r-docker-renv: recreating the r-docker environment (essentially on Windows but in theory on Intel Macs and x86_64 Linux as well)

This repo recreates the r-docker environment on your local installation of R.

Only the package versions and version of R will be the same. The operating system and architecture may be different, hence you may see some very small differences in results.

Setup on Windows

Install R 4.0.5 for Windows, installer here
Install Rtools 4.0 from here
Open RStudio using R 4.0.5 by holding Ctrl when you clicking the RStudio icon. This will show the RStudio R version selection window
Open this project in that RStudio running R 4.0.5, by clicking File | Open Project...
Run setup.R (technically we should just be able to run renv::restore(prompt = FALSE) but some hackery is needed for a few packages first)
```
source('setup.R')
```
Run restore.R or
```
renv::restore(prompt = FALSE)
```
Here's proof that this ran successfully on Windows

A warning you can ignore: After successfully running the restore if you run renv::status() it will report the following.

renv::status()
#> The following package(s) do not appear to be used in this project:
#>                       _
#> BH                      [1.72.0-3]
#>... (I deleted alot of output here) 
#> zoo                     [1.8-8]
#> 
#> Use `renv::snapshot()` to remove them from the lockfile.

You can ignore the advice to remove the packages because all this means is that there is no .R/.Rmd/.qmd script/s in the project that uses the all of the packages.

To do some work in an OS project, change the working directory to that of your OS project using setwd(), and use RStudio's menus, e.g., File | Open File..., to open/create/modify/save R scripts as usual.

Setup on macOS (Intel processor only)

macOS users should install Xcode developer tools by issuing in the Terminal
```
xcode-select --install
```
A Fortran compiler is required to be installed in the location R is expecting (note that the version from Homebrew will not work).
- R 4.0.5 was released on 31/3/2021 so my guess is try installing the dmg file from either (depending on which macOS version your Intel Mac has: Catalina or Big Sur or older??)
- Catalina: try this
- Big Sur: try this
- Or for even older macOS versions: try this
Install R 4.0.5 using the relevant installer or use rig
R installer for macOS
- Intel processor, here
- Apple processor - I don't think R 4.0.5 was released for Apple Silicon, so run the Intel version above (it will automatically run under rosetta2)
rig https://github.com/r-lib/rig#readme, first install Homebrew, then install rig
```
brew tap r-lib/rig
brew install --cask rig
```
- (Only works on Windows and Intel Macs) Install R 4.0.5
```
rig add 4.0.5
```
On macOS launch the Rig.app menu bar app which will now be in your Applications directory
On Apple Silicon Macs run
```
rig system make-orthogonal
```
Use the Rig.app menu bar app to launch RStudio under R 4.0.5 as follows
Open this project, e.g., by double-clicking the renv-r-docker.Rproj file
Run setup.R (again we should just be able to run renv::restore(prompt = FALSE) but some hackery is required for a few packages)
```
source('setup.R')
```
Run restore.R or
```
renv::restore(prompt = FALSE)
```

Updating the environment when r-docker is updated

Update the renv.lock file by running in a Bash shell
```
./fetch.sh
```
Then run
```
renv::restore(prompt = FALSE)
```

If this fails try upgrading renv first, i.e.,

renv::upgrade(prompt = FALSE)
renv::activate()
renv::restore(prompt = FALSE)

remlapmot/renv-r-docker

r-docker-renv: recreating the r-docker environment (essentially on Windows but in theory on Intel Macs and x86_64 Linux as well)

Setup on Windows

Setup on macOS (Intel processor only)

Updating the environment when r-docker is updated