Sigmaepsilon.math

Changed

- The `BinaryGeneticAlgorithm` class was updated with a new optional `minimize` parameter to be pprovided at object instantiation. With this, users can signal if they want their objective to be minimized or maximized.
- The docstring of the `BinaryGeneticAlgorithm` class is updated to clarify that by default, the class performs optimization, unless otherwise specified by the user.

Changed

- The `LinearProgrammingProblem` class has changed. Now it uses SciPy as a solver and the class only serves as a preprocessor. Usage of the modified class should be straightforward from the documentation.
- The `MLSApproximator` class got refactored.

Fixed

- Fixed the `has_full_rank` function in `sigmaepsilon.math.linalg`
- Fixed initialization isssues related to relation classes in `sigmaepsilon.math.function.relation`.

Removed

- The `VariableManager` class was removed, although it was not for the public, it only served as a helper for the `LinearProgrammingProblem` class.

Added

- Protocols for PointData, CellData, PolyData and PolyCell classes
- Cell interpolators (7)

Fixed

- Copy and deepcopy operations (29).
- Cell class generation (36).

Changed

- Class architecture of data structures. Now the geometry is a nested class.

Removed

- Intermediate, unnecessary geometry classes.