Skip to content

Draft: split systematics in EgammaCalibrationAndSmearingAlg (ATLASG-2358)

Ryan Quinn requested to merge rquinn/athena:split-egamma-calib-tool into main

This is a follow up to !57320 (merged) which implements the new functionality and adds some basic validation in a CI test.

Running EgammaAnalysisAlgorithmsTest_jobOptions.py with the argument --split-systematics will cause EgammaCalibrationAndSmearingAlg to be run twice: once with noToolSystematics set to False, and then again with it set to True. The default behaviour is unchanged, so running without the new argument will cause EgammaCalibrationAndSmearingAlg to only be run once with noToolSystematics = True.

The file test_split_calibration_from_systematics.py contains a basic unit test which runs Athena twice on the MC ASG test file, once with the original behaviour and then again with the new behaviour. The kinematic histograms are then compared, and the test will fail if any bin contents differ. The tests run over 2000 events, which can easily be changed if the tests take too much time.

Tagging @krumnack as Nils created the first MR.

Edited by Ryan Quinn

Merge request reports