Unintended behaviour reading timeseries values?

[mihlema]

I just came across some unintuitive behaviour when using SpineInterface time-series parameter values. My minimal example is attached
here.

Basically, if I access a TimeSeries that extends beyond the last DateTime, I sometimes get a value, but only if the start DateTime coincides with the last given timestep (here: 2031-01-01T00:00). In my particular example, I'm currently circumventing this by only using the year of the corresponding timeslice (e.g. t=year(start(t))) but I feel in the long run we need some more options on how to deal with timeseries end values.

julia> t = TimeSlice(DateTime(2031,1,1),DateTime(2032,1,1))
2031-01-01T00:00~>2032-01-01T00:00

julia> test_param(testing=testing()[1],t=t)
0.0

julia> t = TimeSlice(DateTime(2031,1,1,0),DateTime(2031,1,1,1))
2031-01-01T00:00~>2031-01-01T01:00

julia> test_param(testing=testing()[1],t=t)
0.0

julia> t = TimeSlice(DateTime(2031,1,1,1),DateTime(2031,1,1,2))
2031-01-01T01:00~>2031-01-01T02:00

julia> test_param(testing=testing()[1],t=t)
#(nothing)