This MR aims at implement a common framework to persist in the extra_output selections of particle associated to isolation, useful in several physics analysis to suppress offline the partial reconstructed background.
The goal is to have a common configurable tool to be implemented in different lines to persist a reduced amount of information, necessary to make offline studies with the isolation. This supports the idea of saving space for HLT2 BW in Turbo lines.
ATM, these are the persisted TES locations:
To get these object persisted, we can use the method select_parts_for_isolation in isolation.py as follows:
set a list of names of particles wrt we want to calculate isolations,
set a list of containers of particles wrt we want to calculate isolations,
set the predicate that will filter the selection of extra particles,
set to True the flags correspondent to the kind of isolation that we want to evaluate
The following list of extra particles are accessible to be persisted and to compute offline isolation:
make_long_pions for long track isolation (by default: True)
make_ttrack_pions for t-track isolation (by default: False)
make_down_pions for downstream track isolation (by default: False)
make_up_pions for upstream track isolation (by default: False)
make_photons for neutral isolation (by default: True)
make_merged_pi0s for neutral isolation with pi0s (by default: False)
Furthermore, the MR implements the method find_in_decay in isolation.py, which returns a selection of particles in the decay tree a particle container matching a given id, passed as string parameter. In this way it could be possible to save extra information only for signal children candidate and limit the space for the HLT2 BW.
Example of application could be found in b_to_xtaul_hlt2.py file.
Further developments and improvements are discussed in #494
TODO:
Add the possibility to save merged pi0 and set by default to False
Change ~ with not when excluding signal candidates
Check what happen when empty array is saved in extra_outputs
Have a rough estimation of HLT2 BW saved by implementing isolation
Have a look at isolation variables on persisted selections on MC and data
