Write .get_feature() member method for all feature objects.

Right now, all feature objects are queried from the features table by getting the column that corresponds to the feature name.

For categorical features, it may be necessary to expand the feature column into multiple columns for modeling, with a boolean 1/0 for every categorical value. The best way to implement this is to use the Pandas built-in methods for expanding categories, and to have this happen at modeling time.

Thus, a `get_feature()` method would automatically expanding categorical feature using the built-in Pandas methods to do so.