Improve scikit-learn compatibility for CleanTransformer + tests

cleantext/sklearn.py

@@ -68,7 +68,13 @@ def __init__(
         self.replace_with_punct = replace_with_punct
         self.lang = lang
 
-    def fit(self, X: Any):
+    def fit(self, X: Any, y=None):
+        """
+        This method is defined for compatibility. It does nothing.
+        """
+        return self
+
+    def partial_fit(self, X: Any, y=None):
         """
         This method is defined for compatibility. It does nothing.
         """
@@ -89,3 +95,14 @@ def transform(self, X: Union[List[str], pd.Series]) -> Union[List[str], pd.Serie
             return X.apply(lambda text: clean(text, **self.get_params()))
         else:
             return list(map(lambda text: clean(text, **self.get_params()), X))
+
+    def get_feature_names_out(self, feature_names_out=None):
+        """
+        For compatibility with scikit-learn Pipeline objects.
+        This transformer will only return one column, which is ``'Clean Text``.
+        Args:
+            feature_names_out: Defined for compatibility with scikit-learn.
+        Returns:
+            list[str]: List with one element (i.e. The cleaned text column).
+        """
+        return ["Clean Text"]

tests/test_clean_transformer.py

@@ -1,7 +1,6 @@
 try:
     from cleantext.sklearn import CleanTransformer
     import pandas as pd
-
     import pytest
 
     transformer = CleanTransformer()
@@ -25,5 +24,12 @@ def test_set_params():
         assert transformer.get_params()["no_line_breaks"]
         assert transformer.get_params()["no_digits"]
 
+    def test_get_feature_names_out():
+        feature_names = transformer.get_feature_names_out()
+        assert feature_names == ["Clean Text"]
+
+    def test_fit():
+        transformer.fit(["sample1", "sample2"], [0, 1])
+        transformer.partial_fit(["sample1", "sample2"])
 except ImportError:
     pass