weight of time slices in in optimization

[kais-siala]

Currently, urbs gives the same weight for all time steps in the duration of the optimization.
It would be useful if we could define the weights to give to each time step (I think of it as a time series, like demand):
Weights = 0 could be ignored, the optimization will only look at weights > 0.
This should deliver the same result as repeating the time steps as often as their weight (if given in integers), but should be faster and less data-intensive.
Many models using the clustering of representative time steps do it based on weights, so this is state-of-the-art.
The current feature where the user sets the offset and the number of timesteps is useful if you want to test if the code runs, but the results of optimizing one week in winter on its own as if the whole year was like that are not very useful. I strongly recommend dealing with this issue as an upgrade.