Allocation renewal

[SouthEndMusic]

Fixes #2255
Fixes #2149
Fixes #2147

[jarsarasty]

Hi @SouthEndMusic,

I see good progress with the new implementation. For now, I only have minor comments about the way the bounds are handled in the model.

[SouthEndMusic]

Here's a status update for the new allocation implementation:

• Documentation (Overhaul allocation documentation (technical reference) #2252): in progress
• Basin (Linear Basin implementation for allocation #2060): implemented (with 'universal bottom' fix by @simulutions), not tested in detail. There is some work to be done on the piecewise linear basin profile; how much sections do we use and how do we determine them, what extrapolation of the profile should we support
• LevelDemand (Reformulating level demand in allocation #2225, LevelDemand in allocation #2319): implemented (support for multiple priorities per level demand still to be done), existing tests need to be adapted/extended
• FlowDemand (Support multiple priorities per FlowDemand node #2320): implemented (support for multiple priorities per flow demand still to be done), existing tests must be adapted/extended
• UserDemand (no standalone issue): implemented, tested somewhat, other tests need to be adapted. We should evaluate whether linear interpolation of the return factor makes sense
• TabulatedRatingCurve (Linearize TabulatedRatingCurve curve for allocation #2148): implemented, not tested. Here we also have to think about how many sections we use in the piecewise linear Q(h) relationship
• LinearResistance & ManningResistance (Linearize LinearResistance and ManningResistance for allocation #2149): implemented (infeasibilities avoided with 'universal bottom' fix), not tested
• Control of Pump and Outlet from allocation (Allow Pump and Outlet nodes to be controlled by allocation #714): implemented, not tested. See Control from allocation #2139 for a test model
• Allocation taking discrete control into account (Let allocation take DiscreteContol into account #2321): not implemented
• Allocation taking continuous control into account (Let allocation take ContinuousControl into account #2322): not implemented
• Reducing fixed abstractions during drought (Reduce fixed abstractions in allocation during drought #2323): not implemented
• Allocation taking node activity into account (Take node activity into account in allocation #728): not implemented
• Taking source priorities into account (Test source priority optimization in allocation #2051): in progress, discussed, not tested
• Running optimizations for subnetworks in parallel (Run optimizations for subnetworks in allocation in parallel #2325): not implemented
• Output files (Test allocation output files #2326): implemented, not tested. Stays (mostly) the same as described in https://ribasim.org/reference/usage.html#results, with an addition of which bounds are hit after optimization for all goals
• Connection between primary network (formerly main network) and secondary networks (formerly subnetworks) (The connection between the primary network and secondary networks in allocation #2327): in progress, tested somewhat
• Handling infeasibilities (Allocation infeasibility handling #2328): not implemented
• Provide initial guess for solution?

[SouthEndMusic]

@jarsarasty here are some updates

Because I have done quite some refactoring for Ribasim NL issues this branch accumulated quite some merge conflicts. I therefore decided to bring this branch up to date and merge it soon. I marked the remaining tests that are not green yet as broken, and allocation will be categorized as a experimental feature. This is technically breaking, but allocation is hardly used by modelers.

I'm also in conversation about my migration to the wflow team, I will have a meeting about that next Tuesday. It probably means that I have to hand over the allocation developments to @simulutions (Lex) in the coming weeks

[jarsarasty]

Thanks for the updates, @SouthEndMusic! I'll review the changes next week.