Added multi-layer functionality to TidalPy

[jrenaud90]

Scripts based on 0.2.x will likely break with this version!

• Major Changes
  • Added the first iteration of a multilayer tidal calculator module in TidalPy.tides.multilayer this module provides basic functionality to calculate tidal dissipation in a semi-homogeneous, shell-based approach. This is more accurate than the pure homogeneous model used throughout the rest of TidalPy. The downside with the current version is that it does not allow for NSR or high eccentricity / obliquity. A future version will attempt to add in a more robust Tidal Potential equation which will allow for addtional physics.
  • Setup.py has been revamped as has the installation process. This is in prep to allow for TidalPy to become available on PyPI.
  • Did away with all of the _array functions. Found a way for njit to compile a function to handle either arrays or floats.
    • Left the self._func_array (in addition to self._func) in the model.py classes just in case we ever do need to define array functions in the future: all the infrastructure is still in place.
  • Added a numba-safe Explicit Runge-Kutta integrator. This is fully wrapped in njit'd functions.
    • On its own this can be 5--20 times faster than scipy.solve_ivp.
    • This also allows the integration function to be used from within another njit'd function(s).
• Minor Changes
  • Numerous bug fixes.
  • Removed the array versions of the dynamic functions.
    • use_array is still tracked in OOP and some quick calculation functions. These may all be not necessary now.
  • New cookbook to showcase the multilayer calculations.
  • Added surface area slices to base physical class.
  • Fixed some issues with how radius slices are tracked within layers and worlds.
    • <world/layer>.radii[0] is never the radius at the bottom of the object, it is always one dx up.
    • An interpolation had to be added to Burnman layers since Burnman radii starts at 0.
  • Improved various docstrings.
  • Refactored the TidalPy.tools to TidalPy.toolbox.
  • Refactored Cookbooks to Demos.
  • conversions.semi_a2orbital_motion and orbital_motion2semi_a now always return np.nan where they used to return complex numbers.

[codecov[bot]]

Codecov Report

Merging #27 (750a8fa) into master (6e9e6d5) will increase coverage by 13.97%. The diff coverage is 85.32%.

@@             Coverage Diff             @@
##           master      #27       +/-   ##
===========================================
+ Coverage   60.79%   74.77%   +13.97%     
===========================================
  Files         168      216       +48     
  Lines       12394    15238     +2844     
===========================================
+ Hits         7535    11394     +3859     
+ Misses       4859     3844     -1015     

Impacted Files	Coverage Δ
TidalPy/cooling/__init__.py	87.50% <ø> (ø)
TidalPy/dynamics/__init__.py	100.00% <ø> (ø)
TidalPy/tides/methods/global_approx.py	89.20% <ø> (+10.79%)	:arrow_up:
TidalPy/utilities/classes/config/json_utils.py	23.52% <ø> (ø)
TidalPy/utilities/graphics/planet_plot.py	75.00% <ø> (ø)
TidalPy/toolbox/multilayer/liquid_solid.py	9.81% <9.81%> (ø)
TidalPy/utilities/integration/helpers.py	31.81% <31.81%> (ø)
TidalPy/utilities/math/special.py	38.88% <38.88%> (ø)
...ltilayer/numerical_int/initial_solution_dynamic.py	55.77% <55.77%> (ø)
TidalPy/toolbox/multilayer/homogen_solid.py	57.14% <57.14%> (ø)
... and 189 more

---

Continue to review full report at Codecov.

Legend - Click here to learn more Δ = absolute <relative> (impact), ø = not affected, ? = missing data Powered by Codecov. Last update 6e9e6d5...750a8fa. Read the comment docs.