UPP: Umami PreProcessing

This is a modular preprocessing pipeline for jet tagging. It addresses several issues with the current umami preprocessing workflow, and uses the atlas-ftag-tools package extensively.

Documentation is under construction here

Comparisons with umami

Main changes

• modular, class-based design
• h5 virtual datasets to wrap the source files
• 2 main stages: resample -> merge -> done!
• parallelised processing of flavours within a sample
• support for different resampling "regions", which is usefull for Xbb preprocessing
• ndim sampling support, which is also useful for Xbb
• "new" improved training file format (which is actually just the tdd output format)
  • structured arrays are smaller on disk and therefore faster to read
  • only one dataloader is needed and can be reused for training and testing
  • other plotting scripts can support a single file format
  • normalisation/concatenation is applied on the fly during training
  • training files can contain supersets of variables used for training
• new "countup" samping which is more efficient than pdf (it uses more the available statistics and reduces duplication of jets)
• the code estimates the number of unique jets for you and saves this number as an attribute in the output file

Performance and LOC

Compared with a comparable preprocessing config from umami:

1. train file size decreased by 30%
2. train read speed improved by 30% (separate from file size reduction, by using read_direct)
3. only one command is needed to generate all preprocessing outputs (running with --split=all will produce train/val/test files)
4. lines of code are reduced vs umami by 4x
5. 10x faster than default umami preprocessing (0.06 vs 0.825 hours/million jets)