---

Quick Start (Docker)

The easiest way to run the entire RoadWatch stack (Frontend, Backend, and ML Microservice) is using Docker Compose.

1. Clone the repository:

git clone https://github.com/krishnagoyal099/Roadwatch.git
cd Roadwatch

2. Launch the entire civic platform: Make sure you have Docker installed, then run:

docker-compose up --build

The services will spin up and bind to these ports:

• Frontend Web App: http://localhost:5173
• Backend API: http://localhost:8000
• ML Service: http://localhost:8001

---

Manual Setup (For Development)

If you prefer to run the services individually for active development:

1. Start the ML Service:

cd apps/ml
pip install -r requirements.txt
uvicorn api.analyse_image:app --port 8001

2. Start the Backend API:

cd apps/backend
pip install -r requirements.txt
uvicorn app.main:app --port 8000

3. Start the Frontend:

cd apps/frontend
npm install
npm run dev

---

Documentation

Dive deeper into the architecture and system design:

• System Architecture - Learn how the microservices communicate.
• API Flow - Sequence diagrams for the Chatbot and CV engine.
• H3 Spatial Routing - How we match coordinates to road segments.
• Supabase Setup - Database schema and row-level security.