Skip to content

Repeated spline points

In some runs I have there are repeated spline points printed to the spline chain files. BayesWavePost crashes when there are repeated spline points with the error:

gsl: interp.c:83: ERROR: x values must be strictly increasing
Default GSL error handler invoked.

And will exit with signal 6. It looks like the MODEL_spline_IFO.dat files are being printed out with only 2 decimal points (or fewer) of accuracy. This will mean that if spline points are close together, then they will be printed with the same value, meaning we will have repeated splines. The issue itself probably lies either there or somewhere in the actual spacing of the spline points.

I have a python script /home/sophie.hourihane/find_repeat_spline.py that will find repeated spline points. This is happening for example here:

/home/sophie.hourihane/public_html/glitch_events/Blip/Blip2/GW150914/September/Blip2_GW150914_events01234/trigtime_1165578732.447999954_0.0_0.0_2/chains/ACTUAL_clean_spline_H1.dat.0

Update: This issue can just be fixed by printing out doubles with more accuracy. C only prints doubles by default to 6 characters.

The call python /home/sophie.hourihane/find_repeat_spline.py /home/sophie.hourihane/public_html/glitch_events/Blip/Blip2/GW150914/September/Blip2_GW150914_events01234/trigtime_1165578732.447999954_0.0_0.0_2/chains/ACTUAL_clean_spline_H1.dat.0 returns

1 Matches found on line 24224 from 236-261: 1008.06 -94.0252 1008.06
Done looking! found 1 repeated splines

Here is an example with 5 repeated spline points

[sophie.hourihane@ldas-pcdev1 f_Blip2_GW150914_events01234]$ python /home/sophie.hourihane/find_repeat_spline.py /home/sophie.hourihane/public_html/glitch_events/Blip/Blip2/GW150914/September/f_Blip2_GW150914_events01234/trigtime_1165578732.447999954_0.0_0.0_0/chains/ACTUAL_clean_spline_L1.dat.0
1 Matches found on line 27992 from 267-292: 1020.02 -96.8681 1020.02
1 Matches found on line 27993 from 266-291: 1020.02 -96.8688 1020.02
1 Matches found on line 27994 from 265-290: 1020.02 -96.8688 1020.02
1 Matches found on line 27995 from 264-289: 1020.02 -96.8671 1020.02
1 Matches found on line 27996 from 265-289: 1020.02 -96.867 1020.02
Done looking! found 5 repeated splines
[sophie.hourihane@ldas-pcdev1 f_Blip2_GW150914_events01234]$

Likely this problem would start in BayesLineInitialize in BayesLine.c

Edited by Sophie Hourihane