Create postInclude file for saving ID clusters & MSOSs to AOD
This MR adds a single python file that can be used with Reco_tf.py --postInclude to save Inner Detector clusters and MSOSs to the AOD.
Example command:
Reco_tf.py --AMI <r-tag> --inputBSFile <RAW file> --outputAODFile <AOD file> --postInclude 'all:InDetConfig.InDetClustersDecorationConfig.fromRunArgs'
Context:
The Exotics Pixel dE/dx analyses require access to pixel clusters, which are not available in the nominal AOD. Event picking is used to get around this. First, ~1M events of interest are chosen from the nominal datasets. Next, the event picking server is used to create DRAW datasets with just these events. Finally, custom reconstruction is performed to create AODs with pixel clusters and MSOSs.
However, the r-tags used for this custom reconstruction must be identical to those used for the nominal AODs, up to some --preExec or --postInclude to save pixel cluster. Any deviation may cause one-sided event migrations: events that were picked may migrate out of the selections, and there's no way to access events that would have migrated in.
Unfortunately, not all required releases have flags to easily write out pixel clusters via a --preExec. Even in the releases where this is possible, bug fixes are required for disappearing tracks. Furthermore, fixing this in all the required releases would be quite difficult due to all the refactoring (e.g. the move to the component accumulator). We had initially hoped that resolving this in 24.0 and relying on the Frozen Tier0 policy would be sufficient. However, we've observed migrations due to explicit Frozen Tier0 violations, as well as the coupling of Athena releases to specific conditions and geometry tags.
This solution of using postIncludes to schedule the few algorithms we need was discussed with experts (tagged) and deemed acceptable. Once this MR is approved, additional MRs will be created targeting the remaining branches needed for each year of data. Eventually, the branches for the different MC sub-campaigns will be targeted as well.
Data:
- Run 2: 22.0-mc20 (pre-CA, handled slightly differently)
- 2022/2023: 24.0.42-patches (this MR)
- 2024: 24.0.88-patches
MC:
- MC20: 22.0-mc20 (pre-CA, handled slightly differently)
- MC23a: 23.0.20-patches
- MC23d: 23.0.32-patches
- MC23e: 24.0.52-patches
Tagging experts from email thread: @jmasik @nstyles @jcatmore @jchapman and analysis team members: @spagan @okovanda