Section metrics don't work on EB data

[sachaMorin]

Partitioning the manifold leads to locally constant time labels because of the lack of time granularity. This raises errors when computing correlations and MI.

[sachaMorin]

We're no longer using this metric, nor EB data