Fix mypy warnings

bico/core.py

@@ -162,11 +162,11 @@ def _fit(
         has_bico_obj = getattr(self, "bico_obj_", None)
         first_call = not (partial and has_bico_obj)
 
-        X = np.array(X, dtype=np.float64, order="C", copy=False)
-        assert isinstance(X, np.ndarray)
+        _X = np.array(X, dtype=np.float64, order="C", copy=False)
+        assert isinstance(_X, np.ndarray)
 
         if first_call:
-            self.n_features_in_ = X.shape[1]
+            self.n_features_in_: int = _X.shape[1]
             _seed = int(time.time()) if self.random_state is None else self.random_state
 
             # In Melanie's thesis, p = d
@@ -181,10 +181,10 @@ def _fit(
             c_random_state = ctypes.c_size_t(_seed)
 
             _DLL.init.restype = ctypes.POINTER(ctypes.c_void_p)
-            self.bico_obj_ = _DLL.init(c_d, c_k, c_p, c_m, c_random_state)
+            self.bico_obj_: Any = _DLL.init(c_d, c_k, c_p, c_m, c_random_state)
 
-        c_array = X.ctypes.data_as(ctypes.POINTER(ctypes.c_double))
-        c_n = ctypes.c_uint(X.shape[0])
+        c_array = _X.ctypes.data_as(ctypes.POINTER(ctypes.c_double))
+        c_n = ctypes.c_uint(_X.shape[0])
         _DLL.addData(self.bico_obj_, c_array, c_n)
 
         if not partial or fit_coreset:
@@ -203,7 +203,7 @@ def fit_predict(
         self._fit(X, partial=False, fit_coreset=True)
         return self.labels_
 
-    def predict(self, X: Sequence[Sequence[float]]) -> np.ndarray:
+    def predict(self, X: Sequence[Sequence[float]]) -> Any:
         self._fit_coreset()
 
         if self.coreset_estimator is None: