A column-vector y was passed when a 1d array was expected (however, y is already a 1d array)

from sklearn.datasets import make_classification
from sklearn.ensemble import RandomForestClassifier, RandomForestRegressor
from econml.dml import NonParamDML

X, y = make_classification(random_state=42)

T = y.copy()

y, T = y.ravel(), T.ravel()

causal_model = NonParamDML(model_y=RandomForestClassifier(), 
                              model_t=RandomForestClassifier(),
                              model_final=RandomForestRegressor(),
                              discrete_outcome=True,
                              discrete_treatment=True,
                              cv=3)

causal_model.fit(y, T, X=X)

successfully fit while displaying the following warning message:

A column-vector y was passed when a 1d array was expected. Please change the shape of y to (n_samples,), for example using ravel().

Is there anything incorrect in the code or are there any ways to avoid the warning?

Thanks a lot for any feedback.

Best,
Yanis

Start by reproducing the provided NonParamDML example and tracing how its fit inputs reach the scikit-learn models. Identify which input produces the column-vector warning, then confirm the expected behavior with a focused regression test; done means the valid 1D inputs no longer emit the misleading warning.