cameronc137
commented
4 years ago I have tested the results of placing a mandatory Max-burst index on JAPAN pass 2 ROOT files, within the disable-tree-correlator branch (technically I actually tested it in my old copy of the longer-burst branch with operations from today pulled into it).
It looks like the mechanics of it all work out.
Pass 0.5 -> do this:
- disable-trees = 1
- disable-histos = 1
- print-burst-index-map = 1
- min-burstlength = 9000 (or make this shorter depending on how small of a burst you want to have, 9000 forces every single run to define a pass 2 max burst counter, which may be non-ideal)
Pass 2 -> do this:
- add-config = max_burst_index.conf which contains a blank line, and then max_burst_index.####.conf is printed out by the 0.5 pass into ./ directory, and will contain "max-burst-index = #" which sets the max burst index to iterate through in Pass 2 analysis. For example, a run with only one full burst (<18000 good patterns according to the above prescription) will have the number 0 as the max-burst-index.
We should see that the results should be exactly identical when no burst minimum limits are applied, and that does work. Additionally we should see that even the long bursts raw data (not the LRB corrected data though, as slopes will be changed as well with a pass1 using the longer bursts) should match, with appropriate cut to make the number of entries the smaller burst size. This is also seen, in the attachment.
I think this is ready to go: NewBurstsWork.pdf
These changes should allow us to run the "pass 0.5" with no tree outputs to determine where the minirun transitions will occur, and then run later analysis passes which combine the last minirun with the previous one if it has "too few" events.