Implementing the algorithm

Docker container

We provide a docker container based on an Ubuntu hosting an R-Studio Server. To run it you need to have docker installed on your Server/Device where you want to run the analysis.
Follow this guide (https://docs.docker.com/engine/install/) to install it. The required amount of memory is dependent on the number of patients that you want to analyze. We recommend using a system with a sufficient number of cores 16+ and memory (64+ GB). Due to the dynamic chunking option, it is possible to run it on a machine with fewer cores and memory, but it will be slower.
When you want to compute your own correlation matrices for the exclusion, use at least double the amount but we recommend using a system with around 64+ cores and 256+ GB of memory.

Download & import the container

The container is available via the GitHub Container registry. You can pull it with the follow command from the commandline: docker pull ghcr.io/clai-group/post_covid_ai_scripts:2024-04-04

Start the docker container

Navigate to your nearly created directory in the terminal and execute the following command to start the container(replace everything in < >):
docker run --rm -ti -e DISABLE_AUTH=true -e PASSWORD=<YOUR_PASS> -e RSTUDIO_WHICH_R=/usr/local/bin/R -p 127.0.0.1:8787:8787 -v <.path/to/data>:/home/rstudio/data -v <.path/to/output>:/home/rstudio/output ghcr.io/clai-group/post_covid_ai_scripts:2024-04-04

(when running on the server use -p 8787:8787 instead, since the command above makes the web interface only accessible from the device where the container is running;)

If you are using a MAC with an ARM chip(no intel) enable rosetta 2 by executing the following command from the command line:
softwareupdate --install-rosetta

Depending on your setup you might want to use a docker-compose file to orchestrate your setup.
Access the container via the web browser $IP:8787, where $IP is localhost if docker is running on your own device or the server ip address when running it on a server and accessing it remotely.

Github Repository

The codes can also be accessed at https://github.com/clai-group/long_covid_ai_scripts.