ATLAS RECAST Tutorial: Main Page

Welcome to the UVic ATLAS tutorial on analysis preservation with RECAST! RECAST combines docker and gitlab to fully preserve your ATLAS search analysis so that it can be trivially re-interpreted with a new signal model.

Credits: This tutorial is adapted from a RECAST tutorial given at the first US-ATLAS computing bootcamp held at LBNL in August 2019, and as such makes heavy use of other materials developed for the bootcamp by Matthew Feickert, Lukas Heinrich, Karol Krizka, Samuel Meehan, Adam Parker, and Giordon Stark.

Prerequisites

  • Make sure you’ve installed docker on your computer and downloaded the files and docker images described in the setup page.

Schedule

Setup Install docker and download images and files for the tutorial
00:00 1. Introduction What is RECAST and how can it help me make the most of my analysis?
How do docker and gitlab work together to preserve your analysis code and operating environment?
What is needed to fully automate and preserve the analysis workflow?
00:15 2. Docker Crash Course Part 1: Images and Containers What’s a docker container, and how is it related to a docker image?
How do I pull docker images onto my machine and see which ones I’ve already pulled?
00:25 3. Docker Crash Course Part 2: Running and Building Containers How can I customize docker images and run them as containers on my machine?
How do I give a running docker container access to files and directories on my machine?
00:45 4. Toy Analysis What type of physics will I be analyzing during the tutorial?
How do I run a simplified event loop to get the dijet invariant mass distribution from our signal sample?
01:15 5. Gitlab CI and Docker for Environment Preservation How does gitlab CI/CD help me continuously keep my containerized analysis environment(s) up-to-date?
What do I need to add to my gitlab repo(s) to enable this functionality?
01:55 6. Coffee break! Get up, stretch out, take a short break.
02:10 7. Introducing Workflows What is the ultimate physics goal in interpreting our analysis of the VHbb signal model?
What is the goal of RECAST in the context of interpreting our analysis?
How does yadage help us preserve our analysis workflow for re-interpretation?
02:30 8. Intermezzo: Yadage Helloworld What is the syntax to define a basic yadage workflow?
03:00 9. Skimming Step for RECASTing the VHbb Analysis How do I use the yadage syntax I’ve learned to preserve the analysis steps needed to prepare my signal for interpretation?
03:25 10. Coffee break! Get up, stretch out, take a short break.
03:40 11. Scaling the MC Signal for (Re)interpretation How do I ensure that my MC histograms are scaled properly relative to one another, and to the data?
03:50 12. Signal (Re)interpretation What is a μ-scan, and how do I interpret it?
How is the final interpretation step for the VHbb analysis encoded in yadage?
04:10 Finish

The actual schedule may vary slightly depending on the topics and exercises chosen by the instructor.