DiabetesPredict

Enviroment

• Ubuntu20.04
• Python3.8
• Firebase
• Windows10

Introduction

We use Python PyQt5 library to to build a GUI plaform to predict users diabetes status from a pretrained multi-class classification model. This model is trainned by using Keras library with the diatabetes data. Users need to register and login in the GUI platform. The authentication is provided by firebase authentication system. Users input their the diabetes status, and they are stored in firebase realdatabase. Users could export their diabetes history into csv. The firebase administration account could also edit their users information. Finally, we use pyinstaller to make the GUI library to be executed in windows environment.

Installation

• Download repository to your-path

Data Setting

• Sign in Kaggle and save pima-indians-diabetes.csv to the folder your-path/repository/GUI

Firebase Setting

• Create a json file to path your-path/repository/GUI/config.json

• Signin firebase and create project your-project

• create admin account "admin@gmail.com" in Authentication page (you may also want to create other user accounts)

• Go to realtime database rule page and set the rule

{
    "rules": 
    {
        "users": 
        {
            ".read": "auth.uid == '<admin-account-user-id>'",
            ".write": "auth.uid == '<admin-account-user-id>'",
                "$uid": 
                {
                // Allow only authenticated content owners access to their data
                ".read": "auth.uid == $uid",
                ".write": "auth.uid == $uid"
                }
        },
    }
}

• Go to project setting, create an application. Copy and paste the firebaseConfig from your application config to your config.json

• config.json

{
    "apiKey": "<your-apiKey>"
    "authDomain": "<your-authDomain>"
    "databaseURL": "<your-databaseURL>"
    ...
}

Run locally in Ubuntu

• In Ubuntu

• Install PyQt5 tools

$ sudo apt-get install qt5-default
$ sudo apt-get install qttools5-dev-tools

• Go to repository directory

$ cd <your-path>/<repository>

• Run virtualenv and install library in virtualenv

$ virtualenv venv

$ source ./venv/bin/activate

$ pip3 install -r requirements.txt

$ cd ./GUI

$ python3 main.py

• Quit virtualenv

$ deactivate

Exports to exe in Windows

• In Windows OS

• Download repository to your-path

• Go to repository directory

cd <your-path>/<repository>

• (Make sure BOTH pima-indians-diabetes.csv and config.json in your-path/repository/GUI)

• Create virtual environment and enter the virtual environment

virtualenv venv

.\venv\Scripts\activate.bat

• Install dependencies and export python file to EXE

pip3 install -r requirements.txt

pyinstaller main.spec

• Move main.exe to GUI directory

mv dist/main.exe GUI/main.exe

• Double click GUI/main.exe for run the program

Program tutorial

• login up and sign up platform

• create more auth user in firebase

• login in with admin account "admin@gmail.com"

• admin could view, add, edit, save any user info

• input the data info in add or edit page to predict diabetes. Please reference pima-indians-diabetes.csv

• users diabetes info is saved in GUI directory as output.csv

Reference

• Python - PyQt5
• Qt Designer
• Code First with Hala
• pima-indians-diabetes.csv