particleflow
particleflow copied to clipboard
MLPF datasets v2.0: track pythia-level genjets, genmet in datasets; add per-particle ispu flag; generate CMS ttbar+noPU [WIP]
Regenerated all relevant datasets as v2.0:
- [x] CMS: new CMSSW generation with CMSSW_14, new postprocessing, new tfds
- [x] add CMS ttbar no-pileup sample as a cross-check
- [ ] improve and cross-check ground truth
- [ ] CLIC: new postprocessing, updated with the features below, crucially a corrected genStatus definition (introduced a while ago but never regenerated)
- [ ] CLD: first generation, consistent setup with CLIC
- [x] Delphes: deprecated and removed
New features and fixes in v2.0 in postprocessing:
- [x] for CMS postprocessing2.py, use only CaloParticles
- [ ] track pythia-level genjets, genmet in datasets as the ultimate jet/MET reconstruction target
- [x] for CMS: generate v3_1 with updated PFAnalysisNtuplizer, added pythia genjets and genmet: https://github.com/jpata/cmssw/commit/eac61923935b6d7f6a9c9a0f636837dcb0e7c78a
- [x] for Key4HEP: compute genjets, genmet on postprocessing using only visible status=1 particles
- [ ] propagate per-particle
ispuflag- [x] for CMS postprocessing2.py, implemented and cross-checked
- [x] for Key4HEP: currently placeholder
We find perfect alignment between genMetTrue and CaloParticleMET in single-particle guns, PF ntuple ROOT level:
and at the MLPF truth level: