Add T computation for Ti and Zr in Rutile and zircon, from Ferry and Watson, 2007

[Iddingsite]

Hi,

I needed this so I've implemented the functions that gives temperature from Zr and Ti in zircon and rutile (only the ones giving Ti and Zr were implemented before).
Tests were also added.

Just a more general question, is there a reason why the package Measurements.jl is not used in this package? I think it would be a nice feature to have uncertainty propagated from this kind of functions.

I was thinking about something like that to not break anything:

function Ferry_Ti_in_zirconT(Ti::Number, aSiO2::Number, aTiO2::Number; uncertainty=false)
    if uncertainty
        1 / ((5.711±0.072) - log10(aSiO2) + log10(aTiO2) - log10(Ti)) * (4800.0±86) .- 273.15
    else
        1 / ((5.711) - log10(aSiO2) + log10(aTiO2) - log10(Ti)) * (4800.0) .- 273.15
    end
end

What do you think?

Also, I have local tests failure unrelated to this PR due to the test of Perplex on line 245 of testGeochemistry.jl (I am on Mac). That would probably be fixed by using the Perple_X_jll at some point.

Cheers,

[brenhinkeller]

Thanks! Seems like a good add!

I'd be happy to have a Measurements.jl dependency added, and dispatching on that seems like a nice way to do it

The Mac Perplex test failures are probably just because the package currently only knows how to automatically install perplex on linux, so one has to manually install perplex in the right location otherwise. That's a bit of a pain but could be solved if we manage to switch to using perplex_jll (#28)

[Iddingsite]

Ok, I will give it a shot next week then. Have a good wk-end!