Data Processing Deployment

This repo is to build a basic deployment app for Machine Learning process developed for analysing membership data.

Environment

• fastapi 0.63.0
• gunicorn 20.0.4
• Jinja2 2.11.2
• pandas 1.2.4
• seaborn 0.11.0
• sklearn 0.23.2
• requests 2.22.0
• uvicorn 0.13.3
• wbgapi 1.0.5

complete list canbe found in requirements.txt

Usage

There are two possible ways to use this app

1. Run Locally
   • To run the app locally create a clone of the app. Then change the port number in app.py to any other than 80 (default port for internet access).
   • The default port is set to 80. If running locally change that to any port of your choice other than 80
   • Install requirements by using "pip install requirements.txt"
   • Start app by typing "python app.py" in terminal or command shell.
2. Run Using Docker It is also possible to create a docker container.
   • Just Open the folder in terminal or command shell.
   • Use the command docker build -t <app_name> .
   • Once the app is built use the command docker run --rm --name <app_name_running> -p {local_port}:{docker_image_port} -ti <container_name>

Deployment

Before deplyment make sure to change the upload folder to a storage bucket. You can pass the storage bucket address as ENV variable of a Docker container.

The current set up would save any uploaded file to the upload folder in the Docker container which can be accessed manually but that is not a good practice.

(GCP instructions can be found here https://cloud.google.com/run/docs/quickstarts/build-and-deploy/python)