Issue with 'finalfit' ignoring some modes (sometimes)
Closed this issue · 1 comments
I've tested running the pipeline with the other provided lightcurve '009332840.dat' as well as some power spectra from kepseismic (https://archive.stsci.edu/prepds/kepseismic/) and noticed that there seems to be an issue after the last peaks_mle
step that includes the finalfit option, taco.peaks_mle(pds_bgr, peaks_mle, data, finalfit = True)
.
It seems that before this step, most of the obvious peaks in the power spectrum are identified (I'm getting 40 identified peaks after running peaks_mle, data = taco.peaks_mle(pds_bgr, peaks_mle, data, removel02 = True, mixed_peaks = mixed_peaks)
) but after running the last step, only two peaks are identified (see screenshot). I tried renaming the 'peaks/peaks_mle' dataframes after each step, in case overwriting them was causing issues, but this did not fix the problem with finalfit. I think this problem is reproduceable if you just replace ts_raw = pd.read_csv('../data/001296068.dat', comment = '#', header = None, delim_whitespace = True)
with ts_raw = pd.read_csv('../data/009332840.dat', comment = '#', header = None, delim_whitespace = True)
in the pipeline.ipynb file and then run all the cells. It's strange because on some input lightcurves (like 001296068.dat) or input power spectra, everything works fine, as far as I can tell, and sometimes the finalfit step does this.
Thanks!
There was an issue in the legacy bash pipeline. I will correct it soon.