Merge pull request #174 from geometric-intelligence/scikitapi

update topology_classifier tests

neurometry/estimators/topology/topology_classifier.py

@@ -1,6 +1,7 @@
 import os
 
 import numpy as np
+import torch
 from gtda.diagrams import PersistenceEntropy
 from gtda.homology import VietorisRipsPersistence, WeightedRipsPersistence
 from sklearn.base import BaseEstimator, ClassifierMixin
@@ -141,6 +142,9 @@ def fit(self, X, y=None):
             Returns the instance itself.
         """
 
+        if not isinstance(X, np.ndarray | torch.Tensor):
+            raise ValueError(f"Expected array-like input for X, but got {type(X).__name__}.")
+
         ref_point_clouds, ref_labels = self._generate_ref_data(X)
         self.ref_labels = ref_labels
         if self.reduce_dim:

tests/neurometry/estimators/topology/test_topology_classifier.py

@@ -1,6 +1,5 @@
 import os
 
-import numpy as np
 import pytest
 from sklearn.exceptions import NotFittedError
 
@@ -17,7 +16,7 @@ class BaseTopologyTest:
     num_points = 700
     encoding_dim = 10
     fano_factor = 0.1
-    num_samples = 100
+    num_samples = 200
     homology_dimensions = (0, 1)
 
     @classmethod
@@ -30,24 +29,10 @@ def setup_class(cls):
             reduce_dim=True,
         )
 
-        cls.null_data = cls.generate_null_data()
         cls.circle_data = cls.generate_circle_data()
         cls.sphere_data = cls.generate_sphere_data()
         cls.torus_data = cls.generate_torus_data()
 
-    @classmethod
-    def generate_null_data(cls):
-        rng = np.random.default_rng(seed=0)
-        task_points = rng.normal(0, 1, (cls.num_points, cls.encoding_dim))
-        noisy_points, _ = synthetic.synthetic_neural_manifold(
-            points=task_points,
-            encoding_dim=cls.encoding_dim,
-            nonlinearity="sigmoid",
-            scales=5 * gs.random.rand(cls.encoding_dim),
-            fano_factor=cls.fano_factor,
-        )
-        return noisy_points
-
     @classmethod
     def generate_circle_data(cls):
         task_points = synthetic.hypersphere(1, cls.num_points)
@@ -86,22 +71,22 @@ def generate_torus_data(cls):
 
 
 class TestTopologyClassifier(BaseTopologyTest):
-    def test_fit(self):
-        """Test that the fit method runs without errors."""
+    """Unit tests for the topology classifier."""
+
+    def test_invalid_input(self):
+        """Test classifier with invalid input data."""
+        invalid_data = "invalid_input"
+        with pytest.raises(ValueError):
+            self.classifier.fit(invalid_data)
+
+    def test_fit_and_predict_circle(self):
+        """Test that the fit method runs without errors and predict on circle data."""
         try:
-            self.classifier.fit(self.null_data)
+            self.classifier.fit(self.circle_data)
         except Exception as e:
             pytest.fail(f"Fit method raised an exception: {e}")
 
-    def test_predict_null(self):
-        """Test prediction on null (noise) data."""
-        self.classifier.fit(self.null_data)
-        prediction = self.classifier.predict(self.null_data)
-        assert prediction[0] == 0, "Prediction for null data should be 0 (null)"
-
-    def test_predict_circle(self):
-        """Test prediction on circle data."""
-        self.classifier.fit(self.circle_data)
+        # If fit is successful, test the prediction
         prediction = self.classifier.predict(self.circle_data)
         assert prediction[0] == 1, "Prediction for circle data should be 1 (circle)"
 
@@ -117,25 +102,6 @@ def test_predict_torus(self):
         prediction = self.classifier.predict(self.torus_data)
         assert prediction[0] == 3, "Prediction for torus data should be 3 (torus)"
 
-    def test_consistent_predictions(self):
-        """Test that predictions are consistent across multiple calls."""
-        self.classifier.fit(self.circle_data)
-        prediction1 = self.classifier.predict(self.circle_data)
-        prediction2 = self.classifier.predict(self.circle_data)
-        assert prediction1[0] == prediction2[0], "Predictions should be consistent across calls"
-
-    def test_reduce_dim_false(self):
-        """Test classifier with reduce_dim set to False."""
-        classifier = TopologyClassifier(
-            num_samples=self.num_samples,
-            fano_factor=self.fano_factor,
-            homology_dimensions=self.homology_dimensions,
-            reduce_dim=False,
-        )
-        classifier.fit(self.sphere_data)
-        prediction = classifier.predict(self.sphere_data)
-        assert prediction[0] == 2, "Prediction for sphere data should be 2 (sphere) with reduce_dim=False"
-
     def test_not_fitted_error(self):
         """Test that predicting before fitting raises an error."""
         classifier = TopologyClassifier(
@@ -147,9 +113,3 @@ def test_not_fitted_error(self):
         with pytest.raises(NotFittedError):
             classifier.predict(self.circle_data)
 
-    def test_invalid_input(self):
-        """Test classifier with invalid input data."""
-        self.classifier.fit(self.null_data)
-        invalid_data = "invalid_input"
-        with pytest.raises(ValueError):
-            self.classifier.predict(invalid_data)