coherent WaveBurst Review Meeting 17 Feb 2009 09:00 Pacific / 12:00 Eastern
Agenda: Tuesday 17 Feb 12:00 European
- Review of the netplot.C and simplot.C post-processing functions for the HFS [wiki].
Dial-in number: 1 800 704 9896 International dial-in: +1 404 920 6472 Conference code: 5374 2349 #
AttendancePatrick (minutes), Stas, Giovanni, Marco, Gabriele, Romain, Francesco, Virginia.
- FS: Added 2 more macros to the wiki: ApplyVeto9.C and LagLoop.C. Will discuss during call.
- FS: Would also like to discuss calibration.
- FS: About half of the wiki links to CIT have been switched to the git repository.
- LagLoop.C is the code that loops over lags in netplot.C, compiled to an external library for speed.
- Action item: Reviewers to read LagLoop.C more carefully offline (~300 lines).
- ApplyVeto9.C applies category 2 vetoes to triggers in post-production. (The cat 2 data quality flags are applied during trigger production, but not the much shorter vetoes.) Cat 3 flags and vetos are not applied at all (not needed).
- Walked quickly through ApplyVeto9.C.
- Action item: Reviewers to read ApplyVeto9.C more carefully offline.
- SB: What is the efficiency of the vetoes? FS showed us the page on the efficiency of cat 2 vetoes for the HFS: wiki. In most cases the LIGO veto efficiency is close to that expected for random vetoing, which implies that the Kleine Welle LIGO vetoes (developed for the LFS) are not useful for the HFS.
- FS: Need to re-calculate the livetime after vetoing to get the correct rate vs. time.
Note added after telecon (PJS): Is this the right procedure? Can count vetoes as loss of livetime or loss of efficiency; must make sure the two calculations are consistent.
Action item: Check on consistency of veto handling.
- Makes standard plots for report pages, also computes rate vs. threshold using lag triggers.
- Fundamentally the same as netplot.C in LFS (minimal changes).
- FS led walk-through.
- line 70: liv = n-tuple of livetimes, get back total livetime.
- Important distinction: wavearray objects are simple arrays in WAT libraries, but with more methods. Can store rates, livetimes associated with various lags. ROOT n-tuple can be used to fill wavearrays or net objects.
- wavearray already contains various info on each event, e.g., GPS time, etc. If n-tuple contains extra info not recognized by wavearray then this info will not be read by wavearray. In this case have the problem that the extra structure for post-proc veto information is not expected by wavearray, so have to carry it around in extra objects.
- Wlag, Tlag, Rlag, Elag, Olag contain other info, such as time per lag and events per lag.
- FS: Suggest reading plots on standard cWB report page to see how these are used.
- line 98+: Initialize histograms (to be filled afterwards).
- 353+ Loops tocalculate start, stop, min run number, max run number, etc. (for later use).
- 476: histogram initialization finished.
- 481: Prepare file for output events.
- 504: net holds all events from the cWB run. It knows nothing of additional KW vetoes applied by AppyVeto9.C, therefore need to get copy of triggers with vetoes applied.
- 521: Loop over events.
- logic: do if (veto_condition) continue; to break out of loop and not count this event since it fails a veto condition. Note that different histograms are filled before and after applying the veto conditions.
- eg: line 534: skip zero-lag events (doing background estimation).
- 538-540: apply vetoes.
- up to 600: define various quantities for the event.
- 601: apply final threshold xcor < T_cor.
- 627-628: H1-H2 energy disbalance.
- 630-648: Fill histograms with events that survive thresholds. Write output file ftrig. Note that penalty factor not yet applied -- those that fail are written with a minus sign.
- BUG? Expected 636 to be used, not 637.
- 650: Apply penalty factor.
- BUG? Expected 651 to not be commented out.
- Action item: FS to check if the code being displayed (with apparent bugs) is the up-to-date version or an old obsolete version.
- 655: Loop over lags to compute rate.
- 660: Same for overall background rate (counting all events and livetime regardless of lag).
- 672: sort events.
- Stop here; will continue on next call.
- SB: Who to ask minor questions about calibration tests? FS and VR.