GH-16147: Detect classifier/regressor via MRO in __sklearn_tags__ on sklearn 1.8

On sklearn 1.8, ClassifierMixin/RegressorMixin no longer carry
`_estimator_type` — the type is exposed only via __sklearn_tags__.
Wrappers built by make_classifier / make_regressor add the matching
mixin to bases but use is_generic=False, so the dynamic __init__
doesn't take an `estimator_type` argument and the base wrapper never
records `_estimator_type` on the instance.

Result on sklearn 1.8: getattr(self, '_estimator_type', None) returns
None, super().__sklearn_tags__() returns BaseEstimator's defaults
(estimator_type=None), and the type-specific branches in our
__sklearn_tags__ override never fire. is_classifier(self) then returns
False, the params_as_h2o_frames decorator skips asfactor() on y, and
algorithms that only support binomial classification (e.g. AdaBoost)
end up training as regression and crashing in predict.

Add a final MRO-based fallback: if neither _estimator_type nor the
upstream tags expose the type, look for ClassifierMixin / RegressorMixin
in type(self).__mro__.

Author: tomasfryda

h2o-py/h2o/sklearn/wrapper.py

@@ -596,6 +596,18 @@ def __sklearn_tags__(self):
         # NOT inherit ClassifierMixin (see h2o.sklearn._order_estimator_mixins
         # for type='estimator'), so an isinstance() check misses them.
         est_type = getattr(self, "_estimator_type", None) or getattr(tags, "estimator_type", None)
+        if est_type is None:
+            # sklearn 1.8 removed `_estimator_type = "classifier"` from ClassifierMixin
+            # and the same for RegressorMixin. Wrappers built via make_classifier /
+            # make_regressor add the respective mixin to bases but rely on the mixin
+            # for the type tag (is_generic=False, so __init__ doesn't take an
+            # estimator_type arg). Detect the mixin in the class MRO as the final
+            # fallback so is_classifier(self) returns the right answer on sklearn 1.8.
+            mro = type(self).__mro__
+            if ClassifierMixin in mro:
+                est_type = "classifier"
+            elif RegressorMixin in mro:
+                est_type = "regressor"
         if est_type == "classifier":
             tags.estimator_type = "classifier"
             try: