This project aims to industrialize a Kaggle notebook from scratch.

A machine learning python notebook with training done with Tensorflow/Keras package is processed. The model is saved and used in an API that can be requested to perform ML predictions

This project includes following steps:

1. Find a notebook corresponding to the criteria requested on Kaggle

We have chosen the notebook in the following link: https://www.kaggle.com/michalbrezk/x-ray-pneumonia-cnn-tensorflow-2-0-keras-94
The corresponding dataset is in the following link: https://www.kaggle.com/paultimothymooney/chest-xray-pneumonia
We save the trained model to be able to load it into the API

2. Develop an API in Flask to serve the model

Using PyCharm and a virtual environment for development Coding to load the model at API startup Creating a route for inference Use a WSGI HTTP web server (Gunicorn)

3. Add Swagger documentation to the API

4. Contain your API with Docker :

Use docker-composite to define the complete stack

Step to run the project

• Clone the repo
• Install Docker and Docker-compose component
• Build new image by typing this in terminal :
• On linux :$ run_docker.sh
• On Windows : > run_docker_win.bat
• To test the API, using Swagger :
  • On a browser (Chrome or Firefox) open https://editor.swagger.io/
  • Click on File, then Import url
  • Insert this adress : http://localhost:8000/static/openapi.json
  • Play with the different routes.