👚 ZEROFIT_NODEJS_SERVER

This project serves as the backend for providing advanced wardrobe management and virtual fitting solutions powered by Google Cloud Platform (GCP) and cutting-edge AI models (SAM2, KOLORS).

Users can seamlessly organize their wardrobe, engage in buying and selling clothing items through a dedicated marketplace, and experience virtual fitting with personalized style recommendations generated by AI.

👥 ZeroFit contributors

Name	Roles
Jo Mal-Geum	Team Leader / Planning and Design
Lim Sun-Jong	AI
Kwon Tae-Eun	Frontend
Kim Kyung-Mo	Frontend
Kim Jun-Ho	Backend

📺 Demo Video

Experience the functionality of this project by watching the demo video:
Watch the Demo

🚀 Start

1. Requirement

This project was carried out in the following environment.

• Node.js v23.1.0
• npm v10.9.0
• MySQL v8.0
• UTF-8 (includes some Korean content)
• Google Cloud Platform (GCP) account

2. Install

Copy your local environment

1. Clone this repository

   git clone https://github.com/your-username/ZEROFIT_NodeJS_Server.git
   cd ZEROFIT_NodeJS_Server

2. Install dependancy

   npm install

3. Set environmental variables

   COOKIE_SECRET = your-cookie-secret
   
   # Connect to GCP MySQL - PROD
   SEQUELIZE_USERNAME = your-db-username
   SEQUELIZE_PASSWORD = your-db-password
   SEQUELIZE_DB_PROD = your-db-name
   
   SEQUELIZE_HOST = your-cloud-sql-host
   SEQUELIZE_PORT = your-cloud-sql-port
   
   # Connect to local MySQL - DEV
   SEQUELIZE_DEV_USERNAME = your-db-username
   SEQUELIZE_DEV_PASSWORD = your-db-password
   SEQUELIZE_DEV_DB = your-db-name
   
   SEQUELIZE_DEV_HOST = your-cloud-sql-host
   SEQUELIZE_DEV_PORT = your-cloud-sql-port
   
   # Connect to local MySQL - TEST 
   SEQUELIZE_TEST_USERNAME = your-db-username
   SEQUELIZE_TEST_PASSWORD = your-db-password
   SEQUELIZE_TEST_DB = your-db-name
   
   SEQUELIZE_TEST_HOST =  your-cloud-sql-host
   SEQUELIZE_TEST_PORT = your-cloud-sql-port
   TEST_SERVER_URL = your-cloud-sql-url
   
   # FLASK SAM
   FLASK_SAM_HOST = your-flask-sam-host
   FLASK_SAM_PORT = your-flask-sam-port
   
   # FLASK KOLORS
   FLASK_KOLORS_HOST = your-flask-kolors-host
   FLASK_KOLORS_PORT = your-flask-kolors-port
   
   # JWT SECRET
   JWT_SECRET = your-jwt-secret
   JWT_EXPIRES_IN= your-jwt-expires-in

4. Run production environment of Node.js server

   npm run prod

5. Monit Node.js with command below

   npx pm2 monit

---

📡 How to Use API

This project provides a comprehensive API for seamless integration. For detailed instructions and examples, please refer to the API Usage Guide.

🛠️ CI/CD pipeline

GCP Cloud Build & GitHub Actions

For GCP Cloud Build, you must follow these steps.

1. Open cloudbuild.yaml

2. 'your-*' part needs to be modified

   steps:
   # 1. Install dependancy
   - name: 'gcr.io/cloud-builders/npm'
       args: ['install']
   
   # 2. Update your repository using 'git pull'
   - name: 'gcr.io/cloud-builders/gcloud'
      args:
         - 'compute'
         - 'ssh'
         - 'your-vm-instance'
         - '--zone'
         - 'your-vm-instance-timezone'
         - '--command'
         - |
           git config --global --add safe.directory /home/your-github-account/ZEROFIT_NodeJS_Server && \
           cd /home/your-github-account/ZEROFIT_NodeJS_Server && git pull
   
   # 3. Restart your server using pm2
   - name: 'gcr.io/cloud-builders/gcloud'
      args:
         - 'compute'
         - 'ssh'
         - 'your-vm-instance'
         - '--zone'
         - 'your-vm-instance-timezone'
         - '--command'
         - |
           cd /home/your-github-account/ZEROFIT_NodeJS_Server && \
           npx pm2 reload all
   
   # logsBucket
   logsBucket: 'gs://your-bucket-name' 

🗂️ Project architecture

1. Client request

   • When a user sends a request, it is routed to the Express.js server through the GCP Load Balancer.

2. GCP Express.js server

   • The Express.js server handles CRUD operations by communicating with MySQL Cloud SQL.
   • Image processing requests are forwarded to the GPU-powered Flask server.

3. Inha University Artificial Intelligence Convergence Research Center Flask server

   • The Flask preprocessing server utilizes the SAM2 model to generate preprocessed garment data.
   • The Flask virtual fitting server processes API requests from the Express.js server and interacts with the KlingAI server to generate the final virtual try-on image. This image is then sent back to the GCP Node.js server for storage and management and ultimately delivered to the Flutter application.

4. Note: What is GCP Load Balancer?

   • A load balancer distributes incoming client requests across multiple server instances to ensure optimal performance and reliability.
   • Our current setup is configured to limit each backend instance's utilization to a maximum of 80%.
   • If the utilization exceeds 80%, GCP's autoscaling feature automatically provisions a new server instance to handle the additional load.

⚙️ Tech Stacks

• Backend: Node.js, Express.js, Flask
• Database: Cloud SQL (MySQL)
• Cloud Provider: Google Cloud Platform (GCP)
  • GCP VM Instances (with Auto-scaling)
  • GCP Cloud Load Balancer
  • GCP Cloud Build (for CI/CD)
• AI Models: SAM2, KOLORS
• GPU Server: NVIDIA A40 (Inha University AI Convergence Research Center)
• Code Formatting: Clang-format
• Style Guide: Google C++ Style Guide (adapted where applicable)

🧑‍💻 How to contribute

1. Fork this repository
2. Create a new branch

   git checkout -b feature/your-new-feature-name your-remote-name/feature/your-new-feature-name

3. Commit your change logs

   git commit -m "feat: add your-change-logs"

4. Push your branch

   git push your-remote-name feature/your-new-feature-name

5. Create a pull request at github

📄 Licesne

This project is licensed under the MIT License. For more details, please refer to the LICENSE file.

📝 Questions or Support

If you have any questions or need support, feel free to open an issue on GitHub or reach out via the following contact methods:

• Email: logicallawbio@gmail.com
• GitHub: logicallaw