Prototype ML

Machine Learning For Everyone

Django app to expose interface of scikit-learn through API

Update : Refactored code to dynamically fetch model classes mentioned by the user in API. Theoretically, all models in scikit learn can be tested now.

Features

• Independent login for users
• Dashboard for users to manage models
• Train and save models through API
• Run predictions through API

Installation

git clone https://github.com/ramansah/ml_webapp.git  

Configure credentials for MySQL client at ~/mysql.cnf

[client]  
database = ml_webapp  
user = username  
password = ****  
default-character-set = utf8  

Install mysqlclient-python

https://github.com/PyMySQL/mysqlclient-python

Install MongoDB

https://www.digitalocean.com/community/tutorials/how-to-install-mongodb-on-ubuntu-16-04

Create virtual environment and run locally

python -m venv myenv  
source myenv/bin/activate  
  
cd ml_webapp  
  
pip install --upgrade pip  
pip install -r requirements.txt  
  
python manage.py makemigrations  
python manage.py migrate  
python manage.py runserver  

Usage

Visit http://localhost:8000 and register a new user

Fetch the JWT for current user

POST /api/login/
Content-Type: application/json
{
  "username": "username",
  "password": "password"
}

Response
{
  "token": "abcd12345"
}

Create a model and save in the DB

Consider the

POST /api/model/
Content-Type: application/json
Accept: application/json
Authorization: JWT abcd12345

{
  "model_path": "sklearn.linear_model.LinearRegression",
  "action": "new_model",
  "name": "Compute Final Score",
  "input_x": [[95, 87, 69], [99, 48, 54], [85, 57, 98], [90, 95, 91]],
  "input_y": [291, 200, 254, 326]
}

Response
{
  "status": "Trained",
  "model_id": "randommodelid"
}

Use this model to predict your score

POST /api/model/
Content-Type: application/json
Accept: application/json
Authorization: JWT abcd12345

{
  "action": "predict",
  "model_id": "randommodelid",
  "input_x": [[90, 95, 91]]
}

Response
{
  "status": "OK",
  "prediction": [
      326
  ]
}

Check out your trained models at Dashboard