Refine timeindex

[p-snft]

As different types of periods will be used for multi-period optimisation (long-term investment planing) and working with aggregated time-series (e.g. using tsam), I suggest to change the indexing of the flows:

• [x] flows have index [i, o, t] (types Node, Node , int), we use TIMESTEPS for t. #1077
• [ ] Storage content should use TIMEPOINTS
• [ ] The model has a lookup function to find out which period(s) a time step belongs to:
  • [ ] Multi-period: Each time step belongs to one investment period. Limits can be a function of the period. (#1109)
  • [ ] Aggregation: Each time step can belong to multiple aggregation periods (cf. #980).
• [ ] TIMEINDEX [int, int] for period, timestep is removed

[jokochems]

I think, this was agreed upon at the oemof meeting in November 2024 exactly as you suggested it here.