Make (old) PowerAnalytics work with PSY4

[GabrielKS]

Draft, the simulation that generates test data is still failing. I'm using the following patches to get this far:

• In HydroPowerSimulations.jl/src/hydro_generation.jl (this one should probably be committed):

-PSI.proportional_cost(cost::PSY.HydroGenerationCost , ::HydroEnergySurplusVariable, ::PSY.HydroGen, ::AbstractHydroReservoirFormulation)=PSY.get_energy_surplus_cost(cost)
-PSI.proportional_cost(cost::PSY.HydroGenerationCost , ::HydroEnergyShortageVariable, ::PSY.HydroGen, ::AbstractHydroReservoirFormulation)=PSY.get_energy_shortage_cost(cost)
+PSI.proportional_cost(cost::PSY.StorageCost , ::HydroEnergySurplusVariable, ::PSY.HydroGen, ::AbstractHydroReservoirFormulation)=PSY.get_energy_surplus_cost(cost)
+PSI.proportional_cost(cost::PSY.StorageCost , ::HydroEnergyShortageVariable, ::PSY.HydroGen, ::AbstractHydroReservoirFormulation)=PSY.get_energy_shortage_cost(cost)

• In PowerSystems.jl/src/models/cost_functions/StorageCost.jl (this one is just a kludge until something gets updated in PSI):

+get_variable(value::StorageCost) = get_discharge_variable_cost(value)

[GabrielKS]

The simulation and all old PA tests pass now, at the cost of removing simulation of hydro. The two patches above are still required.

[jd-lara]

We have methods in PSY to handle this, horizon us now always a period

[GabrielKS]

Is there a way to get_time_series_values passing in a Period?

[jd-lara]

Update the time series handling using PSY4 correctly.

[GabrielKS]

@daniel-thom is there a way to get a subset of a time series using a horizon Period as @jd-lara is requesting? I couldn't find one.

[daniel-thom]

@daniel-thom is there a way to get a subset of a time series using a horizon Period as @jd-lara is requesting? I couldn't find one.

Nope, we missed this one. You can only pass horizon_count as len. It could be fixed relatively easily.