v2 API: consider specifying datatypes for pandas series

[dweindl]

So far, there the Series of the different Dataframes had varying types, this makes code unnecessary complex.
For example, depending on the contents, the conditions table with either be float or object.
We should consider having a well-defined type for the different columns. Either via the get_*_df, or via custom pandas accessors. The latter would also allow filling in default values on the fly which would simplify things a lot. However, it could get a bit clunky to always write something like problem.parameter_df.petab_parameters.parameterScale...