Add default tunable models

Create a GeneticLGBMClassifier, GeneticLGBMRegressor, GeneticXGBClassifier, etc. that creates a logical param space and automatically tunes the hyperparameters upon fitting.

No files, tests, or entry points are identified in the issue, so first determine where existing classifier and regressor APIs and parameter-space definitions live. Clarify which model variants are required and what constitutes a logical parameter space; done means the named tunable models automatically optimize their hyperparameters when fitting.

Written by the indexing model from the issue text.