Skip to content

RECAST for ATLAS

Welcome to the documentation pages for RECAST. The resources available on these pages are intended to help analysis teams to prepare their analysis such that they can be used within the RECAST framework.

TL;DR

The recommended way to start preparing an analysis for RECAST is to

  1. Build Docker Images for each of the software environments used in the analysis.
  2. Prepare Scripts for the individual Steps of the analysis.
  3. Write a Workflow Description for your analysis.
    • **Development: ** We suggest developing your workflow on your local computer with small test input, and maintaining the workflow description files in a dedicated GitLab project.
    • **Preservation: ** To preserve your completed workflow, ask the analysis preservation contact to upload your input files to the RECAST project area on eos (/eos/project/r/recast/atlas/). Then add a GitLab CI job to your workflow gitlab project to automate your workflow execution, as described in Running Workflows Within CI.

Example Run

You can run an example RECAST workflow on your local machine. This should take approximately 3 Minutes

python -m pip install recast-atlas
recast run examples/rome

at the end you should see an output like

... some earlier logs ...
2019-05-31 12:37:29,955 |                adage |   INFO | workflow completed successfully.
2019-05-31 12:37:29,955 |  yadage.steering_api |   INFO | done. dumping workflow to disk.
2019-05-31 12:37:29,968 |  yadage.steering_api |   INFO | visualizing workflow.
2019-05-31 12:37:30,321 | recastatlas.subcomma |   INFO | RECAST run finished.

RECAST result examples/rome recast-b6746f1c:
--------------
- name: CLs 95% based upper limit on poi
  value:
    exp: 0.8924846399964371
    exp_m1: 0.6377501820447065
    exp_m2: 0.4731739008380644
    exp_p1: 1.2720762961732819
    exp_p2: 1.7545752712294322
    obs: 1.3352971254860764
- name: CLs 95% at nominal poi
  value:
    exp: 0.25999040745937085
    exp_m1: 0.10547655600578199
    exp_m2: 0.03889040527686523
    exp_p1: 0.5345040672498215
    exp_p2: 0.8276574946063575
    obs: 0.574709475331039

Resources

Where to ask questions

If you have questions about the process of recasting and using recast-atlas there are multiple sources of documentation and help to address them. In order to try and get your question answered quickly, and make sure it contributes to the RECAST knowledge base, it is recommended that you work your way through these resources in the following order:

  1. Check the docs to see if the question has been covered since the last time you looked
  2. Ask on the RECAST section of the ATLAS Talk Discourse
  3. Ask on the ATLAS analysis preservation atlas-ap Mattermost Team
  4. Ask on the ATLAS analysis preservation users mailing list: atlas-phys-exotics-recast@cern.ch
  5. Ask the ATLAS analysis preservation contacts by email: atlas-sw-analysis-preservation-contacts@cern.ch

Talks

Tutorials

Past tutorials (unmaintained)

Last update: June 23, 2022