Add `dispatch_production` method for using production data directly from the solver

Currently both the scipy and adhoc solvers calculate production internally, but then cache this data ([here](https://github.com/EnergySystemsModellingLab/MUSE_OS/blob/d7b1bc622f0a4cb7781432d42ea32f8e9b678c8b/src/muse/investments.py#L102)) before production is recalculated again. I imagine the internally calculated production data is nonsense, hence why it has to be calculated again, but for development purposes it would be really useful to have an easy way to output this data.

My proposal is that we add a `dispatch_production` method that retrieves production data from the cache (if that's even possible) and returns this, rather than recalculating. Once we have this, we can then investigate what needs to be changed so that this production data makes sense, before hopefully making this the default and dropping the `dispatch_production` feature entirely.

Any other ways of outputting the cached production data can be considered