[13-1, 14] Change definition of scaling factor `alpha` , add center and approximations for Hermite functions, add Fourier transform utility functions

[MothNik]

This pull request

💥 Breaking changes

• transitions the definition of the scaling parameter alpha of the dilated Hermite functions from x / alpha to x * alpha which is not backwards-compatible
• updates all documentation and test files for the new definition of alpha
• renames hermite_functions._interface to hermite_functions._func_interface

🦾 Features/Enhancements

• adds the module fourier_transform._fft_utils (with test suite) for an object-oriented interface to Fourier transforms as well as easy conversions between discrete and continuous Fourier transforms
• implements support for all kinds of x, n, alpha, and x_center for the Hermite functions by making the type checking less rigid (especially worth to mention are Pandas Series and Python Array support for x)
• makes tests for type handling of input validation for Hermite functions more dynamic and safer
• adds an x_center to the Hermite functions to shift them in x-direction and adds it to documentation and tests

• adds hermite_functions._approximations (with test suite) that offers 3 functions to easily approximate the positions of

  • the largest zero
  • the largest extremum
  • the numerical fadeout point

  of the first 100 000 Hermite functions together with scripts (developer-only (see ✍️ Documentation)) to generate the reference data

✍️ Documentation

• adds a custom Matplotlib style to make the plots look more consistent
• updates the content of the example plots to be more informative
• puts proper titles on the example plots and aligns their visual appearance
• gives example plots better names that include the index of the example they are from
• adds a section to the README that explains the new _approximations module
• made all plot saves in example scripts developer-only and added a note to the README that they need to be enabled by setting an environment variable

🪛 Internal changes

• removed redundant Numba-jitted code for Hermite functions and made Numba jit-compile the existing NumPy-function instead
• moves input validation of Hermite functions to hermite_functions._validate
• makes input validation of Hermite functions smarter in terms of type handling

[codecov[bot]]

Codecov Report

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

Files	Coverage Δ
...rc/robust_hermite_ft/fourier_transform/__init__.py	100.00% <100.00%> (ø)
.../robust_hermite_ft/fourier_transform/_fft_utils.py	100.00% <100.00%> (ø)
...rc/robust_hermite_ft/hermite_functions/__init__.py	100.00% <100.00%> (ø)
...st_hermite_ft/hermite_functions/_approximations.py	100.00% <100.00%> (ø)
...st_hermite_ft/hermite_functions/_func_interface.py	100.00% <100.00%> (ø)
...rmite_functions/_hermite_largest_extrema_spline.py	100.00% <100.00%> (ø)
...hermite_functions/_hermite_largest_roots_spline.py	100.00% <100.00%> (ø)
...obust_hermite_ft/hermite_functions/_numba_funcs.py	100.00% <100.00%> (ø)
...obust_hermite_ft/hermite_functions/_numpy_funcs.py	100.00% <ø> (ø)
...c/robust_hermite_ft/hermite_functions/_validate.py	100.00% <100.00%> (ø)