Megan Bedell

For the moment I've added a long docstring to `make_data.write_data` to specify what things are & which ones are optional. Not an ideal solution but hopefully it helps!

maybe incorporate continuum fitting into the model itself with a GP and forget about initial normalization??

I'm marking this as low priority right now because I think in general the BERVs calculated by the instrument pipeline will be better than what we can get ourselves. This...

ah ok, maybe a little bit down the line we want to build a GP component into the model and let it handle all continuum fitting (including the initial normalization...

I started looking into this a bit and can confirm (a) this is an issue and (b) it's happening somewhere within the astroquery MAST query-- if I run the following...

Some further digging... this isn't an astroquery bug, it happens at the MAST level. Searching the [MAST Portal](https://mast.stsci.edu/portal/Mashup/Clients/Mast/Portal.html) for the KOI, its coordinates, or even the KIC all return both...

I also cannot reproduce this bug using the latest version (lightkurve 2.4.0)

Thanks, this is useful info! Currently I take every pixel that has unnormalized flux count < 1 and set it to 1, along with setting the inverse variance on that...