user_pparameters.C

The pipeline has a default parameter file called cwb_pparameters.C which contains the list of needed variables for the analysis set at a certain value.
The user can modify some of these values adding them in the config/user_pparameters.C file.
We report here all the variable contained in the cwb_parameters.C file, the user can change each of them accordingly to his/her preferences.

This is the complete file:

We divide the file in different sections for simplicity:

Frequency bands  
Labels  
Rho settings (only for Background analysis)
Thresholds  
Data Quality and veto  
Benchmark parameters  
Others  

Frequency bands

The burst analysis is generally divided into 3 main frequency ranges,

  • [64; 200] Hz;
  • [200; 2048] Hz;
  • [1600; 5000] Hz.
Usually production stage is made into two bands (less and more than 2048 Hz), but postproduction is performed separately for each frequency band.
To select one frequency band just use one of the following:
#define PP_64_200_HZ     //for [64; 200] Hz
#define PP_200_2048_HZ   //for [200; 2048] Hz
#define PP_1600_5000_HZ  //for [1600; 5000] Hz
Only one choice should be selected.
If no choice is selected, results show the complete frequency band analysed in the production stage.

Moreover, it is possible to discard small frequency bands (for example to cut a persistent line). At this purpose, there are the two arrays lowFCUT and highFCUT.

int    nFCUT = 0;           // number of frequency cuts
double lowFCUT[100];        // low frequency cut
double highFCUT[100];       // high frequency cut

Example This code:

nFCUT = 2;
lowFCUT[0]=150;
lowFCUT[0]=151;
highFCUT[1]=235;
highFCUT[1]=242;

discard event with central frequency inside the two frequency bands: [150,151] and [235,242].

Rho settings

For Background results, the rho threshold is never applied, because it is the principal variable for calculating False Alarm Rate. However, there are some variables that defines the output behaviour of the results:

#define nRHO 150 //defines the range of the RHO statistic to be used in plots
double pp_rho_max = 10.;
int pp_max_nloudest_list = 25; // maximum number of loudest event in the report list
float T_out = 3.5;
  • nRHO and pp_rho_max define the rho ranges of rate vs rho figure and text files.
    Rho min is usually 2 for 1G and 4 for 2G.
    Rho step is usually 0.1 for 2G and 0.2 for 2G.
    Rho max = Rho min + Rho step * nRHO.
    The figure is zoomed at the value of pp_rho_max if it is less than Rho max.
  • pp_max_nloudest_list is the number of triggers reported at the bottom of web page.

  • T_out: Threshold on rho for the final trigger list

Labels

These are labels that appear in the web page:

#define RUN_LABEL ""
user_pp_label = "";
  • RUN_LABEL: Subtitle on the page
  • user_pp_label: postfix on the name of the directory containing the figures

obsolete:

pp_label = "";
char title[256]="";
char subtitle[256]="";
char  RunLabel[256]="";

Thresholds

This is a collection of thresholds on the trigger parameters:

  • network correlation coefficienct: netCC[0] > T_cor
  • subnet: netCC[3] > T_scc
  • energy disbalance: netED[0] < t_vED
  • penalty: penalty > T_pen

Only for simulation phase:

  • Amplitude correlation coefficient: rho[pp_irho ] > T_cut
  • Injection selection: abs(time[0]-time[nIFO]) < T_win
// obsolete parameters
T_cut_veto = 6.9;           // rho high frequency cut ...
T_CUT_veto = 6.9;        // rho low frequency cut ....

T_acor     = 0.0;        // output threshold
T_hrss     = 0.0;        // penalty threshold
i_hrss1    = 0;
i_hrss2    = 1;

T_rms      = 0.;         // penalty threshold

ratio_hrss  = 1;          //CORRECTION FREQUENCY
pp_freq_offset = 0.;      //CORRECTION FREQUENCY  (SIM)

Data Quality and veto

It is possible to apply Data Quality and Veto to reconstructed triggers. This is done using the same syntax as user_parameters.C. The only difference that the user can decide which DQ/Veto to apply:

#define CAT2_VETO
#define HVETO_VETO
#define CAT3_VETO
#define PEM_VETO

 // {ifo, dqcat_file, dqcat[0/1/2], shift[sec], inverse[false/true], 4columns[true/false]}
  int nDQF = 0;
  dqfile DQF[20];

Selecting one or more of the define above, it is possible to apply DQ/Veto, according to how the file is defined in the user_pparameters.C, i.e. the 4th column, for instance:

{"V1", "input/S6A_OFFLINE_V1_DQCAT2SEGMENTS.txt", CWB_CAT2,  0., true, false},
{"V1", "input/S6A_OFFLINE_V1_DQCAT3SEGMENTS.txt", CWB_CAT3,  0., false, false},
{"V1", "input/V1_KW_HVETO_S6A.txt",               CWB_HVETO, 0., false, false},

Note: The application is different between CAT2 and the others. If a trigger is discarded by CAT2, it is deleted from the list, however, if the trigger is discarded by CAT3/HVETO/PEM, it is mantained in the list, just flagging it. See data quality for details on how to write the data quality

Benchmark parameters

There is the possibility to produce an integrated page with informations on finished/unfinished job and histograms of elapsed time and memoery usage of ech job, as produced by the command cwb_condor benchmark

pp_jet_benckmark = 0;
pp_mem_benckmark = 0;
  • pp_jet_benchmark and pp_mem_benckmark are related to the x-axis of the two histograms.
    If set to 0, the integrated page is not produced.
    If set to a value greater than 0, this is used as maximum value for x-axis.
    If set to a value less than 0, the maximum value of x-axis is automatically set by the pipeline

Others

Some additional parameters :

  • hours: Bin size (in hours) for rate vs time plot

  • pp_eff_vs_thr: Used for ROC calculation

  • pp_show_eff_fit_curve: enable/disable draw efficiency fit curves

    * true (default) : show fit curve, eff@hrss10/50/90 are derived from fit
    * false          : show lines connecting points, eff@hrss10/50/90 are not showed
    
  • pp_sreport: array used to add auxiliary reports to the current report

    * --link LINK   : www link of the auxiliary report
                      --link <tab> open a sub tab
                      --link </tab> close sub tab
    * --label LABEL : report label - used in the html tabs
    * --high HIGH   : report window height (optional : def=4000)
    * --name NAME   : report html name (optional : def = "header.html, body.html")
    
  • pp_fad: enable/disable FAD report (default=”” -> disabled)

    * --bkgrep  : directory of the background report
                  - used to lacate rate_threshold_veto.txt and live.txt
    * --hrss    : - if it is a number then it is used as normalization constant for all MDC types (def=0)
                  - if =0 then hrss rescale is not applied
                  - if it is a file then it is the list of hrss used for each MDC type
                    format : for each line -> hrss
    * --gfit    : true/false (def=false)
                  - enable/disable fit in the output plots
    * --rhomin  : minimum rho value selected for plots (default = 5)
    * --nzbins  : integer number [<0, 0, >0] - (def=0)
                  - if nzbins=0 the standard FAD statistic is used
                  - if nzbin>0 the FAD statistic is computed until there are nzbins
                    consecutive bins with zero events inside
                  - if nzbin<0 the FAD statistic is computed
                    with classical FAD and min-hold (un-biased FAD)
    * --units   : K/M (def=M)
                  - K -> Kpc,Kyr : M -> Mpc,Myr
    * --distr   : formula/mdc  (def=MDC)
                  - radial distribution is computed from formula or from mdc injections
    * --nbins   : number of bins used in hist to computed
                  the radial distribution from the mdc injections
    * --header  : true/false  (def=false)
                  - if true -> add cwb header to fad html file
    * --multi   : true/false  (def=false)
                  if true -> FAD multi plot for each mdc set are created and substituted
                  in the sim report page to the eff_freq plots (def=false)
    
ROC are calculated for a unique factor of each waveform. The factor can be automatically calculated from the pipeline (the one nearest fo hrss50%) or set by the user.
Example:
pp_eff_vs_thr = "--mode auto --rho_min 6.";

set the automatic selection of the factors (setting a minimum of 6 in rho).

pp_eff_vs_thr = "--mode eff_threshold_factors.txt";

set the amplitude factors defined in the eff_threshold_factors.txt file