Gaussian structure learning

[pawel-czyz]

This PR adds the utilities for learning the precision matrix in Gaussian graphical models using block Gibbs samplers with two types of prior:

1. The continuous spike-and-slab prior from H. Wang, Scaling it up: Stochastic search structure learning in graphical models (2015)
2. The graphical horseshoe prior from Y. Li, B.A. Craig and A. Bhadra, The graphical horseshoe estimator for inverse covariance matrices (2019)

List of changes:

• The continuous spike-and-slab sampler.
• The graphical horseshoe sampler.
• Quarto notebook with an example.