CRiSp Prototyping

This repository contains initial prototyping material, such as initial input datasets, scripts, and notebooks.

Installation

R

R (version 4.4) should be installed (see e.g. instructions on CRAN). Setup the R environment using renv from the lock file:

Rscript -e "install.packages('renv')"
Rscript -e "renv::restore()"

or "from scratch" using the dependencies in the DESCRIPTION file (note that the renv.lock file needs to be removed):

Rscript -e "renv::init(bare=TRUE)"
Rscript -e "renv::install()"

Note for Linux users: By default, R will install packages from the CRAN repository, which, unfortunately, does not maintain Linux binaries. The commands above will thus build packages from source, taking a very long time to complete (and potentially raising issues with missing system libraries). Instead, one can instruct R to install packages from the RStudio package manager (Posit), which includes binaries for Linux distributions. On this web page select the Linux distribution, then copy the following lines to the .Rprofile file, replacing the URL below with the one formatted under "Repository URL":

options(repos = c(CRAN = "https://packagemanager.posit.co/cran/__linux__/XXXXXXX/latest"))

Place the .Rprofile file in the project root folder (or in your home directory to add the option for all user projects).

Python

Setup the Python environment using venv:

python -m venv venv
source venv/bin/activate
python -m pip install -r requirements.txt

Alternative installation using Conda

An environment with both the R and Python dependencies can be created from the provided environment.yml file using Conda (or its faster implementation Mamba). Conda can be installed using the Miniforge scripts provided here (download one of the Mambaforge scripts in order to install Mamba as well), then run:

# replace `conda` with `mamba` if using Mambaforge
conda env create -f environment.yml
conda activate crisp

Adding dependencies

Add R dependencies to the DESCRIPTION file, then update the renv.lock file using:

Rscript -e "renv::snapshot(type='explicit')"

Add Python dependencies to the requirements.txt file.

Both R and Python dependencies should also be added to the conda environment.yml file.