Provide more informative warning messages in InputDataWarning (facebook#2713)

Summary:
Pull Request resolved: facebook#2713

X-link: meta-pytorch/botorch#2489

Provide more informative warning messages when `InputDataWarning`s are raised to specify whether it pertains to input features or output targets.

Updated unit tests accordingly to ensure coverage.

Differential Revision: D61797434

Author: ltiao

ax/modelbridge/base.py

@@ -904,7 +904,7 @@ def cross_validate(
             # users with this warning, we filter it out.
             warnings.filterwarnings(
                 "ignore",
-                message="Data is not standardized",
+                message=r"Data \(outcome observations\) not standardized",
                 category=InputDataWarning,
             )
             cv_predictions = self._cross_validate(

ax/modelbridge/cross_validation.py

@@ -184,7 +184,7 @@ def cross_validate(
                 # To avoid confusing users with this warning, we filter it out.
                 warnings.filterwarnings(
                     "ignore",
-                    message="Data is not standardized",
+                    message=r"Data \(outcome observations\) not standardized",
                     category=InputDataWarning,
                 )
                 cv_test_predictions = model._cross_validate(

ax/modelbridge/tests/test_base_modelbridge.py

@@ -225,7 +225,7 @@ def warn_and_return_mock_obs(
             nonlocal called
             called = True
             warnings.warn(
-                "Data is not standardized",
+                "Data (outcome observations) not standardized",
                 InputDataWarning,
                 stacklevel=2,
             )