[newchem-cpp] factor out scale_fields_dust

[mabruzzo]

This PR was originally proposed as brittonsmith#41

---

This should be reviewed after #385 has been merged

---

This factors out some the logic from calc_tdust_3d for scaling densities from comoving form to "proper" form (and back again). This consists of 3 parts:

1. I renamed the file that already held scale_fields_table (so that the new function could be put into the files)
2. I actually factored out the function
3. I did a little cleanup of calc_tdust_3d.

All in all, this is pretty simple.

Note

This PR probably changes the precise answer of the answer-tests (I had to convert a division to multiplication by an inverse). In other words, the answers drift a tiny, nominal amount. Back when I originally proposed this change (as brittonsmith#41) all of the tests still passed