Remove references to deprecated NumPy type aliases.

PiperOrigin-RevId: 496324612

tensorflow_model_analysis/eval_saved_model/integration_test.py

@@ -178,19 +178,19 @@ def testPredictList(self):
 
     # Check that SparseFeatures were correctly populated.
     self.assertAllEqual(
-        np.array([b'cat'], dtype=np.object),
+        np.array([b'cat'], dtype=object),
         features_predictions_labels_list[0].features['animals'][
             encoding.NODE_SUFFIX].values)
     self.assertAllEqual(
-        np.array([b'dog'], dtype=np.object),
+        np.array([b'dog'], dtype=object),
         features_predictions_labels_list[1].features['animals'][
             encoding.NODE_SUFFIX].values)
     self.assertAllEqual(
-        np.array([b'cat', b'dog'], dtype=np.object),
+        np.array([b'cat', b'dog'], dtype=object),
         features_predictions_labels_list[2].features['animals'][
             encoding.NODE_SUFFIX].values)
     self.assertAllEqual(
-        np.array([], dtype=np.object),
+        np.array([], dtype=object),
         features_predictions_labels_list[3].features['animals'][
             encoding.NODE_SUFFIX].values)
 

tensorflow_model_analysis/eval_saved_model/util.py

@@ -200,7 +200,7 @@ def _dense_concat_rows(arrays: List[np.ndarray]) -> np.ndarray:
     raise ValueError('arrays must be a non-empty list.')
 
   shape_max = np.amax(np.array([a.shape for a in arrays]), axis=0)
-  if arrays[0].dtype == np.object:
+  if arrays[0].dtype == object:
     # Assume if the dtype is object then the array contains strings.
     padding_value = ''
   else:

tensorflow_model_analysis/eval_saved_model/util_test.py

@@ -141,25 +141,24 @@ def testSplitTensorValueSparseTypesPreserved(self):
     split_tensor_values = util.split_tensor_value(
         tf.compat.v1.SparseTensorValue(
             indices=np.array([[0, 0], [0, 1], [2, 0], [3, 1]]),
-            values=np.array(['zero0', 'zero1', 'two0', 'three1'],
-                            dtype=np.object),
+            values=np.array(['zero0', 'zero1', 'two0', 'three1'], dtype=object),
             dense_shape=np.array([4, 3])))
     expected_sparse_tensor_values = [
         tf.compat.v1.SparseTensorValue(
             indices=np.array([[0, 0], [0, 1]]),
-            values=np.array(['zero0', 'zero1'], dtype=np.object),
+            values=np.array(['zero0', 'zero1'], dtype=object),
             dense_shape=np.array([1, 2])),
         tf.compat.v1.SparseTensorValue(
             indices=np.zeros([0, 2], dtype=np.int64),
-            values=np.zeros([0], dtype=np.object),
+            values=np.zeros([0], dtype=object),
             dense_shape=np.array([1, 0])),
         tf.compat.v1.SparseTensorValue(
             indices=np.array([[0, 0]]),
-            values=np.array(['two0'], dtype=np.object),
+            values=np.array(['two0'], dtype=object),
             dense_shape=np.array([1, 1])),
         tf.compat.v1.SparseTensorValue(
             indices=np.array([[0, 1]]),
-            values=np.array(['three1'], dtype=np.object),
+            values=np.array(['three1'], dtype=object),
             dense_shape=np.array([1, 2])),
     ]
     for expected_sparse_tensor_value, got_sparse_tensor_value in zip(

tensorflow_model_analysis/evaluators/metrics_plots_and_validations_evaluator_test.py

@@ -2740,8 +2740,7 @@ def check_result(got_sliced_metrics):
            tn=[0.0, 2.0, 2.0],
            fn=[0.0, 1.0, 2.0]), False),
       ('BytesNotDiffable', b'some bytes', False),
-      ('NumpyObjectDtypeNotDiffable', np.array(['obj'],
-                                               dtype=np.object), False),
+      ('NumpyObjectDtypeNotDiffable', np.array(['obj'], dtype=object), False),
   )
   def testIsMetricDiffable(self, metric_value, expected_is_diffable):
     self.assertEqual(

tensorflow_model_analysis/metrics/metric_util_test.py

@@ -703,7 +703,7 @@ def testPrepareLabelsAndPredictionsWithVocab(self):
     self.assertAllClose(got_preds, np.array([0.2, 0.8]))
 
   def testPrepareLabelsAndPredictionsWithVocabUsingObjectType(self):
-    labels = np.array(['e', 'f'], dtype=np.object)
+    labels = np.array(['e', 'f'], dtype=object)
     preds = {'probabilities': [0.2, 0.8], 'all_classes': ['a', 'b', 'c']}
     got_labels, got_preds = metric_util.prepare_labels_and_predictions(
         labels, preds, label_vocabulary=['e', 'f'])

tensorflow_model_analysis/model_agnostic_eval/model_agnostic_predict_test.py

@@ -90,19 +90,19 @@ def testExtractFplExampleGraph(self):
     expected_language = [
         tf.compat.v1.SparseTensorValue(
             indices=np.array([[0, 0]]),
-            values=np.array([b'english'], dtype=np.object),
+            values=np.array([b'english'], dtype=object),
             dense_shape=np.array([1, 1])),
         tf.compat.v1.SparseTensorValue(
             indices=np.array([[0, 0]]),
-            values=np.array([b'chinese'], dtype=np.object),
+            values=np.array([b'chinese'], dtype=object),
             dense_shape=np.array([1, 1])),
         tf.compat.v1.SparseTensorValue(
             indices=np.array([], dtype=np.int64).reshape([0, 2]),
-            values=np.array([], dtype=np.object),
+            values=np.array([], dtype=object),
             dense_shape=np.array([1, 0])),
         tf.compat.v1.SparseTensorValue(
             indices=np.array([[0, 0]]),
-            values=np.array([b'chinese'], dtype=np.object),
+            values=np.array([b'chinese'], dtype=object),
             dense_shape=np.array([1, 1]))
     ]
     expected_probabilities = [

tensorflow_model_analysis/utils/util_test.py

@@ -1219,7 +1219,7 @@ def testSplitThenMergeAllowingScalarNumpyArray(self):
     self.assertEqual(remerged_got['predictions'].dtype, np.float64)
     self.assertEqual(remerged_got['labels'].dtype, np.float64)
     self.assertEqual(remerged_got['array_scalar'].dtype, np.float64)
-    self.assertEqual(remerged_got['objects'].dtype, np.object)
+    self.assertEqual(remerged_got['objects'].dtype, object)
 
   def testSplitThenMergeDisallowingScalars(self):
     extract = {