Refactor to directly support higher order derivatives

[pulsipher]

This refactor makes it so we can store derivatives of arbitrary order instead of only using 1st order derivatives that are nested for higher orders. Additions/fixes include:

• Derivative objects now store order to be any order desired
• Where possible, InfiniteOpt will unnest derivatives (e.g., deriv(deriv(y, t), t) == deriv(y, t, t))
• Finite difference schemes of arbitrary order are supported
• For derivative methods that do not directly support higher order derivatives, InfiniteOpt will automatically reformulate them into nested 1st order derivatives
• Printing now distinguishes between ordinary derivatives and partial derivatives
• Fixes unnoticed bug with derivative deletion that left deleted derivatives inside constraints

Closes #197

[codecov[bot]]

Codecov Report

All modified and coverable lines are covered by tests :white_check_mark:

Project coverage is 99.74%. Comparing base (f1b41c2) to head (0d7748e).

Additional details and impacted files

```diff @@ Coverage Diff @@ ## master #341 +/- ## ======================================= Coverage 99.74% 99.74% ======================================= Files 36 36 Lines 6589 6670 +81 ======================================= + Hits 6572 6653 +81 Misses 17 17 ```

:umbrella: View full report in Codecov by Sentry.
:loudspeaker: Have feedback on the report? Share it here.