add computation of smoothness for graph signals

[hugofluhr]

I want to add a function the computes the smoothness of a signal using the graph laplacian.

Detailed Description

compute the smoothness of any signal defined at each node of the graph using the graph laplacian.

Context / Motivation

Smoothness is one of the basic metrics of GSP and is a good starting point to add other GSP operations to nigsp and start contributing to the package.

Possible Implementation

Straightforward matrix multiplication to get the quadratic form.

[hugofluhr]

I'm not sure how to add the smoothness computation to objects.py as it is assumed there that operations are applied to timeseries and not static signals. Any input on how you want to handle this is welcome.

[smoia]

I would add a function with mtx, signal as inputs (so no SCGraph) in either timeseries or metrics.

You can also add it then as an object method in this way (until we find a better way to do it):

nigsp/nigsp/objects.py

Line 156 in 4ffa7cb

def nodestrength(self, mean=False): # pragma: no cover

And yes, we'll start renaming timeseries as signals! ;)