Add option to run the full HLT2 reconstruction in HltEfficiencyChecker jobs

FYI @rmatev @sstahl @mvesteri

As titled. There is a new run_reconstruction flag that can be specified in the yaml, and the "by-hand" hlt2_eff_example.py uses with reconstruction.bind(from_file=False). As such all the HLT2 HltEfficiencyChecker examples run the reconstruction by default.

On a side note, the possibility of plotting with logx is added as requested by Mike Solokoff in RTAGeneral on Friday.

TODO:

• Update the doc examples (add myself as CODEOWNER for the docs while I'm there)
• Make the rate examples use the reconstruction too
• Add a test that still runs reco from file.

@rjhunter Regarding Mike's request: do we need to add another option to the script or could we always generate the log and linear versions of the plots?

added 1 commit

• bdd65e50 - Apply suggestion to HltEfficiencyChecker/options/hlt2_eff_example.py

Compare with previous version

added 2 commits

• c2010f26 - Use reconstruction for rates examples
• b528ee06 - Merge branch 'rjhunter-use-reco' of...

Compare with previous version

added 1 commit

• ee5de1a7 - Always make logx and x plot

Compare with previous version

mentioned in issue #13 (moved)

added 1 commit

• 5f572884 - Gamble and put the legend on the left if logx

Compare with previous version

added 1 commit

• b88ae671 - Fixed formatting

Compare with previous version

marked the checklist item Make the rate examples use the reconstruction too as completed

resolved all threads

I'm happy with this now and the tests pass for me locally. I can go ahead and change the code examples in the documentation to reflect the changes here. Before I do, does anyone have any more comments/things I should address @apearce @mvesteri @rmatev @sstahl?

LGTM!

unmarked as a Work In Progress

changed the description

@msaur this is ready for testing now, I can update the documentation in the meantime.

assigned to @msaur and unassigned @rjhunter

/ci-test --merge

• [2020-07-08 17:52] Validation started with lhcb-master-mr#1056

@rjhunter There are several failures related to x86_64-centos7-gcc9-dbg and x86_64-centos7-clang8-dbg platforms. Can you check that? Specifically: On x86_64-centos7-gcc9-dbg: RecoConf.hlt2_reco_full_geometry RecoConf.mc_cheated_pattern_recognition On x86_64-centos7-clang8-dbg: RecoConf.hlt2_reco_calo_resolution_pi0 --- here some files can't be accessed RecoConf.hlt2_reco_baseline_multi_threaded HltEfficiencyChecker.test_hlt2_example HltEfficiencyChecker.test_hlt2_wizard_effs HltEfficiencyChecker.test_hlt2_wizard_rates

• avx2 issue is currently being investigated, it also created problems in many others MRs.
• regarding the timeout, for this would be beast to contact relevant maintainer and check if increased timeout will not be an issue due to some other reasons -> solved by comment form Sascha.

marked as a Work In Progress

changed the description

assigned to @rjhunter and unassigned @msaur

added 2 commits

• 294dc586 - Add new test and corresponding yaml
• a885144b - Increase timeouts to 20 mins for reco-on jobs, 40 mins therefore for...

Compare with previous version

assigned to @msaur

unassigned @msaur and @rjhunter

assigned to @msaur

unmarked as a Work In Progress

marked the checklist item Add a test that still runs reco from file. as completed

• [2020-07-09 16:27] Validation started with lhcb-master-mr#1060

mentioned in merge request Moore!571 (merged)

marked the checklist item Update the doc examples (add myself as CODEOWNER for the docs while I'm there) as completed

• [2020-07-13 09:30] Validation started with lhcb-master-mr#1076

assigned to @jonrob

mentioned in commit Moore@a23be6d8

resolved all threads

approved this merge request

merged

mentioned in commit 433aa553

unassigned @jonrob and @msaur

mentioned in merge request !16 (merged)

mentioned in issue Moore#198 (closed)

mentioned in issue Moore#202