When loading a model, the default rn is "Group by observables". This should be run through a check:
- If a vis_df is existent, use that
- Use the one that has less unique values from observables and conditions
TBD: Should we have a Setting that users can use for their Default?