Alternative scenarios

[Leonhard-B]

These scenarios save time by reuse of existing model instances. In a first step the dictionaries containing the data to be changed will be updated. Afterwards the scenario function also adapts the corresponding constraints of the model instance. Once the model is solved all changes need to be undone. This allows future utilization of the model instance in other scenarios.
Creation of completely new scenarios types is very complex now.

[Leonhard-B]

I think the complexity of the new scenario creation should be discussed. For people with little urbs knowledge it will be quite hard to built a new scenario function. If you have any wishes or doubts: Let me know. I will try to find a solution for it.
I also changed the documentation for the new scenario type. If important, please have a look at my changes.

[Leonhard-B]

Something I just found: The .h5 file saves prob._data and prob._result. prob._data does not contain valid data any more, since only the data of the base scenario is stored there. This is due to the structure of the alternative scenarios.
I will fix this soon, but the save function will probably look worse afterwards.

[maledo]

If you change anything regarding the input of the data dict, keep in mind the copy vs. pointer problematic in python. Sometimes you want to create a copy not just a pointer to the old object, that if you change something, the original variable is changed, too.

[Leonhard-B]

Saving of the h5 file should work properly, now.

[Leonhard-B]

It is possible now, to use both original and new version of scenario creation. If "alternative" appears in the scenario name, an existing model instance is being reused to speed up the process of scenario creation. If no scenario_base or another alternative scenario had run before, a new model instance is created first.