[matthias-kleiner]

When ALIEN_JDL_DOTPCRESIDUALEXTRACTION=1, the residual extraction (o2-tpc-scdcalib-interpolation-workflow) and aggregation (o2-calibration-residual-aggregator) tasks now run with --enable-mc. This stores the MC truth of the residual tracks (labels, ITS/TPC/TRD truth parameters, origin) as branch trkMC of the trackData tree.

The option follows --no-mc-labels instead of a separate variable, since the MC truth needs the MC labels of the ITS, TPC and ITS-TPC tracks.
It is passed only if the O2 version used for the reconstruction provides it (option_if_available, which also respects the async reco tag), so older O2 tags are unaffected.
The MC truth needs an O2 version with the residual MC truth commits plus #15943, which writes and reads it through the ROOT files.
Cost on PbPb TFs: interpolation peak RSS about 2.8 GB vs 1.1 GB without MC, within the declared 4 GB; output size about +5 %.
Tested with pp, 2 TFs × 50 events, up to tpcresidmerge:

With current O2 dev, --enable-mc is set on both tasks and every residual track has its MC truth (743 tracks, 743 trkMC entries). With --no-mc-labels it is not set.
With an older O2 that lacks the option, --enable-mc is not set and workflow creation succeeds.

[github-actions]

REQUEST FOR PRODUCTION RELEASES:
To request your PR to be included in production software, please add the corresponding labels called "async-" to your PR. Add the labels directly (if you have the permissions) or add a comment of the form (note that labels are separated by a ",")

+async-label <label1>, <label2>, !<label3> ...

This will add <label1> and <label2> and removes <label3>.

The following labels are available
async-2023-pbpb-apass4
async-2023-pp-apass4
async-2024-pp-apass1
async-2022-pp-apass7
async-2024-pp-cpass0
async-2024-PbPb-apass1
async-2024-ppRef-apass1
async-2024-PbPb-apass2
async-2023-PbPb-apass5

[jackal1-66]

I thought a bit about this: the o2dpg_sim_workflow.py is supposed to work only with Monte Carlo simulations, meaning that --enable-mc should always be enabled. However it seems this is not the case as that flag will work only if MC labels are enabled, right @matthias-kleiner? I would propose to use args.no_mc_labels as if condition for the flag instead of adding an additional environment variable, given that ALIEN_JDL_DOTPCRESIDUALEXTRACTION already exists.

[matthias-kleiner]

I thought a bit about this: the o2dpg_sim_workflow.py is supposed to work only with Monte Carlo simulations, meaning that --enable-mc should always be enabled. However it seems this is not the case as that flag will work only if MC labels are enabled, right @matthias-kleiner? I would propose to use args.no_mc_labels as if condition for the flag instead of adding an additional environment variable, given that ALIEN_JDL_DOTPCRESIDUALEXTRACTION already exists.

Thanks for the suggestion! I updated the PR: --enable-mc is now set automatically whenever MC labels are on (i.e. without --no-mc-labels) and the extra environment variable is removed. It is only added if the O2 version provides the option, so older tags are unaffected.