coherent WaveBurst Review Meeting 19 Feb 2009 09:00 Pacific / 12:00 Eastern
Agenda: Thursday 19 Feb 12:00 Eastern
- Review of the netplot.C and simplot.C post-processing functions for the HFS
- Currently at line 683 of netplot.C
Dial-in number: 1 800 704 9896 International dial-in: +1 404 920 6472 Conference code: 5374 2349 #
AttendancePatrick (minutes), Giovanni, Marco, Gabriele, Romain, Francesco, Igor, Sergey, Virginia, Michele.
Update on Action Items from Previous Call
- FS: netplot.C code reviewed during previous call was actually an out-of-date version, which is why several apparent bugs were found. The page has been updated.
- PJS: Can you provide diffs between the old and new code? FS: No; old code no longer exists.
netplot.CContinuing review at around line 660.
- H1H2NET, etc. defined in netInit.C.
- 663: cHH is energy disbalance cut.
- 669: incremental count of surviving events for each lag.
- 678: yCUT and ycut variables are redundant counts.
- 685: write time-sorted events to a file.
- 698+: Testing number of events vs. lag for consistency with Poisson distribution. When testing, ignore bins with less than 5 counts.
- 860: Normalize counts to livetime, use sqrt(N) error bars assuming we passed the Poisson test. liveTot = background livetime.
- 877-1168: make plots
- ch2 string and cuts: FS described an important
aspect of netplot.C: the cuts are applied twice, in
two separate ways.
- In the big loop over triggers ending around line 680, the cuts are applied sequentially. Counters are incremented, and the loop is exiting early when a trigger fails a cut. This counting is how the background rate vs lag is computed.
- When making plots with the Draw() method, all of the cuts are applied simultaneously. Draw() can take as input an n-tuple of triggers and a string containing a set of logical conditions to be applied to those triggers.
- PJS: Doing the cuts twice in two different ways is dangerous -- you have to be very careful to make sure that precisely the same cuts are applied. Can you get Draw() to return you a list of triggers that passed the tests?
- IY: Yes. Have new scripts being used for LFS that use Draw() to apply the cuts. Can redirect output to a list that can then be applied as a mask to the existing event tree.
- IY uses a different netplot.C script to make the plots for the LFS second-year analysis. He needs this to apply the final segment lists, since the version used to date by Sergey can't handle extra vetoes.
- Action item: We'll need to look at IY's scripts (may have already been reviewed last year).
- 1209: ch1 string just defines the x and y axis quantities to be plotted; see. e.g., 1125.
- 1200+: Just making plots; nothing else.
- Finished netplot.C.
- Action item: Reviewers should look at it again offline.
- Sample Results Page:
FS pointed us to a sample results page:
- The scatter plots shown before the statement of post-proc cuts are made before those cuts are applied.
- The scatter plots shown after the statement of post-proc cuts are made after those cuts are applied.
- While other (older) results pages show the plots and cuts in a different arrangement, the logic of which cuts were made before and after the cuts is the same.
- E.g.: rho(ecor) vs cc histogram filled at line 586 of netplot.C.
- Next week: simplot.C.