Skip to content

Installing EpiBenchmark on the UNC Longleaf Cluster

This guide describes how to install and use EpiBenchmark on the UNC Longleaf HPC cluster.

Prerequisites

Before installing EpiBenchmark, verify or load a compatible Python module.

Verify the Python version:

python --version

List the available Python versions:

module avail python

Load a supported version (Python 3.10 or newer). For example:

module load python/3.11

Note The exact module name may change over time. Use module avail python to see the versions currently available on Longleaf.

Clone the repository:

git clone https://github.com/ACCIDDA/EpiBenchmark.git
cd EpiBenchmark

Option 1: Install with uv

First, verify whether uv is available:

uv --version

If uv is installed, create the virtual environment and install EpiBenchmark package:

uv sync

Activate the environment:

source .venv/bin/activate

Verify the installation:

epibench --help

If uv is not installed

Install uv into your user account:

python -m pip install --user uv

Make sure your user-level executable directory is on your PATH. Users can run echo $PATH to check user-level directory. If you see /nas/longleaf/home/<onyen_ID>/.local/bin in the output list, then your PATH is already configured correctly.

Then run:

uv sync

Verify the installation:

epibench --help

Option 2: Install with pip

Create a virtual environment:

python -m venv .venv

Activate it:

source .venv/bin/activate

Upgrade pip:

python -m pip install --upgrade pip

Install EpiBenchmark package:

python -m pip install -e .

Verify the installation:

epibench --help

R Requirements for Scoring

To run epibench score command, users need:

  • R
  • Rscript
  • the CRAN package scoringutils

Verify the R version:

Rscript --version

If R is not installed, check that R is available:

module spider r

Load the default R

module load r

R library setup

Before installing scoringutils, check whether R already has a personal library configured:

Rscript -e '.libPaths()'

This is because R will try to install packages into the system library if no users' personal library configured. That system directory is owned by the Longleaf administrators, ordinary users cannot write to it.

If there isn't a user library, create one.

mkdir -p ~/R/x86_64-pc-linux-gnu-library/<R_version_installed_in_longleaf>
e.g. mkdir -p ~/R/x86_64-pc-linux-gnu-library/4.4

Note: 4.4 will be used in the following command examples.

Add the created user library to libPaths and verify:

Rscript -e '.libPaths(c("~/R/x86_64-pc-linux-gnu-library/4.4", .libPaths())); print(.libPaths())'

Users should see the following: 

[1] "/home/<onyen_ID>/R/x86_64-pc-linux-gnu-library/4.4"
[2] "/nas/longleaf/rhel9/apps/r/4.4.0/lib64/R/library"

Install package scoringutils

Once the user-level library exists, run:

Rscript -e '.libPaths(c("~/R/x86_64-pc-linux-gnu-library/4.4", .libPaths())); install.packages("scoringutils", repos="https://cloud.r-project.org")'

Verify the installation:

Rscript -e '.libPaths(c("~/R/x86_64-pc-linux-gnu-library/4.4", .libPaths())); library(scoringutils)'
if no error appears, the scoringutils is installed correctly.

Make R configuration permanent

Create a profile file: 

nano ~/.Rprofile

Put the following line inside and save the file: 

.libPaths(c("~/R/x86_64-pc-linux-gnu-library/4.4", .libPaths()))
Now users can simply load the R module and run EpiBenchmark without repeating the library configuration in Longleaf.

Running EpiBenchmark

Activate the virtual environment whenever you begin a new session:

source .venv/bin/activate

Example commands:

epibench
epibench --help
epibench setup --help
epibench score --help
epibench plot --help

Deactivate the environment when finished:

deactivate