Update filtered input data type warning type in BoTorchTestCase

Summary:
Since #2304, these warnings use `InputDataWarning`. The current filter does not filter these out: https://github.com/pytorch/botorch/actions/runs/8904363989/job/24453525662

Also audited the test classes to make sure they call `super().setUp()` if they overwrite `setUp`.

Reviewed By: Balandat

Differential Revision: D56805187

fbshipit-source-id: a0899a7b799f4cbee13f9b5dff7f36b00bf01c19

Author: saitcakmak

botorch/utils/testing.py

@@ -51,7 +51,7 @@ def setUp(self, suppress_input_warnings: bool = True) -> None:
             warnings.filterwarnings(
                 "ignore",
                 message="The model inputs are of type",
-                category=UserWarning,
+                category=InputDataWarning,
             )
             warnings.filterwarnings(
                 "ignore",

test/acquisition/multi_objective/test_monte_carlo.py

@@ -702,6 +702,7 @@ def get_zeros_like_last(Z: Tensor) -> Tensor:
 
 class TestQNoisyExpectedHypervolumeImprovement(BotorchTestCase):
     def setUp(self):
+        super().setUp()
         self.ref_point = [0.0, 0.0, 0.0]
         self.Y_raw = torch.tensor(
             [
@@ -718,7 +719,6 @@ def setUp(self):
             ],
             device=self.device,
         )
-        super().setUp()
 
     def test_q_noisy_expected_hypervolume_improvement(self):
         for dtype in (torch.float, torch.double):

test/acquisition/test_integration.py

@@ -31,6 +31,7 @@
 
 class TestObjectiveAndConstraintIntegration(BotorchTestCase):
     def setUp(self) -> None:
+        super().setUp()
         self.q = 3
         self.d = 2
         self.tkwargs = {"device": self.device, "dtype": torch.double}