Merge pull request #549 from CronosC/tdl_minor_fixes

Addressed some minor issues in the TDL implementation

Author: Demirrr

ontolearn/learners/tree_learner.py

@@ -30,7 +30,10 @@
  OWLObjectIntersectionOf,
  OWLClassExpression,
  OWLObjectUnionOf,
+ OWLObjectOneOf,
+ OWLObjectHasValue
 )
+from owlapy.utils import HasFiller, HasOperands
 from owlapy.owl_individual import OWLNamedIndividual
 import ontolearn.triple_store
 from ontolearn.knowledge_base import KnowledgeBase
@@ -132,19 +135,32 @@ def concepts_reducer(
  dl_concept_path = reduced_cls((dl_concept_path, c))
  return dl_concept_path
 
+def contains_nominal(expr: OWLClassExpression) -> bool:
+ """Returns True if the OWL expression contains a nominal (OWLObjectOneOf, OWLDataOneOf)."""
+ if isinstance(expr, (OWLObjectOneOf, OWLObjectHasValue)):
+ return True
+
+ if isinstance(expr, HasFiller):
+ return contains_nominal(expr.get_filler())
+
+ if isinstance(expr, HasOperands):
+ return any(contains_nominal(op) for op in expr.get_operands())
+
+ return False
 
 class TDL:
  """Tree-based Description Logic Concept Learner"""
 
  def __init__(self, knowledge_base,
  use_inverse: bool = False,
  use_data_properties: bool = False,
- use_nominals: bool = False,
+ use_nominals: bool = True,
  use_card_restrictions: bool = False,
  kwargs_classifier: dict = None,
  max_runtime: int = 1,
  grid_search_over: dict = None,
  grid_search_apply: bool = False,
+ kwargs_grid_search: dict = {},
  report_classification: bool = True,
  plot_tree: bool = False,
  plot_embeddings: bool = False,
@@ -166,15 +182,21 @@ def __init__(self, knowledge_base,
  "min_samples_leaf": [1, 2, 3, 4, 5, 10],
  "max_depth": [1, 2, 3, 4, 5, 10, None],
  }
+ elif grid_search_apply and grid_search_over is not None:
+ pass
  else:
  grid_search_over = dict()
+
+ kwargs_grid_search.setdefault("cv", 10)
+ 
  assert (
  isinstance(knowledge_base, KnowledgeBase)
  or isinstance(knowledge_base, ontolearn.triple_store.TripleStore)
  or isinstance(knowledge_base)
  ), "knowledge_base must be a KnowledgeBase instance"
  print(f"Knowledge Base: {knowledge_base}")
  self.grid_search_over = grid_search_over
+ self.kwargs_grid_search = kwargs_grid_search
  self.knowledge_base = knowledge_base
  self.report_classification = report_classification
  self.plot_tree = plot_tree
@@ -209,11 +231,12 @@ def extract_expressions_from_owl_individuals(self, individuals: List[OWLNamedInd
  verbose=self.verbose,
  desc="Extracting information about examples"):
  for owl_class_expression in self.knowledge_base.abox(individual=owl_named_individual, mode="expression"):
- str_dl_concept=owl_expression_to_dl(owl_class_expression)
- individuals_to_feature_mapping.setdefault(owl_named_individual.str,set()).add(str_dl_concept)
- if str_dl_concept not in features:
- # A mapping from str dl representation to owl object.
- features[str_dl_concept] = owl_class_expression
+ if self.use_nominals or not contains_nominal(owl_class_expression):
+ str_dl_concept=owl_expression_to_dl(owl_class_expression)
+ individuals_to_feature_mapping.setdefault(owl_named_individual.str,set()).add(str_dl_concept)
+ if str_dl_concept not in features:
+ # A mapping from str dl representation to owl object.
+ features[str_dl_concept] = owl_class_expression
 
  assert len(features) > 0, "First hop features cannot be extracted. Ensure that there are axioms about the examples."
  if self.verbose > 0:
@@ -372,7 +395,7 @@ def fit(self, learning_problem: PosNegLPStandard = None, max_runtime: int = None
  if self.grid_search_over:
  grid_search = sklearn.model_selection.GridSearchCV(
  tree.DecisionTreeClassifier(**self.kwargs_classifier),
- param_grid=self.grid_search_over, cv=10, ).fit(X.values, y.values)
+ param_grid=self.grid_search_over, **self.kwargs_grid_search).fit(X.values, y.values)
  print(grid_search.best_params_)
  self.kwargs_classifier.update(grid_search.best_params_)
  # Training