WebMLOpenEO - Web-based Machine Learning for OpenEO

About this Project

This project was created for the course "Geosoftware II" at the University of Münster, Institute for Geoinformatics. The task was to create a web based application building upon the openeocubes-Framework. It should enable researchers and remote sensing experts to perform machine-learning based Land-Use-Land-Cover (LULC) classifications.

In our application we provide a web-app with an adjoined Express-API to expose functionality and handle requests. Computations are performed on a third container runnning the framework openeocubes.

Installation Guide

Prerequisites

This respository uses a submodule, so some extra steps may be necessary.

It can be cloned with git clone --recurse-submodules https://github.com/wittrockscode/WebMLOpenEO.git. This ensures that also the newest version of https://github.com/TimCi/openeocubes is fetched.

If this repository was cloned without the recurse-submodules flag, use git submodule update --init --recursive to update the state of the submodule.

For further information please refer to the git documentation.

Local deployment with docker

Requirements

Docker and docker compose have to be installed on your system. The application operates on the ports 80, 3000 and 8000 locally, which have to be free while using it.

Starting

To start the whole application navigate to the project directory and use:

docker compose up -d --build

Local deployment without docker

Please refer to the individual application segments to start them locally without docker:

• Frontend

• Node-Backend

• R-Backend

Remote deployment

Requirements

Docker and docker compose have to be installed on the machine.

Configuration

The frontend depends on some environment variables which hold relevant URLs of the application:

• Navigate to ./app/frontend.

• Open the file .env with a text editor (for example nano).

• Replace the VITE_NODE_BACKEND_URI value with http://your_server_url.com/api (note the /api).

• Replace the VITE_BASE_URL value with http://your_server_url.com.

• Replace the VITE_ENV value with production.

• Finally, save the file.

We have provided a sample file with the data for the AWS instance used while developing this application, it can be found here.

Starting

Execute the docker compose up -d --build command in the terminal.

Testing

Please also refer to the individual application segments to get information about testing.