Merge pull request #540 from gchq/feature/experimental-warnings

Add warnings to experimental parts of the codebase

Author: pc532627

pyproject.toml

@@ -237,7 +237,7 @@ filterwarnings = [
 
     # Other warnings
     "ignore::pytest.PytestCollectionWarning",
-    "ignore::vanguard.warnings.ExperimentalFeatureWarning",
+    "ignore::vanguard.warnings.ExperimentalFeatureWarning",  # we already know what parts of the code are experimental!
     "ignore:The input matches the stored training data. Did you forget to call model.train\\(\\)\\?:gpytorch.utils.warnings.GPInputWarning",
     "ignore:To copy construct from a tensor, it is recommended to use sourceTensor.clone\\(\\).detach\\(\\):UserWarning",
     "ignore:Pass auto_restart=True to the controller to automatically restart training up to the last stable iterations:UserWarning",

vanguard/classification/kernel.py

@@ -30,6 +30,7 @@
 from vanguard.classification.mixin import Classification, ClassificationMixin
 from vanguard.classification.models import InertKernelModel
 from vanguard.decoratorutils import Decorator, process_args, wraps_class
+from vanguard.warnings import warn_experimental
 
 ControllerT = TypeVar("ControllerT", bound=GPController)
 SAMPLE_DIM, TASK_DIM = 0, 2
@@ -41,6 +42,11 @@ class DirichletKernelMulticlassClassification(Decorator):
 
     Based on the paper :cite:`MacKenzie14`.
 
+
+    .. warning::
+        This decorator is EXPERIMENTAL. It may cause errors or give incorrect results, and may have breaking changes
+        without warning.
+
     .. warning::
         Fuzzy classification (with `classify_fuzzy_points`) is not supported.
 
@@ -80,6 +86,7 @@ def __init__(self, num_classes: int, **kwargs: Any) -> None:
         :param num_classes: The number of target classes.
         :param kwargs: Keyword arguments passed to :class:`~vanguard.decoratorutils.basedecorator.Decorator`.
         """
+        warn_experimental("The DirichletKernelMulticlassClassification decorator")
         self.num_classes = num_classes
         super().__init__(framework_class=GPController, required_decorators={}, **kwargs)
 

vanguard/features.py

@@ -47,6 +47,10 @@ class HigherRankFeatures(Decorator):
     This decorator modifies the gp model class to make it compatible with higher
     rank features.
 
+    .. warning::
+        This decorator is EXPERIMENTAL. It may cause errors or give incorrect results, and may have breaking changes
+        without warning.
+
     :Example:
         >>> @HigherRankFeatures(2)
         ... class NewController(GPController):