Bug when passing featurizer that requires more than 1 row in SparseLinearDrLearner

[itamarfaran]

in file econml/utilities.py:75 (function name check_high_dimensional):

d_x = clone(featurizer, safe=False).fit_transform(X[[0], :]).shape[1]

when passing a transformer such as SplineTransformer, it raises an error:

ValueError: Found array with 1 sample(s) (shape=(1, 1)) while a minimum of 2 is required by SplineTransformer.

from what I understand this line only runs to infer the number of rows out, but it fails the whole fit process.
this would also fail for transformers as OneHotEncoder as the number of new rows is dependent on the data

[itamarfaran]

#906