Install R and Python packages from the local Sandbox repository

Users can install R and Python packages from the local Sandbox repository.

Users can request new packages to be added to the local Sandbox repository (please see policy).

R packages

List R packages available in the local Sandbox CRAN repository using R

The Sandbox IVM has been pre-configured to use the local Sandbox CRAN repository. The available.packages() R command can be used to list the available packages in the local Sandbox CRAN repository.

Use the following command to list all available packages and their versions:

available.packages()[,c("Package","Version")]

You can search available packages by name. For example, use the following command to list all packages that start with 'Bi':

p <- available.packages()[,c("Package", "Version")] 
c <- grepl("^Bi", p[,c("Package")], ignore.case=TRUE) 
p[c,]

The available.packages()command returns other columns as well. You can list them all using the colnames command:

colnames(available.packages())

List R packages available in the local Sandbox CRAN repository using the terminal

You can list packages using the R command line program with the -e option following other instructions in the previous section.

List all packages:

R -e 'available.packages()[,c("Package","Version")]'

Search packages by name:

R -e 'p <- available.packages()[,c("Package", "Version")];c <- grepl("^Bi", p[,c("Package")], ignore.case=TRUE);p[c,]'

You can also directly list the packages in the local Sandbox CRAN repository:

ls /usr/finngen-repos/cran/source/src/contrib

Install R packages from the local Sandbox CRAN repository

The Sandbox IVM has been pre-configured to use the local Sandbox CRAN repository. Use the install.packages("package name") R command to install packages from the local Sandbox CRAN repository.

For example, the following command installs the Biobase package:

install.packages("Biobase")

Packages are installed in your Sandbox IVM persistent disk. The exact location depends on the R version, for example, /home/ivm/Rx86_64-pc-linux-gnu-library/4.3. Installed packages are automatically detected by R only from this location. When using a different version of R, you may need to use the .libPaths() variable to define the location of the package (more information is available here). You can also install the same package for different R versions.

List available R packages in the local Sandbox CRAN repository outside the Sandbox

It is possible to list the available R packages outside Sandbox by using buckets.

R packages are available in the following bucket:

gs://fg-production-master_cran/source/src/contrib

You can see the contents of the bucket using the cloud console:

https://console.cloud.google.com/storage/browser/fg-production-master_cran/source/src/contrib;tab=objects?project=fg-production-master

or the gsutil command line program:

gsutil ls gs://fg-production-master_cran/source/src/contrib

Python packages

List Python packages available in the local Sandbox PyPI repository using the terminal

The Sandbox IVM has been pre-configured to use the Sandbox PyPI repository on GCP Artifact Registry: https://europe-west1-python.pkg.dev/finngen-sandbox-v3-containers/pypi/simple/. It contains packages for Python 3. FinnGen uses the pip program to install Python packages in the Sandbox.

Use the following command to list available Python 3 packages:

python3 -m pip list --index-url https://europe-west1-python.pkg.dev/finngen-sandbox-v3-containers/pypi/simple/

Install Python packages from the local Sandbox PyPI repository using the terminal

Use the following command to install a Python 3 package:

python3 -m pip install <packagename>

Notice for Debian 12 Users: To comply with PEP 668 and prevent system instability, global package installation via pip is blocked by default. The recommended and safest fix is to isolate your packages using a virtual environment before executing your pip install commands. Alternatively, if you explicitly require a global installation and accept the risk of breaking system-managed utilities, you can bypass this restriction for a single command by appending the --break-system-packages flag.

Python lmod modules with preinstalled scientific python packages are available in the sandbox. You can read about them here.