starterkit-lessons
starterkit-lessons copied to clipboard
lhcb
Basics of analysis preservation
We would like to teach people from the beginning good practices that can make an analysis reproducible. What would be the minimum set of skills / tricks people would need to know? Here are some ideas:
- basic snakemake intro
- usage of the containerization template https://gitlab.cern.ch/lhcb-analysis-preservation/containerization-cookie (this template will receive updates before the next starterkit)
- gitlab WG groups and eos WG space, usage of xrootd for file access
I love the idea, although I would make it more about good practices and "common tools" (xrootd, eos spaces, etc) than specific tools such as snakemake. For those, I would certainly make a list and discuss them, but I would not lean towards one or the other (unless the AP group has converged on a single tool, of course).
The point is to demonstrate automation and processing pipelines. Snakemake has gained most traction in the collaboration so far. CERN also had a very successful meeting with the Common-Workflow-Language people. I agree that the tool is less important and the pattern counts, but to make a hands-on lesson we need to choose an example. In the AP roadmap document snakemake is recommended.
