🔋 Battery Anomaly Detector

An end-to-end ML/DL + MLOps project for detecting anomalies in battery behavior using LSTM Autoencoders.
This project demonstrates how to go from model training → API service → Streamlit UI → CI/CD deployment on AWS EC2 🚀

---

✨ Features

• 🧠 LSTM Autoencoder for anomaly detection on battery cycles
• 📊 Preprocessing pipeline (scaling + sequence creation)
• ⚡ REST API using FastAPI
• 🎨 Interactive Streamlit UI
• 🐳 Dockerized setup for portability
• 🔄 Automated CI/CD with GitHub Actions → builds & deploys to AWS EC2
• 🔑 Secrets managed via GitHub Secrets (secure deployment)

---

🏗️ Project Structure

  Battery-Anamoly-Detector/
  │── app/
  │ ├── fastapi_app.py # FastAPI backend (model inference API)
  │ ├── streamlit_app.py # Streamlit frontend (UI)
  │
  │── src/
  │ ├── model.py # LSTM Autoencoder definition
  │ ├── train.py # Training pipeline
  │ ├── utils.py # Helper functions
  │ ├── config.py # Configs (sequence length, paths, features, etc.)
  │
  │── artifacts/
  │ ├── best_model.pth # Saved trained model
  │ ├── scaler.pkl # Pre-fitted scaler
  │
  │── Dockerfile # Docker build for API + Streamlit
  │── requirements.txt # Dependencies
  │── .github/workflows/cd.yml # CI/CD pipeline
  │── README.md # Project documentation

---

⚙️ Workflow

1️⃣ Training

• Run train.py to train the model on battery data.
• Saves best_model.pth & scaler.pkl in artifacts/.

2️⃣ API Service (FastAPI)

• Exposes endpoints for anomaly detection.
• Example: http://<EC2-IP>:8000/docs

3️⃣ Streamlit UI

• Interactive frontend for predictions.
• Example: http://<EC2-IP>:8501

4️⃣ Dockerization

• Single container runs both FastAPI & Streamlit.
• Lightweight & portable deployment.

5️⃣ CI/CD Deployment

• On push to GitHub:
  • 🏗️ Build Docker image on GitHub runner
  • 📦 Push image → DockerHub
  • ☁️ Connect to AWS EC2 & deploy container
  • 🔄 Stops old container & runs latest image

---

🚀 Deployment

🔑 GitHub Secrets Required

• DOCKERHUB_USERNAME → your DokcerHUB username
• DOCKERHUB_TOKEN → your login token , generated from DockerHub
• EC2_HOST → your instance IP
• EC2_USER → usually ubuntu or else the Cloud OS you choosed
• EC2_KEY → your private SSH/RSA key for connecting this github to cloud machine

🌍 Access After Deployment

• FastAPI → http://<EC2-IP>:8000/docs
• Streamlit → http://<EC2-IP>:8501

---

📈 Tech Stack

• Python 🐍
• PyTorch 🔥
• FastAPI ⚡
• Streamlit 🎨
• Docker 🐳
• GitHub Actions 🤖
• AWS EC2 ☁️

---

🤝 Contributions

Pull requests are welcome! 🎉
For major changes, open an issue first to discuss what you’d like to change.

---

📜 License

MIT License © 2025 Milan Kumar Singh