Typo in Q-q in BWPost
In anderson_darling_test in BayesWavePost, the section that makes MODELNAME_residual_P-p_IFO.dat
, 'samples' is being indexed from [0, Nsamp), but samples holds complex data (of length 2 * Nsamp) where the even numbers index the reals, and odds the imaginary. gsl_cdf_ugaussian_P(samples[n]) is thus making a unit gaussian cdf of some mixture of the 1st half of the real and imaginary FFT data, which I do not believe is meant. Link to problem line here: