SFF Corrector Pages

[ashpags]

Hello! First, thank y'all for all of the helpful tutorials and documentation for Lightkurve -- as someone who is using both Lightkurve and Kepler data for the very first time, they have been invaluable! My undergrad research student and I are working with some K2 data and trying to replicate the SFF correction done by Vanderburg & Johnson 2014 to remove the photometric variations caused by the spacecraft drift in some short-cadence data. (We can't find anywhere this is already done by a pipeline for short-cadence data, just long-cadence. But if it exists, we'd love to know about it!) When we run Lightkurve's built-in SFF corrector, those variations are still there, so we were going to try to replicate the V&J steps by hand. There are two "further detail" pages linked in the SFF Corrector tutorial, but neither page exists anymore. They are https://docs.lightkurve.org/tutorials/04-replicate-vanderburg-2014-k2sff.html and https://docs.lightkurve.org/tutorials/04-replicate-vanderburg-2014-lightkurve.html. I found archived versions on the Wayback Machine, but I just wanted to check and see if there's a reason those pages don't exist anymore, if there are known errors, or if there's some other reason we shouldn't use them. Thank you in advance!

[orionlee]

Tutorial 04-replicate-vanderburg-2014-k2sff.html was removed as part of the lightkurve v2 documentation revamp in PR #926 . So maybe @barentsen could shed some light?

[barentsen]

@ashpags Good question! The Jupyter notebook you mention was part of the Lightkurve v1 docs. You can view an archived version here:

https://github.com/lightkurve/lightkurve/blob/v1.11.3/docs/source/tutorials/04-replicate-vanderburg-2014-k2sff.ipynb

This notebook did not make it into the Lightkurve v2 docs yet because it does not directly demonstrate a Lightkurve feature. It does of course provide helpful background information to help understand the inner workings of the SFFCorrector class, so I'd be open to see it re-instated. There is nothing wrong with it, though beware that it does not provide a complete explanation of the internals of SFFCorrector as it is implemented right now.

Note: the SFF algorithm is more difficult to use on short-cadence data (for reasons I do not remember right now). I'd be very interested to hear about any and all experiences in using it on K2 short cadence data. Alternatively, the PLDCorrector may work well out of the box for such data.

Does this help?

[ashpags]

Yes, that helps! Thank you! And yes, we're having issues with the SFF algorithm on short-cadence data, but have some more playing around to do first before we declare it a lost cause. We'd just really like to be able to use that short-cadence data but without removing the jitter from the spacecraft motion, it's not super useful. We'll figure something out though! Thank you again!