$\alpha_{5000}$ and line opacity in M stars

[ajwheeler]

In Korg, by default, the formal solution (radiative transfer) uses $\tau{5000}$, the optical depth at 5000 Å. In order to do this it must calculate the absorption coefficient at the same wavelength, $\alpha{5000}$.

At present, it ignores contributions to $\alpha_{5000}$ from lines. This is an unproblematic choice, except for M stars, where you get line blanketing across the visible.

I think the right way to handle this is to use the line opacity when it's available (when the linelist encompasses 5000 Å), and to not worry about it (for now) when it isn't. Plots to come.

[ajwheeler]

As with many detail of how RT is done (see #257), this impacts the continuum more than the rectified flux. (It's defintiely still important for both though. (And, as a side note, it's interesting to see that RT choices in general are crucial for cool stars, easily at the multi percent level.)

At slightly warmer Teff, things are less dicey, but still a but dicey:

[ajwheeler]

Here's what this looks like in the IR:

[ajwheeler]

Closed by #313