Feature: Graph Laplacian box

The box should compute the "Graph Laplacian" from the connectivity matrix, computed by the Connectivity Measurement box.

The Graph Laplacian consists of computing L = D - A, with A being the connectivity (adjacency) matrix, and D being the matrix whose diagonal is the sum of the rows of A.

Then, a "denoising algorithm" described by T.CATTAI & F.VICO DE FALLANI can be applied, which does a sub selection of N eigenvalues of L (N<M, M being the number of eigenvalues of L). If M are sorted from smallest to largest value, then CATTAI et al. show that the smallest or largest values (& associated eigenvectors) can be selected for different purposes of filtering/enhancing the data. Once the N eigenvalues/vectors have been selected, the matrices L' & A' can be reconstructed from this subselection.