Move to Interpolations.jl

[Sbozzolo]

Note that the order of interpolation is now first instead of cubic.

Is this acceptable?

Closes #43
Closes #39

[charleskawczynski]

Hmm, the docs are breaking, and the tests have all sorts of issues. Maybe we can fix the tests first.

[Sbozzolo]

Hmm, the docs are breaking, and the tests have all sorts of issues. Maybe we can fix the tests first.

What issues do you see in tests? Aren't they all passing?

[Sbozzolo]

The error in the docs is because by default Interpolations.jl does not allow extrapolation, and something is trying to evaluate a point outside of the domain of defition of the interpolant.

[charleskawczynski]

What issues do you see in tests? Aren't they all passing?

Sorry, I meant that the design of the tests in this repo need improvement-- Right now, the things that are tested are implicit, based on names(AtmosphericProfilesLibrary; all = true) / getproperty(AtmosphericProfilesLibrary, name) of those names, which requires ignoring/skipping names-- it's way more complicated than it needs to be. Also, the inference tests are just wrong (@test @inferred Float32 iscallable(prof(Float32))).

[charleskawczynski]

I think #45 fixes all of this.

[charleskawczynski]

Sorry about all the conflicts, I can rebase if you'd like

[charleskawczynski]

Rebased

[charleskawczynski]

Results in the docs look visually indistinguishable.