Feat/cpp gradient interface enhancements

[curiosity54]

Implemented "get number of coefficients" for spherical expansion and invariants on the C++ side with Alex.

Tasks before review:

• [x] Add tests, examples, and complete documentation of any added functionality
  • [x] Make sure the autogenerated documentation appears in Sphinx and is formatted correctly (ask @max-veit if you need help with this task).
  • [x] Write additional documentation in the Sphinx (not docstrings!) to explain the feature and its usage in plain English
  • [x] Make sure the examples run (!) and produce the expected output
  • [x] For bugfix pull requests, list here which tests have been added or re-enabled to verify the fix and catch future regressions as well as similar bugs elsewhere in the code.
• [x] Run make lint on the project, ensure it passes
  • [x] Run make pretty-cpp and make pretty-python, check that the auto-formatting is sensible
  • [x] Review variable names, make sure they're descriptive (ask a friend)
  • [x] Review all TODOs, convert to issues wherever possible
  • [x] Make sure draft, in-progress, and commented-out code is moved to its own branch
• [x] If committing any code in Jupyter/IPython notebooks, install and run nbstripout
• [x] Merge with master and resolve any conflicts
• [x] (anything else that's still in progress)

[agoscinski]

Because the CRTP made the error messages hard to understand and I thought I just try this implementation, since we anyway don't use any other stacking for gradients (except the upcoming LODE stuff). In terms of development time this was faster than now figuring out the within structure manager solution.

The problem was that I used a const function, which works in the adaptors and structure manager collection but not in the structure, since we are not changing any member variable. But in structure manager one needs to create a non const Iterator passing this to it which changes the this object, so it cannot be const.