Skip to content

README.md

Latest commit

 

History

History
128 lines (78 loc) · 3.82 KB

README.md

File metadata and controls

128 lines (78 loc) · 3.82 KB

🌸 FlowerPredictor: MLOps Major Project

Welcome to MLOps Project 6. This repository demonstrates a production-grade machine learning pipeline. The core focus of this project is mastering CircleCI for cloud-native automation, replacing traditional tools like Jenkins to deploy a containerized application to Google Kubernetes Engine (GKE).

📖 Introduction

This project marks a major milestone in our MLOps journey. While previous projects focused on complex data engineering (like the 200,000-row Australia Weather dataset), this project uses a streamlined dataset to prioritize Deployment Infrastructure.

The goal is to move from "learning to drive in a parking lot" (local testing) to "driving a massive truck on the highway" (scalable cloud deployment).

🛠️ The Technology Stack

Tool Icon Purpose
CircleCI ⭕ Cloud-hosted automation for building and testing code on every push.
GitLab CI 🦊 Manages the step-by-step "pipeline" journey from laptop to server.
GitHub Actions 🐙 Automates tasks like unit testing directly within the repository.
Docker 🐳 Containerizes the application for consistent environments.
GCP (GKE) ☁️ Orchestrates the deployment using Google Kubernetes Engine.

🏗️ Project Workflow

The project follows a modular, step-by-step professional engineering cycle:

  1. Project Setup: Environment configuration () and directory structuring (src, pipelines, artifacts).
  2. Jupyter Testing: Prototyping data processing and model logic in blocks.
  3. Modularity: Refactoring code into Classes and Objects for production readiness.
  4. UI Development: Building a user-friendly interface using Flask and HTML.
  5. GCP Infrastructure: * Setting up GKE Clusters.
  • Configuring Artifact Registry (GCR).
  • Managing IAM service account JSON keys.
  1. CI/CD Pipeline: Automating the flow from GitHub → CircleCI → GCR → GKE.

🔄 How CircleCI Works (The Connection)

The moment you push code to the main branch, the following automated sequence triggers:

  1. Trigger: CircleCI detects the update via GitHub SCM.
  2. Checkout: CircleCI pulls the latest code into a cloud container.
  3. Build & Push: A Docker image is created and pushed to Google Container Registry (GCR).
  4. Deploy: CircleCI communicates with GKE to roll out the new image to the Kubernetes cluster.

🥊 Comparison: CircleCI vs. Jenkins

Why choose CircleCI for this project?

Feature ⭕ CircleCI 🦊 Jenkins
Hosting Cloud-Based (SaaS) Local/Self-Managed
Setup Fast (Mostly UI clicks) Complex (Manual installation)
Scalability Native & Automated Manual Scaling
Customization Standardized Infinite (3,000+ plugins)
Typical User Startups (Spotify, Pinterest) Big Tech (Netflix, Uber)

🚀 Getting Started

Prerequisites

  • Google Cloud Platform Account
  • CircleCI Account linked to GitHub
  • Python 3.8+

Installation

  1. Clone the repo:
git clone https://github.com/Mahul777/FlowerPredictor.git
  1. Install Dependencies:
pip install -r requirements.txt
  1. Run Locally:
python app.py

📜 License

Distributed under the MIT License.

📧 Contact

Mahul - Your GitHub Profile

Project Link: https://github.com/Mahul777/FlowerPredictor

Would you like me to help you write the specific config.yml file for CircleCI to match this workflow?

## Pipeline
![alt text](image.png)

## UI 
![alt text](image-1.png)

## Clustering
![alt text](image-2.png)