remove X_train and X_test from parameters

hazardous/metrics/_concordance_index.py

@@ -12,13 +12,11 @@
 def concordance_index_incidence(
     y_test,
     y_pred,
+    y_train=None,
+    ipcw_estimator="km",
     time_grid=None,
     taus=None,
-    y_train=None,
-    X_train=None,
-    X_test=None,
     event_of_interest=1,
-    ipcw_estimator="km",
     tied_tol=1e-8,
 ):
     r"""Time-dependent concordance index for prognostic models with competing risks \
@@ -94,34 +92,26 @@ def concordance_index_incidence(
         Cumulative incidence for the event of interest, at the time points
         from the input time_grid.
 
+    y_train : array, dictionnary or dataframe of shape (n_samples, 2), default=None
+        The train target, consisting in the 'event' and 'duration' columns.
+        Only used when ipcw_estimator = "km".
+
+    ipcw_estimator : None, 'km', or fitted estimator, default="km"
+        The inverse probability of censoring weighted (IPCW) estimator.
+        - Pass None to set uniform weights to all samples.
+        - Pass "km" to use the Kaplan-Meier IPCW estimator. It fits using y_train,
+          which must be set.
+
     time_grid: array of shape (n_time_grid,), default=None
         Time points used to predict the cumulative incidence.
 
     taus: array of shape (n_taus,), default=None
         float or vector, timepoints at which the concordance index is
         evaluated.
 
-    y_train : array, dictionnary or dataframe of shape (n_samples, 2)
-        The train target, consisting in the 'event' and 'duration' columns.
-
-    X_train: array or dataframe of shape (n_samples_train, n_features), default=None
-        Covariates, used to learn a censor model if the inverse probability of
-        censoring weights (IPCW) is conditional on features (for instance Cox).
-        Unused if ipcw is None or 'km'.
-
-    X_test: array or dataframe of shape (n_samples_test, n_features), default=None
-        Covariates, used to predict weights a censor model if the inverse probability
-        of censoring weights (IPCW) is conditional on features (for instance Cox).
-        Unused if ipcw is None or 'km'.
-
     event_of_interest: int, default=1
         For competing risks, the event of interest.
 
-    ipcw_estimator : {None or 'km'}, default="km"
-        The inverse probability of censoring weighted (IPCW) estimator.
-        - None set uniform weights to all samples
-        - "km" use the Kaplan-Meier estimator
-
     tied_tol : float, default=1e-8
         The tolerance range to consider two probabilities equal.
 
@@ -149,13 +139,11 @@ def concordance_index_incidence(
     c_index_report = _concordance_index_incidence_report(
         y_test,
         y_pred,
+        y_train=y_train,
+        ipcw_estimator=ipcw_estimator,
         time_grid=time_grid,
         taus=taus,
-        y_train=y_train,
-        X_train=X_train,
-        X_test=X_test,
         event_of_interest=event_of_interest,
-        ipcw_estimator=ipcw_estimator,
         tied_tol=tied_tol,
     )
     return np.array(c_index_report["cindex"])
@@ -164,13 +152,11 @@ def concordance_index_incidence(
 def _concordance_index_incidence_report(
     y_test,
     y_pred,
+    y_train=None,
+    ipcw_estimator="km",
     time_grid=None,
     taus=None,
-    y_train=None,
-    X_train=None,
-    X_test=None,
     event_of_interest=1,
-    ipcw_estimator="km",
     tied_tol=1e-8,
 ):
     """Report version of function `concordance_index_incidence`.

hazardous/tests/test_cindex.py

@@ -457,7 +457,12 @@ def test_concordance_index_incidence_report_competitive():
     time_grid = [10, 20, 30]
 
     res = _concordance_index_incidence_report(
-        y_test, y_pred, time_grid, taus=None, y_train=y_test, event_of_interest=1
+        y_test,
+        y_pred,
+        y_train=y_test,
+        time_grid=time_grid,
+        taus=None,
+        event_of_interest=1,
     )
     assert res == {
         "cindex": [1.0],
@@ -472,7 +477,12 @@ def test_concordance_index_incidence_report_competitive():
     }
 
     res = _concordance_index_incidence_report(
-        y_test, y_pred, time_grid, taus=None, y_train=y_test, event_of_interest=2
+        y_test,
+        y_pred,
+        y_train=y_test,
+        time_grid=time_grid,
+        taus=None,
+        event_of_interest=2,
     )
 
     assert res == {