Standardize and modularize slurm scripts

[f-hafner]

goal:

• make slurm modules easier to maintain and load
• use consistent file names for logs
• make it flexible for running jobs on OSSC and bare-metal snellius (what is the best way to find out whether a script is being run on the ossc or on snellius?)

Steps:

1. on OSSC and in this repository, create directory slurm_setup/, and then create files modules.sh and venv.sh
   • modules.sh should call module load xyz in order
   • venv.sh calls source "$PREFIX"/virtual_envs/"$VENV"/bin/activate
2. on OSSC, move the virtual environments from root into dedicated folder virtual_envs

3. in this repo, the slurm scripts should follow the same conventions

   • (common location for logs -- in the future) -- see #27
   • common names for log files, ie jobname.jobid.err and *.out as is currently done in some scripts (I'm open to alternatives)

   • declare local variables and load

     declare PREFIX="/gpfs/ostor/ossc9424/homedir/"
     declare VENV="ossc_new" # the name of the virtual environment
     
     source "$PREFIX"/slurm_setup/modules.sh
     source "$PREFIX"/slurm_setup/venv.sh
     
     # run scripts
     

work on this once branch pipeline_parallel is merged.

[f-hafner]

I have tested the logic above on the ossc and it works