Swayog

Lightweight pose-correction app that runs a Python CV/ML backend (FastAPI), a React + Vite frontend, and an optional Spring Boot security backend. Includes live webcam pose analysis and voice feedback for corrective guidance.

Project Structure

• api/ — FastAPI backend that processes webcam frames, runs pose checkers from logic/, and exposes a WebSocket at /ws/{client_id}.
• frontend/ — React + Vite single-page app (TypeScript) used for live camera feed and UI (src/pages/Practice.tsx handles the WebSocket client).
• logic/ — Pose-checker implementations and voice_feedback.py utilities used by the backend (legacy TTS available here).
• Poses/ — Landmark/pose data (e.g. yoga_pose_landmarks.json).
• sec/ — Spring Boot project (optional) providing auth/MFA and APIs used by the frontend.

Quick Start (Windows PowerShell)

1. Open PowerShell in the repository root (d:\YogaFix).

2. Create and activate a Python virtual environment (recommended name: .venv):

python -m venv .venv
# Activate for this PowerShell session
.\.venv\Scripts\Activate.ps1

If PowerShell blocks the activation script because of execution policy, run (temporary for the session):

Set-ExecutionPolicy -ExecutionPolicy RemoteSigned -Scope Process
.\.venv\Scripts\Activate.ps1

3. Install Python dependencies:

pip install -r requirements.txt

4. Start the FastAPI backend (from d:\YogaFix\api):

cd api
# dev: use uvicorn for autoreload
py -m uvicorn main:app --reload --host 0.0.0.0 --port 8000

5. Start the frontend (from d:\YogaFix\frontend):

cd frontend
npm install
npm run dev

6. (Optional) Start the Spring Boot server (from d:\YogaFix\sec):

cd sec
./mvnw.cmd spring-boot:run

License & Contributing

This repository is provided as-is. Feel free to open issues or contribute improvements via pull requests.