BH comments

Author: paroma

snorkel/labeling/model/label_model.py

@@ -455,7 +455,7 @@ def score(
         L
             An [n,m] matrix with values in {-1,0,1,...,k-1}
         Y
-            Gold labels associated with datapoints in L
+            Gold labels associated with data points in L
         metrics
             A list of metric names
         tie_break_policy
@@ -479,7 +479,7 @@ def score(
         """
         if tie_break_policy == "abstain":  # pragma: no cover
             logging.warning(
-                "Metrics calculated over datapoints with non-abstain labels only"
+                "Metrics calculated over data points with non-abstain labels only"
             )
 
         Y_pred, Y_prob = self.predict(

test/analysis/test_scorer.py

@@ -76,8 +76,7 @@ def test_abstain_labels(self) -> None:
 
         # Test abstain=-1 for preds and gold
         abstain_preds = np.array([-1, -1, 1, 1, 0])
-        abstain_probs = np.array([0.5, 0.5, 0.9, 0.7, 0.4])
-        results = scorer.score(golds, abstain_preds, abstain_probs)
+        results = scorer.score(golds, abstain_preds)
         results_expected = dict(accuracy=0.5)
         self.assertEqual(results, results_expected)
 

test/labeling/model/test_label_model.py

@@ -240,6 +240,7 @@ def test_predict_proba(self):
         np.testing.assert_array_almost_equal(probs, true_probs)
 
     def test_predict(self):
+        # 3 LFs that always disagree/abstain leads to all abstains
         L = np.array([[-1, 1, 0], [0, -1, 1], [1, 0, -1]])
         label_model = LabelModel(cardinality=2, verbose=False)
         label_model.fit(L, n_epochs=100)