ATClean appears to download incorrect data for large batch requests

[jackstrophysics]

Hi there!

I have been using ATClean to mass-download lightcurves for Type Ia supernovae from the ATLAS database. I recently noticed that some (although I'm not sure how many) lightcurves downloaded from batch requests using:

./download.py <long list of supernovae> --sninfo_file <sninfo_file> -l 500 -o

are not downloading the correct lightcurves as shown in the ATLAS database. As an example, here is the lightcurve that was downloaded and cleaned for 2021nyt as part of the batch download:

whereas downloading and cleaning the lightcurve as an individual request using:

./download.py 2021nyt --sninfo_file <sninfo_file> -l 500 -o

yields the following lightcurve:

which much more clearly matches the forced photometry data (https://star.pst.qub.ac.uk/sne/atlas4/candidate/1162313711262836100/).

I have used the exact same sninfo_file both times, also attached here:
sninfo_100_to_200_Mpc.txt

This is just one example from the batch request performed using this list, but there were many instances where it would download and fail to clean a lightcurve, but then successfully clean it after downloading the data individually.

Is the list of requested supernovae too long such that the server is getting confused? I'm not too sure what is happening here.

Thanks!

[srest2021]

Hi, thank you for bringing my attention to this! Would you mind sharing the command(s) you used to clean and average the light curves after downloading them, and/or the cuts you applied?

[jackstrophysics]

Yes of course. My cleaning command was:

./clean.py <supernovae> -x -p -g -o

and I have not changed any of the defaults in the clean.py section of config.ini.