Task #11872

Establish a standard way to represent particle identification results

Added by Gianluca Petrillo almost 5 years ago. Updated over 3 years ago.

Target version:
Start date:
Due date:
% Done:


Estimated time:


Robert Sulej prompts us to discuss about

... a uniform way of saving PID results. [...] dE/dx PID result is a vector of probabilities for each tested "pdg" hypothesis (and optionally probability of "none of known pdg"), we do the same for e/$\gamma$ and stopping p/_K_/$\pi$/$\mu$ (and for test "stopping or not"). Now the algorithm outputs std::map<pdg, probability> - maybe it is reasonable format. We did not add yet any data product to save the result. I have a feeling that PID algorithm outputs in larana can be a bit more standardized, if their authors agree of course, instead of hardcoded names like kaon_prob, kaon_chi2, kaon_mva (just for example, the actual names in AnalysisBase are different).
The agreement on the PID output convention can be very useful since people are starting work on modules for higher level analysis.

Jim Kowalkowski also pointed out work on a tuple-like data product he's working on supporting, that might be useful in this context.


#1 Updated by Katherine Lato over 4 years ago

From Jim K: the description of needs during the meeting may have been for a data product that had something that looked like an n-tuple i.e. a bunch of named column of specific types. The generic n-tuple product supports stuff like that.

#2 Updated by Katherine Lato about 4 years ago

  • Status changed from New to Accepted
  • Target version set to 2017-2-quarter

This is tied to the track redesign effort.

#3 Updated by Katherine Lato about 4 years ago

  • Target version changed from 2017-2-quarter to 2017-4-quarter

#4 Updated by Katherine Lato over 3 years ago

  • Target version changed from 2017-4-quarter to 2018-3-quarter

Also available in: Atom PDF