Timesteps hard coded for exogenous time series

[SeanMcOwen]

In params.py, there is TIMESTEPS which is being set to figure out the size of the timeseries to use for gas and similar.

There is a problem that it is not dynamic to changing the sim params. However it still runs since it just uses the last value of the time series.

Two potential solutions:

• Get rid of the time series as exogenous and instead make a PSUB that is deciding at each timestep. The downside is that you can't use the time series across sims.
• Try to add something that forces the simulations before running to just-in-time compute the time series based on the current number of timesteps. This can possibly be rolled into #214 to make one wrapper function that handles cartesian sweeps + just-in-time computation of the time series

[danlessa]

@SeanMcOwen option one will likely degrade simulation performance, and I would avoid it.

We can consider option 3: consuming the gas time series from L1 time rather than Timestep units for the index. We would be required to generate more points than timesteps (e.g. maybe by a factor of 5x or 10x). The nice thing is that we would make the gas series invariant in time.