Monotonic constraints for autogluon-tabular

I would find it very useful if autogluon-tabular could add partial support for monotonic constraints. I would be happy to write a PR if that has interest?

Of the underlying models used only Lightgbm and Catboost support monotonic constraints but we could add a warning that the constraints will only be respected by these two models. Monotonic constraints can help in ability to generalize and are often very helpful in establishing user trust.

No files, tests, or entry points are named. Start by tracing how autogluon-tabular selects its underlying models and exposes prediction constraints, then compare the Lightgbm and Catboost integration points. Done should include partial monotonic-constraint support for those models and a warning for unsupported models, with tests covering the behavior.

Written by the indexing model from the issue text.