Cmtj

- Dipole interaction added to the `SB Model`
- Kasdin 1/f noise generator added to the `noise` module and to the solvers
- reworking the solvers for better performance and stability
- added a simple noise model to the `utils` class. It exists outside standard simulation procedures.
- added LLGB bindings and code. The solver is still WIP and doesn't integrate with more advanced features yet.
- added aliases for `ScalarDriver` -- for example, instead of calling `ScalarDriver.getConstantDriver`, you can now call `constantDriver` directly to create a constant driver.

- Adding a basic optimisation script in the `optimization` module.
- Streamlit optimization updates.

- Adding new, dynamic symbolic model compatible with `Solver` class. It is now possible to use the `Solver` class with `LayerDynamic` to solve the LLG equation.
- Added tests for procedures and operators.
- Added missing operators for `CVector` class in Python.
- `CVector` is now subscriptable.
- Added new `CVector` and `AxialDriver` initialisations.
- `VSD` and `PIMM` procedures accept additional new parameters.
- Added some optimization utilities like `coordinate_descent`.
- Added a `streamlit` service for an example PIMM simulation.

- Added new `ScalarDrivers` -- Gaussian impulse and Gaussian step.

- `CVector` got extra functionality in Python bindings. Operators are now supported.
- Domain Wall dynamics is now also for 2 layer systems. Added edge potential.
- SB model generalised for N layers.

- Adding DW dynamics 1D model with dynamic drivers. (Numba optimised)
- Adding SB model for energy-based FMR computation. Gradient computed using Adam optimiser.
- Moving resistance functions from `utils` to `resistance`
- Introducting docs updates for tutorial notebook (dark/light toggle works now).
- Reservoir computing is now exposed in Python in the `reservoir` computing module.