search

StingraySoftware / stingray

Anything can happen in the next half hour (including spectral timing made easy)!
https://stingray.science/stingray
MIT License
166 stars 137 forks source link

Handling gaps in data for Dynamical PS extraction #803

Closed fcotizelati closed 3 months ago

fcotizelati commented 4 months ago

I tried extracting a DPS from a time series with gaps and observed that the "time" attribute of the DPS object overlooks the presence of these gaps. Could a warning or error message be implemented to notify users they're trying to extract a DPS from data with gaps? Alternatively, is it possible to account for the 'gti' attribute of the light curve during the DPS extraction?

matteobachetti commented 4 months ago

@fcotizelati thanks for the Issue! This is weird though, because DPS is based on the standard AveragedPowerSpectrum, and it does take gaps/GTIs into account. Could you provide a minimal working example showing the behavior?

fcotizelati commented 4 months ago

Hi @matteobachetti thank you! Of course, here are some steps:

lc.gti array([[ 0., 2270.], [5570., 7370.]])

dynspec = DynamicalPowerspectrum(lc, segment_size=128, norm='frac') 32it [00:00, 2205.50it/s]

dynspec.time array([ 64., 192., 320., 448., 576., 704., 832., 960., 1088., 1216., 1344., 1472., 1600., 1728., 1856., 1984., 2112., 5634., 5762., 5890., 6018., 6146., 6274., 6402., 6530., 6658., 6786., 6914., 7042., 7170., 7298.], dtype=float128)

In fact, I've realized that the problem is not so much in extracting the DPS, but in my plotting procedure: if I use the range min(dynspec.time), max(dynspec.time) to plot the power spectrum, I should mask the DPS portion that corresponds to the time intervals that are not covered by the GTI to ensure that the plot correctly reflects periods without data (e.g. using blank spaces?). So I can fix this easily on my own plotting code. But if I don't use the masks, the DPS is plotted also in the time periods with gaps, which can be misleading?