Add handing of missing data

[KaWest]

I've run into a use case where my time series sometimes lack data in parts of the grid. Computing a cost for these points obviously becomes challenging but after discussing with @ElektrikAkar, there might be some strategies for handling this.

This issue is to be considered done when:

• The DTW code can compute a similarity metric also between time-series where one or both of the series have gaps in the data