__sklearn_tags__ in TabPFNClassifier

[anshulg954]

File "/home/runner/work/tabpfn-server/tabpfn-server/app/services/tabpfn_prediction_service.py", line 118, in predict
res = model.predict_proba(x_test).tolist()
File "/opt/hostedtoolcache/Python/3.10.15/x64/lib/python3.10/site-packages/tabpfn/scripts/estimator/base.py", line 3117, in predict_proba
return self.predict_full(X, additional_y=additional_y)["proba"]
File "/opt/hostedtoolcache/Python/3.10.15/x64/lib/python3.10/site-packages/tabpfn/scripts/estimator/base.py", line 3066, in predict_full
X_full, y_full, additional_y_full, eval_pos = self.predict_common_setup(
File "/opt/hostedtoolcache/Python/3.10.15/x64/lib/python3.10/site-packages/tabpfn/scripts/estimator/base.py", line 1442, in predict_common_setup
check_is_fitted(self)
File "/opt/hostedtoolcache/Python/3.10.15/x64/lib/python3.10/site-packages/sklearn/utils/validation.py", line 1751, in check_is_fitted
tags = get_tags(estimator)
File "/opt/hostedtoolcache/Python/3.10.15/x64/lib/python3.10/site-packages/sklearn/utils/_tags.py", line 405, in get_tags
sklearn_tags_provider[klass] = klass.sklearn_tags(estimator) # type: ignore[attr-defined]
File "/opt/hostedtoolcache/Python/3.10.15/x64/lib/python3.10/site-packages/sklearn/base.py", line 540, in sklearn_tags
tags = super().sklearn_tags()
AttributeError: 'super' object has no attribute 'sklearn_tags'

The error arises because BaseEstimator in scikit-learn now relies on a method called sklearn_tags for metadata about the estimator, which is missing from TabPFNClassifier.

A possible fix:
Adding the following code to TabPFNClassifier?

def __sklearn_tags__(self):
    """
    Metadata tags required by scikit-learn.
    """
    tags = super().__sklearn_tags__() if hasattr(super(), "__sklearn_tags__") else {}
    tags.update({
        "requires_y": True,  # This estimator requires a target `y`
        "X_types": ["2darray"],  # Input should be a 2D array
        "multioutput": False,  # Not a multi-output classifier
        "allow_nan": True,  # Allows NaN values in `X`
        "binary_only": False,  # Handles multiclass and binary classification
    })
    return tags