Build documentation

Currently done locally.

cd book
quarto render .

Then push to GitHub

2026-02-07

Update pangeo feedstock to latest version 2026-01-21
Set Python back to 3.11, same as CryoCloud
R packages installed with install-r-packages.sh are not using binaries. I was circumventing the behavior in $R_HOME/etc/Rprofile.site that ensures that using repo p3m.dev gets binaries instead of source.
The set -euo meant that unbound variables in scripts throw and error. This is probably correct although now threw some errors when calling install-r-packages.sh

A series of fixes related to VSCode OSS and ssh into JuptyerHub from VSCode

code-server was by default opening in ~ and trying to index all the file = RAM blowup
bash terminals in code-server and vscode are not login and so notebook conda env was not being activated
ssh does not get the Docker ENV so failing when bash startup tried to use CONDA_DIR
the various scripts were not using set -euo so failing but not erroring out

Update R to 4.5.1
Unbuntu updated to 24.04 in base image
Fix JupyterLab Quarto extension not working by adding back in jupytext after Quarto team fixed bug
Add in a few Python packages: conda-lock, pigz, jupyterlab-quarto (was in geospatial-2)

I had removed ms-toolsai.jupyter from the VSCode extensions but turns out that is required for VSCode to work in a Jupyter Hub. Without it, first Jupyter notebooks don't open but more importantly, there is a memory leak that will crash the kernel in about 5-10 minutes.
I added a catch in the Docker file so that the build fails if the extensions script does not run.

Adding IRkernel in Docker file. Added notes to developer section of documentation. Allows one to select R kernel in Jupyter Lab.
Updated to 4.4.3 as 4.4.2 no longer on rocker-org

Fixed problem with tex not building due to dead tinytex link
Fixed problem with Quarto + JupyterLab by pinning Quarto to 1.5.57. quarto-dev/jupyterlab-quarto#13
Updated to 4.4.2