Robustica

- made `scikit-learn-extra` dependency as optional to avoid incompatibilities
- updated default `FastICA` parameters to follow `scikit-learn>=1.3`
- added `pyproject.toml` to keep compatibility with `pip>=25.0`

- added verbose option to RobustICA class
- bug fix: fixed reproducibility seed
- updated documentation

- bug fix: sign correction before computation of standard deviations of centroids
- added web documentation

This first release of `robustica` contains:
- three modules:
- `RobustICA`: contains the class and functions to perform robust independent component analysis (ICA)
- `InferComponents`: contains the class to infer how many components should one consider for ICA
- `examples`: contains helper functions to quickly create examples
- tutorials
- documentation