Evaluate pyfunnel as a user-defined metric

[pstelzig]

Rationale

pyfunnel is a collection of algorithms to compare result timeseries from hybrid simulations (continuous time, discrete event). It is used in BuildingsPy's regression tests. These algorithms are particularly well suited to compare results when there are events and discontinuities in the result timeseries.

CSV Result Compare uses similar algorithms like pyfunnel.

Currently the mopyregtest.metrics collection has no similar built-in metric that allows for error localization in time and that is robust against events occuring at slightly different times.

Proposed feature

Make

• a proof of concept and
• a blueprint for other users that uses pyfunnel as a user-defined comparison metric

I think it makes more sense to fall back to an existing an proven solution instead of reproducing a similar functionality natively in MoPyRegtest.