Image Prediction API Backend for MaggotIn App 🐛

This project implements a backend API for image prediction of maggots. The application uses Flask for handling web requests and TensorFlow for the prediction model.

Cloud Computing Team

Student ID	Name	University
C296B4KY2742	Muhammad Arizaldi Eka Prasetya	Universitas Pembangunan Nasional "Veteran" Jawa Timur
C296B4KX0841	Berlian Viga Septiani	Universitas Pembangunan Nasional "Veteran" Jawa Timur

APIs Services

• Predict Maggot Image

Google Cloud Platform Services

• Cloud SQL with database engine MySQL
• Cloud Storage
• Cloud Run
• Artifact Registry
• Secret Manager

Project Structure

PROJECT ROOT
├── app            
│   ├── server
│   │   ├── handler.py  
│   │   └── routes.py    
│   └── services
│       ├── inferenceServices.py       
│       ├── initModel.py    
│   	├── loadModel.py      
│      	└── storeData.py                                
├── app.py
├── myenv
├── .env                           
├── .gitignore                     
├── env.example                                  
└── requirements.txt                   

README.md

Project Setup

If Python is not yet installed, you can download and install it from official python website

Install

# check python version
$ python --version

# install virtual environment
$ pip install virtualenv

# create virtual environment
$ virtualenv myenv

# for windows, activate use
$ myenv\Scripts\activate

# install flask
$ pip install flask

Clone Project

https://github.com/Capstone-MaggotIn/Cloud-Computing-2.git

Install Project Dependencies

$ pip install -r requirements.txt

Run the Application

• Edit .env files and specify with your env.examples
• Run the app

$ python app.py

Deployment

To set up and deploy the backend model application, first clone the source code from the specified GitHub repository and navigate to the project directory. Create a Dockerfile to define the container configuration details can be found at Dockerfile. Ensure that the required Google Cloud APIs, including Artifact Registry, Cloud Build, and Cloud Run, are enabled. Next, create an Artifact Registry repository to store the Docker image. Build the container image using the Dockerfile, then upload it to Artifact Registry using the gcloud builds submit command. Finally, deploy the application to Cloud Run using the created container image for scalable and serverless hosting with the gcloud run deploy command.

Endpoint

POST /predict

Request

• Endpoint : /predict
• Method : POST
• Example Request

Content-Type: multipart/form-data

image: [maggot_image.jpg]

Response

• Status Code : 200
• Content Type : application/json
• Result Response

{
  "status": "true",
  "message": "Model is predicted successfully",
  "data": {
    "id": "e326e8e5-ce70-46e9-8369-f3dc6fd43740",
    "result": 5,
    "phase": "Prepupa",
    "confidenceScore": 76.68556213378906,
    "created_at": "2024-12-08T17:35:04.827625"
}

Dependency

• Flask
• TensorFlow
• Flask-Cors
• Python-DotEnv
• Mysql-Connector

Resources

• Planning costs via Google Cloud Pricing Calculator
• Cloud Architecture
• Visit the Flask Documentation to learn more about the framework

License

This repository's source code is available under MIT License.