GitLab

In particular add the option to do a 'dry run' which doesn't submit anything
(to quickly debug what will happen) and an option to specify the files to run
over.

The grid doesn't like dataset names longer than 120 charicters. This shortens
the name, and also adds a check to make sure the name will fit before
submitting

Add some b-tagging job options and config files to:
 - Decorate tracks with the 'poormans' impact parameter info
 - Create a new b-tagging object
 - Associate tracks to b-tagging
 - Run dips on the associated tracks

The most recent athena releases drop the default impact parameter prefix
'btagIp_*' and make us specify it. This gives us more flexibility as to our
primary vertex definition but requires some changes to the the code.

Command to do this:

jq --indent 4 '.decorate = {"jet_aug": .run_augmenters, "btag_jes": "true", "soft_muon": .run_augmenters, "track_truth_info": .decorate_track_truth_info, "track_sv_info": .decorate_track_sv_info} | del(.run_augmenters) | del(.decorate_track_truth_info) | del(.decorate_track_sv_info)' $f

Mostly adding data dependencies and cleaning up job options

Samuel Van Stroud authored Jul 07, 2021 and Dan Guest committed Jul 07, 2021

This adds:
* Track truth origin label using `InDetTrackTruthOriginTool` (e.g. integer label for PU/fake/primary/fromB/...)
* Track truth vertex index (e.g. integer uniquely identifying the truth vertex of the truth particle matched to the track (if available)
* Index of JF & SV1 vertex if the track is used in one (currently not turned on - info not yet in derivations)

Bump the Athena version and the stable nightly version. Also remove the
--nFilesPerJob=1 option (hopefully that's fixed now) and add another
validation sample (to check the truth jet -> pileup vertex association samples)

Update .gitlab-ci.yml (run test for trackjets added)

See merge request dmascion/training-dataset-dumper!1