Merge pull request #533 from dice-group/clean_up

alkidbaci · web-flow · commit 40f9f5edcfa3 · 2025-07-09T10:23:45.000+02:00
Small changes/fixes
diff --git a/docs/usage/05_evaluate_ce.md b/docs/usage/05_evaluate_ce.md
@@ -29,6 +29,8 @@ the positive and negative examples for the concept of 'Father'. Our positive exa
 ```python
 from owlapy.owl_individual import OWLNamedIndividual, IRI
 
+NS = "http://example.com/father#"
+
 positive_examples = {OWLNamedIndividual(IRI.create(NS, 'stefan')),
                      OWLNamedIndividual(IRI.create(NS, 'markus')),
                      OWLNamedIndividual(IRI.create(NS, 'martin'))}
@@ -94,8 +96,9 @@ You can now evaluate the class expression you just created:
 <!--pytest-codeblocks:cont-->
 ```python
 from ontolearn.quality_funcs import evaluate_concept
+from ontolearn.metrics import F1
 
-evaluated_concept = evaluate_concept(concept_to_test, F1(), encoded_lp)
+evaluated_concept = evaluate_concept(kg, concept_to_test, F1(), encoded_lp)
 ```
 In this example we use F1-score to evaluate the concept, but there are more [metrics](ontolearn.metrics) 
 which you can use including Accuracy, Precision and Recall. 
diff --git a/examples/concept_learning_cv_evaluation.py b/examples/concept_learning_cv_evaluation.py
@@ -274,7 +274,6 @@ def dl_concept_learning(args):
                 print(f"NCES Runtime: {rt_nces:.3f}")
 
             if not args.learner_types or 'nces2' in args.learner_types:
-                continue
                 start_time = time.time()
                 # () Fit model on training dataset
                 pred_nces2 = nces2.fit(train_lp).best_hypotheses(n=1)
@@ -299,7 +298,6 @@ def dl_concept_learning(args):
                 ##
 
             if not args.learner_types or 'roces' in args.learner_types:
-                continue
                 start_time = time.time()
                 # () Fit model on training dataset
                 pred_roces = roces.fit(train_lp).best_hypotheses(n=1)
diff --git a/ontolearn/knowledge_base.py b/ontolearn/knowledge_base.py
@@ -117,8 +117,10 @@ def __init__(self, *,
         elif reasoner_factory is not None:
             self.reasoner = reasoner_factory(self.ontology)
         else:
-            # default to native Reasoner of owlapy. To use SyncReasoner, pass the reasoner explicitly as an argument.
-            self.reasoner = StructuralReasoner(ontology=self.ontology)
+            if isinstance(self.ontology, Ontology):
+                self.reasoner = StructuralReasoner(ontology=self.ontology)
+            else:
+                self.reasoner = SyncReasoner(ontology=self.ontology, reasoner="Pellet")
 
         if load_class_hierarchy:
             self.class_hierarchy: ClassHierarchy
diff --git a/ontolearn/learners/drill.py b/ontolearn/learners/drill.py
@@ -448,6 +448,11 @@ def compute_quality_of_class_expression(self, state: RL_State) -> None:
                     named_individual = OWLClass(IRI('http://www.w3.org/2002/07/owl#', 'NamedIndividual'))
                     if named_individual in tp:
                         c = named_individual
+                    assert c == named_individual, ("Individuals in your dataset do not explicitly have the type"
+                                                   "owl:Thing or owl:NamedIndividual and this will eventually "
+                                                   "throw an error or provide misleading results because these "
+                                                   "individuals are not recognized as such. SPARQL queries used by "
+                                                   "TripleStore cannot infer this information.")
 
             sparql_query = owl_expression_to_sparql_with_confusion_matrix(expression=c, positive_examples=self.pos,
                                                                           negative_examples=self.neg)
diff --git a/ontolearn/scripts/run.py b/ontolearn/scripts/run.py
@@ -41,6 +41,7 @@
 from ontolearn.verbalizer import LLMVerbalizer
 from owlapy import owl_expression_to_dl
 import os
+import shutil
 
 app = FastAPI()
 args = None
@@ -79,9 +80,13 @@ def get_drill(data: dict):
                   use_nominals=data.get("use_nominals", True),
                   verbose=1)
     # (2) Either load the weights of DRILL or train it.
-    if data.get("path_to_pretrained_drill", None) and os.path.isdir(data["path_to_pretrained_drill"]):
+    if (data.get("path_to_pretrained_drill", None) and os.path.isdir(data["path_to_pretrained_drill"]) and
+            os.path.isfile("path_to_pretrained_drill/drill.pth")):
         drill.load(directory=data["path_to_pretrained_drill"])
     else:
+        # remove the pretrained directory if already exist but does not have the drill.pth file and retrain
+        if data.get("path_to_pretrained_drill", None) and os.path.isdir(data["path_to_pretrained_drill"]):
+            shutil.rmtree('path_to_pretrained_drill')
         # Train & Save
         drill.train(num_of_target_concepts=data.get("num_of_target_concepts", 1),
                     num_learning_problems=data.get("num_of_training_learning_problems", 1))
diff --git a/tests/test_tdl_regression.py b/tests/test_tdl_regression.py
@@ -60,7 +60,7 @@ def test_regression_mutagenesis(self):
             lp = PosNegLPStandard(pos=typed_pos, neg=typed_neg)
             h = model.fit(learning_problem=lp).best_hypotheses()
             q = compute_f1_score(individuals=frozenset({i for i in kb.individuals(h)}), pos=lp.pos, neg=lp.neg)
-            assert q >= 0.85
+            assert q >= 0.80
 
     def test_regression_carcinogenesis(self):
         path = "KGs/Carcinogenesis/carcinogenesis.owl"
