Analysis Mtg 240513
From Moller Wiki
Back to Main Page >> Analysis Software overview >> Analysis Meetings
previous meeting << >> following meeting
Meeting information
Agenda
- Updates
- General discussion
- No meeting on May 20 due to collaboration meeting, and no meeting on May 27 due to holiday
Minutes
Participants: O. Hansen, K. Paschke, P King, B. Uduwaraarachchi, A. Sen, E. King, M. Gericke, O. Hansen, P. Gautam, R. Conaway, W. Deconinck
- MacOS branch should be ready for testing
- Buddhika: starting to work on detector visualization
- One of the possible plots that he shows is showing a difference from expected
- Kent asks to think about how to show the differences to represent failures of a particular channel in a way that doesn't hide the details of other channels too much
- Juliette asks about incorporating python in our analysis chain
- One of the possible plots that he shows is showing a difference from expected
- Discussion of the naming of the rings and modules
- Michael says that there is an existing naming scheme in the remoll for the different tiles
- Instead of "a/b/c", the ring 5 tiles should be "left", "center", "right", with the "left" being on the negative phi side.
- Do we want to keep the "front flush"/"back flush" designation in the analysis?
- Some discussion in the chat about the visualization and plotting thin
- Wouter
- you could also plot KS-test p-values since you care about differences which consider statistical power of both current and reference.
- Wouter
- In terms of speed of panguin, at least prefilling the histograms online based on what's going to be plotted may be useful. I'm looking at TTree::Draw statements in loops in https://github.com/JeffersonLab/japan-MOLLER/blob/main/panguin/macros/PlotBPMSmacros/PlotHallBPMSshort.C and that's just gotta be really slow.
- Juliette
- we used the histograms produced from Japan-MOLLER
- Wouter
- Discussion of panguin
- Juliette asks if panguin can directly plot the histograms
- Ole says that just using the histogram in the panguin config should work, but he will check if it supports histograms in directories.
- The histogram support might be in the "Jefferson Lab/panguin" repository. I will look at removing the existing panguin in japan-MOLLER and linking to the independent repo
- Juliette asks if panguin can directly plot the histograms
- Counting mode
- This will be in an independent repo, built on the Hall A analyzer and using some classes for SBS for example
- Discussion of detector signal levels
- Kent will get a list of rates for the different tiles to Juliette, then she will coordinate with me to enter them into the mock data parameter file