Merge pull request #544 from dice-group/general-fixes

General fixes

Author: Demirrr

README.md

@@ -18,7 +18,8 @@ $E^+$ and $E^-$, learning [OWL Class expression](https://www.w3.org/TR/owl2-synt
 
 $$\forall p \in E^+\ \mathcal{K} \models H(p) \wedge \forall n \in E^-\ \mathcal{K} \not \models H(n).$$
 
-To tackle this supervised learning problem, ontolearn offers many symbolic, neuro-symbolic and deep learning based Learning algorithms: 
+To address this supervised learning problem, OntoLearn provides a diverse suite of learning algorithms, 
+including symbolic, neuro-symbolic, and deep learning-based approaches:
 - **TDL** → [Tree-based OWL Class Expression Learner for Large Graphs](https://rdcu.be/eJDlY)
 - **Drill** → [Neuro-Symbolic Class Expression Learning](https://www.ijcai.org/proceedings/2023/0403.pdf)
 - **EvoLearner** → [EvoLearner: Learning Description Logics with Evolutionary Algorithms](https://dl.acm.org/doi/abs/10.1145/3485447.3511925)
@@ -348,13 +349,22 @@ pytest -p no:warnings -x # Running 76 tests takes ~ 17 mins
 ```
 
 
-
 </details>
 
 ## References
-Currently, we are working on our manuscript describing our framework. 
 If you find our work useful in your research, please consider citing the respective paper:
 ```
+# Ontolearn
+@article{demir2025ontolearn,
+  title={Ontolearn---A Framework for Large-scale OWL Class Expression Learning in Python},
+  author={Demir, Caglar and Baci, Alkid and Kouagou, N'Dah Jean and Sieger, Leonie Nora and Heindorf, Stefan and Bin, Simon and Bl{\"u}baum, Lukas and Bigerl, Alexander and Ngomo, Axel-Cyrille Ngonga},
+  journal={Journal of Machine Learning Research},
+  volume={26},
+  number={63},
+  pages={1--6},
+  year={2025}
+}
+
 # TDL
 @InProceedings{10.1007/978-3-032-06066-2_29,
 author={Demir, Caglar and Yekini, Moshood and R{\"o}der, Michael and Mahmood, Yasir and Ngonga Ngomo, Axel-Cyrille},
@@ -368,17 +378,6 @@ pages={495--511},
 isbn={978-3-032-06066-2}
 }
 
-# Ontolearn
-@article{demir2025ontolearn,
-  title={Ontolearn---A Framework for Large-scale OWL Class Expression Learning in Python},
-  author={Demir, Caglar and Baci, Alkid and Kouagou, N'Dah Jean and Sieger, Leonie Nora and Heindorf, Stefan and Bin, Simon and Bl{\"u}baum, Lukas and Bigerl, Alexander and Ngomo, Axel-Cyrille Ngonga},
-  journal={Journal of Machine Learning Research},
-  volume={26},
-  number={63},
-  pages={1--6},
-  year={2025}
-}
-
 # ROCES
 @inproceedings{kouagou2024roces,
   title     = {ROCES: Robust Class Expression Synthesis in Description Logics via Iterative Sampling},
@@ -446,4 +445,4 @@ address="Cham"
 }
 ```
 
-In case you have any question, please contact: ```[email protected]``` or ```[email protected]```
+In case you have any question or feedback, please contact us: ```[email protected]``` or ```[email protected]```.

docs/usage/01_introduction.md

@@ -20,25 +20,25 @@ One of OntoLearn’s key contributions is its exclusive concept learning algorit
 Logics (DL). The library currently includes nine fully functional algorithms capable of learning complex concepts in DL. 
 For further details and references, relevant research papers can be found [here](09_further_resources.md).
 
-At the core of OntoLearn lies [Owlapy](https://github.com/dice-group/owlapy), a Python package inspired by the OWL API (its Java counterpart) and developed by 
+At the core of OntoLearn lies [OWLAPY](https://github.com/dice-group/owlapy), a Python package inspired by the OWL API (its Java counterpart) and developed by 
 the OntoLearn team. To enhance modularity, readability, and maintainability, we have separated Owlapy from Ontolearn into an 
 independent repository. This modular approach allows Owlapy to serve not only as a framework for representing OWL 2 
 entities, but also as a tool for ontology manipulation and reasoning.
 
 ---------------------------------------
 
-**Ontolearn (including [owlapy](https://github.com/dice-group/owlapy) and [ontosample](https://github.com/alkidbaci/OntoSample)) can do the following:**
-
-- **Use concept learning algorithms to generate hypotheses for classifying positive examples in a learning problem**.
-- **Use local datasets or datasets that are hosted on a triplestore server, for the learning task.**
-- Construct/Generate class expressions and evaluate them using different metrics.
-- Define learning problems.
-- Load/create/save ontologies in RDF/XML, OWL/XML.
-- Modify ontologies by adding/removing axioms.
-- Access individuals/classes/properties of an ontology (and a lot more).
-- Reason over an ontology.
-- Convenient functionalities like converting OWL class expressions to SPARQL or DL syntax.
-- Sample ontologies.
+**Ontolearn offers:**
+
+- **Diverse concept learning algorithms to generate hypotheses for classifying positive examples in a learning problem**.
+- **Support for local datasets and datasets hosted on triplestore servers.**
+- Perform operation on generated class expressions like evaluating them using different metrics, verbalizing, etc.
+- Generate learning problems.
+- An extensible and modular architecture that allows easy integration of new algorithms and functionalities.
+
+Via [OWLAPY](https://github.com/dice-group/owlapy) you can also perform ontology manipulations and reasoning.
+
+[OntoSample](https://github.com/alkidbaci/OntoSample) is another library closely related with the task of concept learning, where sampling is used 
+to accelerate the learning process.
 
 ------------------------------------
 

docs/usage/03_examples.md

@@ -74,7 +74,7 @@ for str_target_concept, examples in settings['problems'].items():
     model.save_best_hypothesis(n=3, path='Predictions_{0}'.format(str_target_concept))
 
     # Get top n hypotheses
-    hypotheses = list(model.best_hypotheses(n=3))
+    hypotheses = list(model.best_hypotheses(n=3, return_node=True))
 
     # Use hypotheses as binary function to label individuals.
     predictions = model.predict(individuals=list(typed_pos | typed_neg),

docs/usage/09_further_resources.md

@@ -24,11 +24,33 @@ Also check Owlapy's documentation [here](https://dice-group.github.io/owlapy/usa
 
 
 ## Citing
-
-Currently, we are working on our manuscript describing our framework. 
 If you find our work useful in your research, please consider citing the respective paper:
 
 ```
+# Ontolearn
+@article{demir2025ontolearn,
+  title={Ontolearn---A Framework for Large-scale OWL Class Expression Learning in Python},
+  author={Demir, Caglar and Baci, Alkid and Kouagou, N'Dah Jean and Sieger, Leonie Nora and Heindorf, Stefan and Bin, Simon and Bl{\"u}baum, Lukas and Bigerl, Alexander and Ngomo, Axel-Cyrille Ngonga},
+  journal={Journal of Machine Learning Research},
+  volume={26},
+  number={63},
+  pages={1--6},
+  year={2025}
+}
+
+# TDL
+@InProceedings{10.1007/978-3-032-06066-2_29,
+author={Demir, Caglar and Yekini, Moshood and R{\"o}der, Michael and Mahmood, Yasir and Ngonga Ngomo, Axel-Cyrille},
+editor={Ribeiro, Rita P. and Pfahringer, Bernhard and Japkowicz, Nathalie and Larra{\~{n}}aga, Pedro and Jorge, Al{\'i}pio M. and Soares, Carlos and Abreu, Pedro H. and Gama, Jo{\~a}o},
+title={Tree-Based OWL Class Expression Learner over Large Graphs},
+booktitle={Machine Learning and Knowledge Discovery in Databases. Research Track},
+year={2026},
+publisher={Springer Nature Switzerland},
+address={Cham},
+pages={495--511},
+isbn={978-3-032-06066-2}
+}
+
 # DRILL
 @inproceedings{demir2023drill,
   author = {Demir, Caglar and Ngomo, Axel-Cyrille Ngonga},
@@ -123,10 +145,10 @@ the project better. Find them in the folders
 
 ## Contribution
 
-We try to keep documentation up to day to the latest changes, but sometimes we may
+We try to keep documentation up to date with the latest changes, but sometimes we may
 overlook some details or make mistakes. If you notice any of such things please let us know :).
 As for coding part, feel free to create a pull request and our developers will take a look 
-on it. We appreciate your commitment. 
+at it. We appreciate your commitment. 
 
 ## Questions
 

examples/concept_learning_with_evolearner.py

@@ -60,7 +60,7 @@
 
     model.save_best_hypothesis(n=3, path='Predictions_{0}'.format(str_target_concept))
     # Get Top n hypotheses
-    hypotheses = list(model.best_hypotheses(n=3))
+    hypotheses = list(model.best_hypotheses(n=3, return_node=True))
     # Use hypotheses as binary function to label individuals.
     predictions = model.predict(individuals=list(typed_pos | typed_neg),
                                 hypotheses=hypotheses)

examples/concept_learning_with_tdl_and_triplestore_kb.py

@@ -6,7 +6,7 @@
 from ontolearn.utils.static_funcs import save_owl_class_expressions
 from owlapy.render import DLSyntaxObjectRenderer
 # (1) Initialize Triplestore- Make sure that UPB VPN is on
-kb = TripleStore(url="http://dice-dbpedia.cs.upb.de:9080/sparql")
+kb = TripleStore(url="https://dbpedia.data.dice-research.org/sparql")
 # (2) Initialize a DL renderer.
 render = DLSyntaxObjectRenderer()
 # (3) Initialize a learner.

examples/concept_learning_with_tdl_local_kb.py

@@ -0,0 +1,26 @@
+import json
+
+from owlapy.owl_individual import OWLNamedIndividual, IRI
+
+from ontolearn.knowledge_base import KnowledgeBase
+from ontolearn.learners import TDL
+from ontolearn.learning_problem import PosNegLPStandard
+from owlapy.render import DLSyntaxObjectRenderer
+
+kb = KnowledgeBase(path="../KGs/Mutagenesis/mutagenesis.owl")
+render = DLSyntaxObjectRenderer()
+model = TDL(kb)
+
+
+with open('../LPs/Mutagenesis/lps.json') as json_file:
+    settings = json.load(json_file)
+p = set(settings['problems']['NotKnown']['positive_examples'])
+n = set(settings['problems']['NotKnown']['negative_examples'])
+typed_pos = set(map(OWLNamedIndividual, map(IRI.create, p)))
+typed_neg = set(map(OWLNamedIndividual, map(IRI.create, n)))
+lp = PosNegLPStandard(pos=typed_pos, neg=typed_neg)
+
+h = model.fit(learning_problem=lp).best_hypotheses()
+str_concept = render.render(h)
+print("Concept:", str_concept)
+

ontolearn/lp_generator/generate_data.py

@@ -29,19 +29,24 @@
 
 
 class LPGen:
-    def __init__(self, kb_path=None, kb=None, storage_path=None, max_num_lps=1000, beyond_alc=False, depth=3, max_child_length=20, refinement_expressivity=0.2,
-                 downsample_refinements=True, sample_fillers_count=10, num_sub_roots=50, min_num_pos_examples=1):
+    def __init__(self, kb_path=None, kb=None, storage_path=None, max_num_lps=1000, beyond_alc=False, depth=3,
+                 max_child_length=20, refinement_expressivity=0.2, downsample_refinements=True, sample_fillers_count=10,
+                 num_sub_roots=50, min_num_pos_examples=1, max_pos_neg_examples_per_lp=None):
         """
-        Args
-        - kb_path: path to the owl file representing the knowledge base/ontology
-        - storage_path: directory in which to store the data to be generated. Not the directory needs not to exists, it would be created automatically
-        - max_num_lps: the maximum number of learning problems to store
-        - beyond_alc: whether to generate learning problems in ALCHIQD
-        - depth, max_child_length, refinement_expressivity, sample_fillers_count, num_sub_roots all refer to the size of the data (learning problems) to be generated
-        - downsample_refinements: whether to downsample refinements in ExpressRefinement. If refinement_expressivity<1, this must be set to True
+        Args:
+            kb_path: path to the owl file representing the knowledge base/ontology
+            kb: an instance of KnowledgeBase class. Can be used instead of kb_path.
+            storage_path: directory in which to store the data to be generated. Not the directory needs not to exists, it would be created automatically
+            max_num_lps: the maximum number of learning problems to store
+            beyond_alc: whether to generate learning problems in ALCHIQD
+            depth, max_child_length, refinement_expressivity, sample_fillers_count, num_sub_roots all refer to the size of the data (learning problems) to be generated
+            downsample_refinements: whether to downsample refinements in ExpressRefinement. If refinement_expressivity<1, this must be set to True
         """
-        self.lp_gen = KB2Data(path=kb_path,knowledge_base=kb, storage_path=storage_path, max_num_lps=max_num_lps, beyond_alc=beyond_alc, depth=depth,
-                      max_child_length=max_child_length, refinement_expressivity=refinement_expressivity,
-                      downsample_refinements=downsample_refinements, sample_fillers_count=sample_fillers_count, num_sub_roots=num_sub_roots, min_num_pos_examples=min_num_pos_examples)
+        self.lp_gen = KB2Data(path=kb_path,knowledge_base=kb, storage_path=storage_path, max_num_lps=max_num_lps,
+                              beyond_alc=beyond_alc, depth=depth, max_child_length=max_child_length,
+                              refinement_expressivity=refinement_expressivity,
+                              downsample_refinements=downsample_refinements, sample_fillers_count=sample_fillers_count,
+                              num_sub_roots=num_sub_roots, min_num_pos_examples=min_num_pos_examples,
+                              max_pos_neg_examples_per_lp = max_pos_neg_examples_per_lp)
     def generate(self):
         self.lp_gen.generate_descriptions().save_data()

ontolearn/lp_generator/helper_classes.py

@@ -21,7 +21,7 @@
 # OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
 # SOFTWARE.
 # -----------------------------------------------------------------------------
-
+from ontolearn.triple_store import TripleStore
 from tqdm import tqdm
 import random
 from ontolearn.knowledge_base import KnowledgeBase
@@ -72,7 +72,7 @@ class KB2Data:
     def __init__(self, path=None, storage_path=None, max_num_lps=1000, beyond_alc=False, depth=3,
                  max_child_length=20, refinement_expressivity=0.2,
                  downsample_refinements=True, sample_fillers_count=10, num_sub_roots=50,
-                 min_num_pos_examples=1,knowledge_base=None):
+                 min_num_pos_examples=1,knowledge_base=None, max_pos_neg_examples_per_lp=None):
         """
         Args
         - kb_path: path to the owl file representing the knowledge base/ontology
@@ -83,8 +83,19 @@ def __init__(self, path=None, storage_path=None, max_num_lps=1000, beyond_alc=Fa
         - depth, refinement_expressivity, sample_fillers_count, num_sub_roots all refer to the size of the data (learning problems) to be generated
         - downsample_refinements: whether to downsample refinements in ExpressRefinement. If refinement_expressivity<1, this must be set to True
         """
-        self.path = path
+        if path and knowledge_base:
+            assert path == knowledge_base.path, "Path argument and knowledge base's path do not match"
+        if path:
+            self.path = path
+        elif isinstance(knowledge_base, KnowledgeBase):
+            self.path = knowledge_base.path
+        elif isinstance(knowledge_base, TripleStore):
+            self.path = None
+        else:
+            raise ValueError("No knowledge base provided or path provided. Please provide a value to 'path' or "
+                             "'knowledge_base' arguments")
         if storage_path is None:
+            assert self.path is not None, "Storage path must be provided. Please provide a value to 'storage_path'"
             self.storage_path = f'{self.path[:self.path.rfind("/")]}/LPs/'
         else:
             self.storage_path = storage_path
@@ -96,7 +107,14 @@ def __init__(self, path=None, storage_path=None, max_num_lps=1000, beyond_alc=Fa
             self.kb = KnowledgeBase(path=path)
         else:
             self.kb = self.knowledge_base
-        self.num_examples = self.find_optimal_number_of_examples()
+        if max_pos_neg_examples_per_lp:
+            count = self.kb.individuals_count()
+            assert max_pos_neg_examples_per_lp <= count, \
+                ("The maximum number of examples per learning problem cannot be greater "
+                 "than the number of individuals in the knowledge base")
+            self.num_examples = max_pos_neg_examples_per_lp
+        else:
+            self.num_examples = self.find_optimal_number_of_examples()
         self.min_num_pos_examples = min_num_pos_examples
         atomic_concepts = frozenset(self.kb.ontology.classes_in_signature())
         self.atomic_concept_names = frozenset(

ontolearn/triple_store.py

@@ -43,7 +43,8 @@
 )
 from owlapy.owl_datatype import OWLDatatype
 from owlapy.owl_individual import OWLNamedIndividual
-from owlapy.owl_literal import OWLLiteral, BooleanOWLDatatype, DoubleOWLDatatype, NUMERIC_DATATYPES, TIME_DATATYPES
+from owlapy.owl_literal import OWLLiteral, BooleanOWLDatatype, DoubleOWLDatatype, NUMERIC_DATATYPES, TIME_DATATYPES, \
+    NonNegativeIntegerOWLDatatype
 from owlapy.owl_ontology import OWLOntologyID
 from owlapy.abstracts import AbstractOWLOntology, AbstractOWLReasoner
 from owlapy.owl_property import (