fix issue 1828 about a warning (skrub-data#1908)

Co-authored-by: Jerome Dockes <jerome@dockes.org>

Author: sandrineh

CHANGES.rst

@@ -14,6 +14,9 @@ New Features
 
 Changes
 -------
+- An unnecessary warning that was raised when passing a numpy array to the
+  TableVectorizer has been removed. :pr:`1908` by
+  :user:`Sandrine Henry <sandrineh>`.
 
 Bugfixes
 --------

skrub/_check_input.py

@@ -160,12 +160,6 @@ def _handle_array(self, X):
                 "Input should be a DataFrame. Found an array with incompatible shape:"
                 f" {X.shape}."
             )
-        warnings.warn(
-            "Only pandas and polars DataFrames are supported, but input is a Numpy"
-            " array. Please convert Numpy arrays to DataFrames before passing them to"
-            " skrub transformers. Converting to pandas DataFrame with columns"
-            " ['0', '1', …]."
-        )
         import pandas as pd
 
         columns = list(map(str, range(X.shape[1])))

skrub/tests/test_check_input.py

@@ -21,15 +21,13 @@ def test_input_is_an_array():
     check = CheckInputDataFrame()
 
     # 2D array in fit
-    with pytest.warns(UserWarning, match="Only .* DataFrames are supported"):
-        d = check.fit_transform(a)
+    d = check.fit_transform(a)
     assert sbd.is_pandas(d)
     assert sbd.is_dataframe(d)
     assert sbd.column_names(d) == ["0", "1"]
 
     # 2D array in transform
-    with pytest.warns(UserWarning, match="Only .* DataFrames are supported"):
-        d = check.transform(a)
+    d = check.transform(a)
     assert sbd.is_pandas(d)
     assert sbd.is_dataframe(d)
     assert sbd.column_names(d) == ["0", "1"]