Overlap improvements
The full set of MRs is now
lhcb-datapkg/PRConfig!509 (merged) DaVinci!1230 (merged) LHCbIntegrationTests!96 (closed) LHCb!4913 (merged) !4327 (merged)
See for reference also !4195 (merged)
Discussion held
https://codimd.web.cern.ch/WvIE4KWAR0q68brTWCnAbg#
Explanation
We now have 4 input_process
Hlt2
Only processed up to Hlt2, no Sprucing run at all
RawEvent at /Event/DAQ/RawEvent
, candidates at /Event/HLT2/Hlt2<line>/Particles
, reco at /Event/HLT2/Rec
TurboPass
(for technical streams)
Processed by passthrough sprucing - DstData rawbank never opened
RawEvent at /Event/<stream>/RawEvent
, candidates at /Event/HLT2/Hlt2<line>/Particles
, reco at /Event/HLT2/Rec
TurboSpruce
(for TURBO stream)
Processed by Sprucing, removes overlap by creating corresponding Sprucing lines and re-streaming. Only the persistency is handled - no physics changes
RawEvent at /Event/<stream>/RawEvent
, candidates at /Event/Turbo/Hlt2<line>/Particles
, reco at /Event/Turbo/HLT2/Rec
Spruce
(for FULL stream)
The exclusive Sprucing - redoing selections from the reco objects (FULL sprucing)
RawEvent at /Event/<stream>/RawEvent
, candidates at /Event/Spruce/Spruce<line>/Particles
, reco at /Event/Spruce/HLT2/Rec
Notes
- Does not affect MC - MC TURBO does not get Spruced in any way
- DaVinci should be able to handle all these
input_process
types