Preinstall needed packages on R Studio workbench

[atheo89]

JIRA: https://issues.redhat.com/browse/RHOAIENG-366

Why you need this feature:

If the bellow packages missing, the majority of the packages will fail installing, either directly or due to missing dependencies.

libsodium-devel.x86_64 
libgit2-devel.x86_64
libcurl-devel
cmake
flexiblas-*
devtools

Rcpp.h
hb-ft.h
lfreetype 
lharfbuzz 
lfribidi 
lpng
tidyverse
tidymodels
plumber
vetiver

Describe the solution you'd like:

Rcpp has to be compiled, add in containefile RUN R -e ‘install.packages(“Rcpp”)’ to get the missing headerfile. OBS! This package has to be installed first, execute it on its own line. It is the C compiler for R. If this is missing, the majority of the packages will fail installing, either directly or due to missing dependencies.

Preinstall needed packages, add to containerfile or similar:
RUN dnf install -y libsodium-devel.x86_64 && dnf install -y libgit2-devel.x86_64 && dnf install -y libcurl-devel && dnf install -y cmake && dnf install -y flexiblas-*

Preinstall these packages to save data scientist time to install. Add this to containerfile:
RUN R -e ‘install.packages(c(“tidyverse”, “tidymodels”, “plumber”, “vetiver”, “devtools”))’

Anything else you would like to add:

Document: https://docs.google.com/document/d/1Hp0cg98FqptMxerB5jtbFAcvaRH2uZZgBjTLt5FojuE/edit#heading=h.i4eajclb9qkj

Acceptance criteria

Dev:

• Explore the package that needs to be installed and provide the methods to install them.
  • Check if these packages are minimal packages.
• Install R-core package based on the exploration along with the OS binaries.
• Test the R-core package installed.

QA:

• Automation on ODS-CI to check on these pre-installed packages.
  Example: here

Docs:

• Document the information about the installed pre-installed.
• Document on how to install additional packages if the user wishes too.
  example: here