reusable analysis To reusability and beyond Daniel Prelipcean
reusable analysis To reusability… and beyond! Daniel Prelipcean Supervisor: Tibor Simko IT-CDA-DR CERN Summer Student Programme: Student Session 6 th August 2019 1
Before @BSc: Jacobs University Bremen Physics and Mathematics CERN @CERN: IT-CDA-DR: Development for the reana project, focusing on Particle Physics analysis examples. After @Ph. D: ~software tools for physicists @MSc: Technical University of Munich Applied Engineering Physics @Best. Experience. At. Cern: Feeling that we are part of something greater than social constructs, e. g. borders and countries 2
Reproducibility crisis “ 50% could not reproduce their own experiments” “ 70% of researchers have tried and failed to reproduce another scientist's experiments” https: //www. nature. com/news/1 -500 -scientists-lift-the-lid-on-reproducibility-1. 19970 3
Long term value of data 4
Thomas Steckler FORCE 2015, Oxford, Jan 2015 taken from https: //slideplayer. com/slide/5818937/ 5
reana approach There are 4 questions to be answered: 1. Input data? 2. Code? * 3. Environment? 4. Analysis workflow? Serial Yadage 6
Simplified and real physics examples 7
ALICE Open Data example 1. Input data: 2. Code: 3. Environment: Ali. Physics/Ali. Root -> reana-env-aliphysics 4. Workflow: serial 8
reana. yaml one workflow file containing all the necessary information for reproducibility 9
Results? 10
Results? ● Someone outside the collaboration has successfully used the collaboration-specific framework to reproduce existing results ● Reproducibility consists in one single workflow file REANA. yaml ● Science should be show me, not trust me It should be help me if you can, not catch me if you can 1 ● Still 7 more weeks to go -> CMS reconstruction demo, . . . 1 https: //www. nature. com/articles/d 41586 -018 -05256 -0 11
Contact us! Additional resources: Reana team Tibor Simko Team leader tibor. simko@cern. ch Rokas Maciulaitis Team member rokas. maciulaitis@cern. ch Diego Rodriguez Team member diego. rodriguez@cern. ch Leticia Wanderley Open. Lab Student leticia. wanderley@cern. ch Jan Okraska Team member jan. okraska@cern. ch Daniel Prelipcean Summer Student daniel. prelipcean@cern. ch Web: http: //www. reana. io Github: @reanahub Twitter: @reanahub Email: info@reanahub. io 12
Backup slides 13
Flexible workflow support Serial Yadage CWL 14
Technology stack 15
reana architecture 16
- Slides: 16