31_Deployment on Cloud Platform.md

<< Day 30 |

🎉 Bonus Day 31: Deployment on Cloud Platform

Welcome to Bonus Day 31 of the 30 Days of Data Science series! 🎉 Today, we’ll explore how to deploy your machine learning models or applications to Cloud Platforms using Flask/FastAPI. By the end, you'll understand the steps to deploy to AWS, Azure, and GCP.

📜 Table of Contents

• 🎉 Bonus Day 31: Deployment on Cloud Platform
  • 📜 Table of Contents
  • 🌐 Introduction
  • 🚀 Preparing the Application for Deployment
  • ⚙️ Deploying with Flask/FastAPI
    • Flask Example
    • FastAPI Example
  • ☁️ Deployment on AWS
    • Steps to Deploy
  • ☁️ Deployment on Azure
    • Steps to Deploy
  • ☁️ Deployment on GCP
    • Steps to Deploy
  • 📝 Practice Exercise
  • 📚 Summary

🌐 Introduction

Deployment is a crucial step in bringing your data science project to life. It allows others to interact with your model or application in real-time. We'll focus on deploying applications using Flask or FastAPI to popular cloud platforms like:

• AWS (Amazon Web Services)
• Azure
• GCP (Google Cloud Platform)

🚀 Preparing the Application for Deployment

Basic Folder Structure

Ensure your project folder is structured properly:

project/
|-- app.py  # Main application script
|-- model.pkl  # Serialized ML model (if applicable)
|-- templates/
|    |-- index.html  # Frontend files (if needed)
|-- requirements.txt  # Python dependencies
|-- Dockerfile  # (Optional) Docker configuration

Creating requirements.txt

List all dependencies in a requirements.txt file:

Flask==2.1.2
pandas==1.5.3
numpy==1.23.5
scikit-learn==1.1.3

Generate automatically:

pip freeze > requirements.txt

⚙️ Deploying with Flask/FastAPI

Flask Example

Here’s a minimal Flask application:

from flask import Flask, request, jsonify

app = Flask(__name__)

@app.route("/", methods=["GET"])
def home():
    return "Welcome to the Flask Deployment!"

@app.route("/predict", methods=["POST"])
def predict():
    data = request.json
    prediction = sum(data["values"])  # Example prediction logic
    return jsonify({"prediction": prediction})

if __name__ == "__main__":
    app.run(debug=True)

Run locally:

python app.py

FastAPI Example

Here’s a minimal FastAPI application:

from fastapi import FastAPI
from pydantic import BaseModel

app = FastAPI()

class InputData(BaseModel):
    values: list[int]

@app.get("/")
def home():
    return {"message": "Welcome to FastAPI Deployment!"}

@app.post("/predict")
def predict(data: InputData):
    prediction = sum(data.values)  # Example prediction logic
    return {"prediction": prediction}

if __name__ == "__main__":
    import uvicorn
    uvicorn.run(app, host="0.0.0.0", port=8000)

Run locally:

uvicorn app:app --reload

☁️ Deployment on AWS

Steps to Deploy

1. Set Up an EC2 Instance:

   • Go to the AWS Management Console.
   • Launch an EC2 instance with an Ubuntu AMI.

2. Install Dependencies on EC2: SSH into the instance and set up the environment:

   sudo apt update && sudo apt upgrade
   sudo apt install python3-pip
   pip3 install -r requirements.txt

3. Run the Application:

   python3 app.py

4. Expose the Application:

   • Open port 5000 (or your application port) in the AWS security group.
   • Access the app using the public IP of the EC2 instance.

☁️ Deployment on Azure

Steps to Deploy

1. Set Up an App Service:

   • Go to the Azure Portal.
   • Create an App Service and select a Python runtime.

2. Deploy Code:

   • Zip your project folder and upload it through the Azure portal.

3. Configure Startup Command: Add the following startup command in the portal:

   gunicorn -w 4 -k uvicorn.workers.UvicornWorker app:app

4. Access the Application:

   • Use the App Service URL provided by Azure.

☁️ Deployment on GCP

Steps to Deploy

1. Set Up a Google Cloud Project:

   • Go to the GCP Console.
   • Enable the App Engine API.

2. Create an app.yaml File:

   runtime: python39
   entrypoint: gunicorn -w 4 -k uvicorn.workers.UvicornWorker app:app

3. Deploy the Application:

   gcloud app deploy

4. Access the Application:

   • Use the provided GCP URL.

📝 Practice Exercise

1. Create a Flask/FastAPI application that:

   • Accepts an input text.
   • Returns the sentiment (positive/negative) using a pre-trained model.

2. Deploy this application to any cloud platform of your choice.

📚 Summary

In this bonus session, we learned how to:

• Prepare a Flask/FastAPI application for deployment.
• Deploy to AWS EC2, Azure App Service, and Google Cloud Platform.
• Configure cloud platforms and expose services to the internet.

🎉 Congratulations on completing the 30 Days of Data Science series with this bonus day! You’re now equipped to deploy your projects to the world. 🌎

---