I am intrigued on how this would perform on astronomical data. If anyone is not aware there are many periodic phenomena in astronomy - e.g. variable stars whic

I am intrigued on how this would perform on astronomical data. If anyone is not aware there are many periodic phenomena in astronomy - e.g. variable stars which can have periods from minutes to hundreds of days. The description of this library sounds like it's very tied to the human world - talking about yearly, weekly and daily seasonality. [Weirdly though, we do sometimes see variability on 'human' timescales in astronomical data series. If maintenance is carried out weekly on a Monday that can add a signal into the data through missing datapoints.]