inclusive_R_D

Inclusive tagging measurement of R(D) and R(D*) at Belle II

General procedure

1. Generate small run independent signal MC and test reconstruction locally in the Samples/Signal_MC_ROEx1
2. Prepare reconstruction scripts to run on the grid in the Recon_script
3. Download Ntuples from the grid to Samples
4. Merge Ntuples by hadd merged.root file1.root file2.root ... fileN.root
5. Run offline scripts 1-8. Script 3-5 are deprecated.

Offline procedure

Procedure	Purpose
1. python3 Apply_DecayHash_BCS.py	Apply decayhash/offline cuts and save parquet files for signal/generic_bb and root files for continuum in Samples
2. python3 Prepare_Training_Samples.py	Use MCtruth and decayhash to select signal and particular background for BDT training save output root files to BDTs
Deprecated	Deprecated
3. cd BDTs basf2_mva_merge_mc	Merge signal and different bkg to prepare train/test samples for BDTs.
4. BDT_Grid_Search.py and BDT_Training.py	Grid search hyperparameters and train the first 3 BDTs (CS,DTCFake, BFake)
5. Add_Labels_or_combine.py basf2_mva_expert	Add labels to all files in BDTs (signal, continuum, DTCFake, BFake) Apply 3BDT weights and get file_applied.root for each type
6. Add_Labels_or_combine.py	Combine 3 BDT outputs with spectators to file_applied.root for the 4th BDT training
7. basf2_mva_merge_mc	Merge signal_applied and all bkg_applied to prepare the 4th BDT train/test
8. BDT_Grid_Search.py and BDT_Training.py	Train the 4th BDT
9. Add_Labels_or_combine.py basf2_mva_expert	Add labels to all files in Samples (bb, qq) Apply 3BDT weights and get file_applied.root for each type
10. Add_Labels_or_combine.py basf2_mva_expert	Combine spectators and 3BDT outputs in file_applied.root Apply 4th BDT weights and get file_applied_2.root for each type
Continue from step 2	Continue from step 2
3. python3 7_LightGBM_tuner.py or 8_XGBoost_tuner.py	Tune hyperparameters with optuna of multiclass models.
4. python3 7_LightGBM_training.py or 8_XGBoost_training.py	Train multiclass models.

Required libraries

1. plotly
2. mplhep-0.3.23
3. pyarrow-10.0.0
4. iminuit-2.18.0
5. cabinetry
6. optuna-3.1.0
7. h2o-3.40.0.1