general variables instead of numpy ones

neurolib/models/multimodel/builder/base/backend.py

@@ -635,11 +635,11 @@ def _init_xarray(self, times, results):
         var_idx = 0
         for node_idx, node_results in enumerate(self.state_variable_names):
             for result_idx, result in enumerate(node_results):
-                xr_results[result][:, node_idx] = results[:, var_idx + result_idx].astype(np.floating)
+                xr_results[result][:, node_idx] = results[:, var_idx + result_idx].astype(floating)
             var_idx += len(node_results)
 
         for state_var, array in xr_results.items():
-            xr_results[state_var] = array.astype(np.floating)
+            xr_results[state_var] = array.astype(floating)
 
         dataset = xr.Dataset(xr_results)
 

neurolib/models/multimodel/builder/base/network.py

@@ -29,7 +29,7 @@ def _sanitize_matrix(matrix, target_shape):
     """
     assert matrix.shape == target_shape
     if isinstance(matrix, np.ndarray) and matrix.dtype.kind == "i":
-        return matrix.astype(np.float)
+        return matrix.astype(float)
     else:
         return matrix
 
@@ -436,7 +436,7 @@ def init_node(self, **kwargs):
         assert self.idx_state_var is not None
         # gather inputs form this node - assumes constant delays
         var_idx = 0
-        self.inputs = np.zeros_like(self.connectivity, dtype=np.object)
+        self.inputs = np.zeros_like(self.connectivity, dtype=object)
         # iterate over masses as `from`, hence columns
         for from_mass, mass in enumerate(self.masses):
             # iterate over indices as `to`, hence rows
@@ -800,7 +800,7 @@ def _construct_input_matrix(self, within_node_idx):
         if isinstance(within_node_idx, int):
             within_node_idx = [within_node_idx] * self.num_nodes
         assert self.num_nodes == len(within_node_idx)
-        inputs = np.zeros_like(self.connectivity, dtype=np.object)
+        inputs = np.zeros_like(self.connectivity, dtype=object)
 
         # iterate over nodes as `from`, hence columns
         for from_node, (node, node_var_idx) in enumerate(zip(self.nodes, within_node_idx)):

neurolib/models/multimodel/model.py

@@ -71,15 +71,15 @@ def _set_model_params(self):
         # all matrices to floats
         for k, v in params.items():
             if isinstance(v, np.ndarray):
-                params[k] = v.astype(np.float)
+                params[k] = v.astype(float)
         params.update(DEFAULT_RUN_PARAMS)
         params["name"] = self.model_instance.label
         params["description"] = self.model_instance.name
         if isinstance(self.model_instance, Node):
             params.update({"N": 1, "Cmat": np.zeros((1, 1))})
         else:
             params.update(
-                {"N": len(self.model_instance.nodes), "Cmat": self.model_instance.connectivity.astype(np.floating)}
+                {"N": len(self.model_instance.nodes), "Cmat": self.model_instance.connectivity.astype(floating)}
             )
         return params
 

neurolib/utils/parameterSpace.py

@@ -134,7 +134,7 @@ def getRandom(self, safe=False):
                 random_value = np.random.choice(value)
                 if isinstance(random_value, np.float64):
                     random_value = float(random_value)
-                elif isinstance(random_value, np.int64):
+                elif isinstance(random_value, int64):
                     random_value = int(random_value)
                 randomPar[key] = random_value
         else:

tests/multimodel/base/test_network.py

@@ -25,7 +25,7 @@
 
 class TestSanitizeMatrix(unittest.TestCase):
     def test_sanitize_matrix_int(self):
-        mat = (np.random.rand(2, 2) * 100.0).astype(np.int)
+        mat = (np.random.rand(2, 2) * 100.0).astype(int)
         result = _sanitize_matrix(mat, (2, 2))
         self.assertTrue(result.dtype.kind == "f")
         np.testing.assert_equal(mat.astype(np.float), result)
@@ -37,7 +37,7 @@ def test_sanitize_matrix_float(self):
         np.testing.assert_equal(mat, result)
 
     def test_sanitize_matrix_wrong(self):
-        mat = (np.random.rand(2, 2) * 100.0).astype(np.int)
+        mat = (np.random.rand(2, 2) * 100.0).astype(int)
         with pytest.raises(AssertionError):
             result = _sanitize_matrix(mat, (3, 3))