Refractive Index Calibration Task for Run3
Note - This MR was accidentally opened against the wrong fork. Closed, and new one opened at lhcb/Panoptes!116 (merged)
Adds a new task, run as a new QM test, that runs the RICH n-1 refractive index calibration task as a dedicated job, together with the tracking and RICH reconstruction it requires. This is potentially to replace the OMALib analysis task (still available as an option) that in Run2 use the Online Brunel Savesets.
Requires lhcb/Moore!741 (merged)
@graven @sstahl Would appreciate you feed back on the tracking options I am using, as I am not 100% sure I have really understand the various options to speed things up.
I have tried four configurations
with TrackMasterFitter.bind(FastMaterialApproximation=False),\
make_hlt2_tracks.bind(use_pr_kf=False, light_reco=True, fast_reco=False):
run_reconstruction(options, standalone_rich_ref_index_calib)
with TrackMasterFitter.bind(FastMaterialApproximation=False),\
make_hlt2_tracks.bind(use_pr_kf=False, light_reco=True, fast_reco=True):
run_reconstruction(options, standalone_rich_ref_index_calib)
with TrackMasterFitter.bind(FastMaterialApproximation=True),\
make_hlt2_tracks.bind(use_pr_kf=False, light_reco=True, fast_reco=True):
run_reconstruction(options, standalone_rich_ref_index_calib)
Run-6718803-LightFastMatFast.pdf
with TrackMasterFitter.bind(FastMaterialApproximation=True),\
make_hlt2_tracks.bind(use_pr_kf=True, light_reco=True, fast_reco=True):
run_reconstruction(options, standalone_rich_ref_index_calib)
Run-6718803-LightFastMatFastPrFk.pdf
Of these, the first three all show the same CK theta resolutions for both RICHes, which means the tracking changes do not impact the RICH performance.
However, in the fourth where I enable the 'PrKf' option (a new track fit I think ?) I see a significant degradation in the RICH1 resolution, from 0.86mrad to 0.95mrad. This implies there is something in this option that is giving worse tracking resolution in the RICH1 region ??
Finally, I would still like if possible to apply some tighter cuts to the tracks (P,Pt) before the track fit to avoid making tracks I don't need for the calibration. Is there a way I can do this ?