Prepare R dependencies for PsN

Overview

This workflow prepares the R dependencies for PsN by installing PsNR along with all of the dependencies required. It uses renv.lock to restore most of the dependencies, and the pacakge cellar for PsNR and vaplot pacakges.

The 2 cellar packages are built with slight tweaks to enable a local workflow. 1. The original PsNR requires vaplot to come from Remotes: rikardn/vaplot@digits, we don't want to rely on a GitHub branch so we removed the remote source. 2. vaplot was built fresh from the digits branch as a local tarball. This way we can use it as a cellar pacakge without having to install from GitHub every time we set up the environment.

rlibs_r43_jammy.tar.xz contains all the R libraries needed (all from ~/your-project/renv/library/R-4.3/x86_64-pc-linux-gnu) to run library(PsNR).

How it works

1. Start an RStudio project and make sure you have renv activated. This workflow was tested on renv 1.0.3.

2. Clone this repo in a terminal.

   git clone https://github.com/A2-ai/misc-assets

3. Copy the .Rprofile from misc-assets to your project root so it replaces the existing file. This ensures the packages we are going to install to not use cache or symlinks, so we can copy them later on.

   cp ./misc-assets/.Rprofile .

   Move the folder cellar as a subfolder of renv in your project, so renv detects it when installing the 2 packages.

   mv ./misc-assets/cellar/ ./renv/

4. Install all the packages in the R console.

   renv::restore()

5. Check whether PsNR loads correctly in R.

   library(PsNR)

6. Copy copy_rlibs.sh to project root and run it.

   cp ./misc-assets/copy_rlibs.sh . && bash copy_rlibs.sh

7. We should see 124 folders in /opt/PsN/5.3.1/PsN_5_3_1/rlibs

   ls /opt/PsN/5.3.1/PsN_5_3_1/rlibs | wc -l