Add gaia passbands

[segasai]

It addresses the #40

[dkirkby]

Thanks @segasai.

I checked the filter curve values and it looks like they need a bit of cleaning up:

• Some wavelengths have transmission values of 99.99. I suspect these lines should just be removed, unless there are special instructions about this in their documentation.
• There are some transmission values < 0. These should probably be dropped (in the tails) or set to zero (in case this occurs within the filter support).
• There are some long tails with very low transmission values. If these are consistent with zero, within the transmission measurement errors, I would drop these to avoid errors when convolving with a flux that is not specified over these long tails. It looks like a threshold around 1e-6 would be about right.
• Each curve should have a single zero transmission point at each end, for a well defined linear interpolation model over the full filter support.

[segasai]

Thanks, I've fixed it. Should I put the python script that preprocess the official curves into these .ecsv somewhere in the repo ?

[dkirkby]

Thanks, this is the result of:

gaiadr2 = speclite.filters.load_filters('gaiadr2-*')
speclite.filters.plot_filters(gaiadr2)

Can you confirm that the abrupt edge in G around 10500 is desired?

[dkirkby]

@segasai Can you allow me to push to your fork so I can add this plot to the sphinx docs?

[dkirkby]

It looks like G goes smoothly to zero around 11,000A in the plot here.

[segasai]

It was a consequence of me cutting by 3sigma uncertainties. I'm updating that. I think I'll do 1 sigma

[segasai]

I've also added you @dkirkby to my repo, so you can make commits there

[dkirkby]

Looks good now:

I added this plot to the docs and I think everything is ready to merge now.

[segasai]

Thanks for merging/help!

[dkirkby]

I just tagged v0.13 and pushed the new version to pypi so this can be easily incorporated into the next pipeline release.

[dkirkby]

Updated v0.13 docs are here.

@sbailey Note that this release also contains updated (DR1) DECam filters.

Thanks for the contribution @segasai!