Goal: Detail how to update an R Conda environment so it can be used as a kernel within ARCC’s Jupyter service.
General Process
The general process involves updating the conda environment to include kernel related packages, and then configuring the kernel spec to allow it to be picked up by the Jupyter service.
Activate you R Conda Environment.
Start R.
Install the
IRkernelpackage.Exit R.
Deactivate your Conda environment.
Copy the created
kernelspecinto your home.local/share/jupyter/kernels/folder.Update the kernel.json.
Install the IRkernel package
Activate you R Conda Environment and start R.
[salexan5@mblog2 ~]$ module load miniconda3/24.3.0
[salexan5@mblog2 ~]$ conda activate /project/arcc/software/conda-envs/r_4.3.3_env/
(/project/arcc/software/conda-envs/r_4.3.3_env) [salexan5@mblog2 ~]$ R
> install.packages('IRkernel')
...
also installing the dependencies ‘fastmap’, ‘fansi’, ‘utf8’, ‘htmltools’, ‘pillar’, ‘base64enc’, ‘repr’, ‘evaluate’, ‘IRdisplay’, ‘pbdZMQ’, ‘crayon’, ‘jsonlite’, ‘uuid’, ‘digest’
...
* DONE (IRkernel)
Created kernelspec folder
Installing the IRkernel package will create a kernel spec that we can use.
This can be found under the Conda environment location, under: lib/R/library/IRkernel/kernelspec/
[salexan5@mblog2 ~]$ ls /project/arcc/software/conda-envs/r_4.3.3_env/lib/R/library/IRkernel/kernelspec/ kernel.js kernel.json logo-64x64.png logo-svg.svg
[salexan5@mblog2 ~]$ cat /project/arcc/software/conda-envs/r_4.3.3_env/lib/R/library/IRkernel/kernelspec/kernel.json
{"argv": ["R", "--slave", "-e", "IRkernel::main()", "--args", "{connection_file}"],
"display_name":"R",
"language":"R"
}
Configure Your Jupyter Environment
If you haven’t used the Jupyter service, then you might not have, and thus will need to create the following folders:
.local/share/jupyter/kernels/
Copy the created kernelspec into your home:
[salexan5@mblog2 ~]$ cd .local/share/jupyter/kernels/ [salexan5@mblog2 kernel]$ mkdir r_4.3.3 [salexan5@mblog2 kernel]$ cd r_4.3.3 [salexan5@mblog2 r_4.3.3]$ cp /project/arcc/software/conda-envs/r_4.3.3_env/lib/R/library/IRkernel/kernelspec/* .
Update the kernel.json file to:
point to the Conda environment’s version of R.
give this kernel a unique display name.
[salexan5@mblog2 r_4.3.3]$ cat kernel.json
{"argv": ["/project/arcc/software/conda-envs/r_4.3.3_env/bin/R", "--slave", "-e", "IRkernel::main()", "--args", "{connection_file}"],
"display_name":"R_4.3.3 (local)",
"language":"R"
}
Start Jupyter
From OnDemand start a Jupyter session.
Notice how the newly configured kernel is available.
Within a Notebook
Within a cell try:
.libPaths()
# Cell Output: '/cluster/medbow/project/arcc/software/conda-envs/r_4.3.3_env/lib/R/library'
write.table(installed.packages()[,c(1,2,3:4)])
# Cell Output: "Package" "LibPath" "Version" "Priority" "base" "base" "/cluster/medbow/project/arcc/software/conda-envs/r_4.3.3_env/lib/R/library" "4.3.3" "base" ... "vctrs" "vctrs" "/cluster/medbow/project/arcc/software/conda-envs/r_4.3.3_env/lib/R/library" "0.6.5" NA
Note the library path is that of the Conda environment: /project/arcc/software/conda-envs/r_4.3.3_env/lib/R/library