Scikit-learn has modified their metrics interface to deprecate the
`SCORERS` dictionary and made it private instead so we need to start
using the private interface. This is a bit risky as the private
interface can change at any time but since we pin scikit-learn, this is
less of an issue.

- Replace `loss` parameter for `SGDClassifier` with `log_loss`.
- Replace `auto` parameter for tree models with `sqrt`.
- Use `get_feature_names_out()` instead of `get_feature_names()`
- Explicitly specify `min_samples` when using RANSACRegressor with an
  estimator that is not LinearRegression.
- Use 'estimator' instead of deprecated 'base_estimator' for `RANSACRegressor`.

- Seaborn v0.12.0 has some breaking API changes that now require the
  following changes.
- Hue levels and keywords should be handled by `pointplot` and
  not `FacetGrid`.
- We need to make sure that the variable in the data frame that maps to
  hue levels is categorical.
- It is now recommended to explicitly assign palette colors to hue
  levels.