tickets/SP-1600: updates to pass ruff checks

[humnaawan]

No description provided.

[rhiannonlynne]

Sorry. - I should have provided more context around ephem ... I was kind of curious to see what you'd do about this particular issue, actually.
We dropped ephem (we had been using it for some almanac-style calculations of sunset and planet positions), and do not want to put it back in.
This particular metric (the star counts) doesn't run in standard use, so it's not worth us adding another dependency to support this particular conversion between RA/Dec and galactic coordinates.
It's also not clear why ephem would have to be used to do this conversion (and I don't know why there are three different methods to convert between ra/dec and galactic l/b here).
In short, this is a community contribution that has bit-rotted over time.
That's the additional context ... which leaves a couple of options:

• add ephem as a dependency, keep using the ephem ra-dec to gal l-b conversion
• check the difference between other conversions and the conversion you get from ephem, and swap to one of the other conversions that is available using some other method (astropy would be my preferred option .. SkyCoord converts between ra/dec and l/b in a widely-used way) -- dropping the ephem pieces of the code
• just use one of the other conversions that's written into this piece of code .. my take on what is provided here is that probably one of the other methods was used first, then the contributor found that ephem provided a more accurate conversion and switched. However, I did check the conversion using the analytic method here (that doesn't use ephem) against astropy's conversion, and it's at most a few degrees incorrect. Which probably doesn't make that much of a difference in the metric output .. especially for a metric we're not actually running. So we could drop the ephem conversion and just use the analytic method and say it's close enough.
• just drop the entire piece of code. I didn't do this in the first place, because having an analytic way to calculate stellar density and distribution over distance at any point in the sky seemed useful .. but tbh, I don't know how accurate it is, and other metrics don't use it. And probably wouldn't use it, without further checking and validation. So I suspect, in the end, this is likely the best option.

[rhiannonlynne]

I don't want to add ephem back into the dependencies, so we should probably either deprecate all of the star_counts code that needs it, or move to a different ra/dec to galactic l/b conversion that doesn't depend on ephem.

[humnaawan]

regarding whether to keep ephem or not, would it helpful to contact the contributor (seems like its Mike Lund)? i wonder if they are running the metric, even if its not run in the regular use.

[rhiannonlynne]

regarding whether to keep ephem or not, would it helpful to contact the contributor (seems like its Mike Lund)? i wonder if they are running the metric, even if its not run in the regular use.

Eek, I thought I'd replied to this. Yes, it's very reasonable to contact Mike Lund.

[rhiannonlynne]

Looks good! Thanks Humna.