Fix bug # 406 which does not support categorical variables

mapie/conformity_scores/residual_conformity_scores.py

@@ -12,6 +12,7 @@
 from mapie._machine_precision import EPSILON
 from mapie._typing import ArrayLike, NDArray
 from mapie.conformity_scores import ConformityScore
+from mapie.wrap_arrays import wrap_ndarray_and_dataframe
 
 
 class AbsoluteConformityScore(ConformityScore):
@@ -377,6 +378,8 @@ def get_signed_conformity_scores(
         (X, y, y_pred,
          self.residual_estimator_,
          random_state) = self._check_parameters(X, y, y_pred)
+        # Wrap numpy or pandas array transparently to handle indexing
+        X = wrap_ndarray_and_dataframe(X)
 
         full_indexes = np.argwhere(
             np.logical_not(np.isnan(y_pred))

mapie/wrap_arrays.py

@@ -0,0 +1,44 @@
+from typing import Union
+import numpy as np
+import pandas as pd
+from mapie._typing import NDArray
+
+
+class wrap_ndarray_and_dataframe:
+
+    def __init__(self, X_array: Union[NDArray | pd.DataFrame]):
+        """
+            This class is a wrapper for numpy arrays and pandas DataFrames.
+            It is used to handle the indexing access to the data
+            in a consistent way.
+
+        Parameters
+        ----------
+        X_array: Union[NDArray | pd.DataFrame]
+            The data to wrap, either a numpy array or a pandas DataFrame.
+        """
+        self.X_array = X_array
+        if isinstance(X_array, pd.DataFrame):
+            self.X_array = pd.DataFrame(X_array, columns=X_array.columns)
+            self.X_array = self.X_array.astype(self.X_array.dtypes.to_dict())
+
+    def __getitem__(self, i: int):
+        """
+            This method is used to handle the indexing access to X_array.
+
+        Parameters
+        ----------
+        i: int
+            Index to access.
+
+        Returns
+        -------
+        NDArray
+            The data at index i.
+        """
+        if isinstance(self.X_array, pd.DataFrame):
+            return self.X_array.iloc[i].values
+        elif isinstance(self.X_array, np.ndarray):
+            return self.X_array[i]
+        else:
+            raise ValueError("Input must be a numpy array or pandas DataFrame")