Add support for Unitful.jl numbers to MonotonicInterpolations

[MatFi]

Now one can do:

using Unitful
using Interpolations

x=-1u"s":0.25u"s":1u"s" ;
y=sign.(0.1u"s".+x)u"m";
itp = interpolate(x, y, FiniteDifferenceMonotonicInterpolation());
julia> itp(-0.125*u"s")
0.0 m

The reason why Unitful.jl did not work here before is that all polynomial coefficients were of the same type TCoeffs. When using Unitful numbers, the different dimensions of the coefficients become relevant. Therefore every coefficient now has its own type.
This pull request additionally contains a fix of a erroneous tangent calculation in FritschCarlsonMonotonicInterpolation.
(Be warned, I have been using julia only since version 1.0, so this may not be implemented optimally)

[timholy]

Sorry for the lack of review. If you're interested in a bigger role (see #365) I'll be happy to review this.

[MatFi]

I'm honored and I'd love to, but giving a hack like me even a shred of responsibility for this package would be the end of it. I think the best way to give something back to the community is to stay on the ground where I'm reasonably proficient in what I do. Of course I like to help out where I can.

[timholy]

Sounds good. I'll leave this for my successor to review as good practice! Sorry for the likely delay.

[MatFi]

No problem, I'm glad to hear that this PR becomes useful even before the merge

[mkitti]

Successor in training here. Could we develop some tests and add Unitful as a test dependency to ensure integration going forward? Also see #367

[timholy]

Note that this does add some Unitful tests (see xa and y). Unitful is already in the test deps, see https://github.com/JuliaMath/Interpolations.jl/blob/master/Project.toml

[timmyfaraday]

Is this feature available in the latest release, as I'm not able to recreate this behavior.
If not, any indication on when would it be available?

[mkitti]

Is this feature available in the latest release, as I'm not able to recreate this behavior.
If not, any indication on when would it be available?

This branch as not been merged yet. It's a high priority at the moment since @timholy already looked at it. I'll take a look this weekend and hopefully merge within two weeks.

[mkitti]

I intend to merge this and eventually release as v0.13.1