Merge pull request #110 from dice-group/csv_kg_categorical

A CSV file is mapped to an RDF KGs containing only literals

README.md

@@ -25,11 +25,34 @@ pip3 install owlapy
 ```shell
 # To download RDF knowledge graphs
 wget https://files.dice-research.org/projects/Ontolearn/KGs.zip -O ./KGs.zip && unzip KGs.zip
-pytest -p no:warnings -x # Running  142 tests ~ 30 secs
+pytest -p no:warnings -x # Running  147 tests ~ 35 secs
 ```
 
 ## Examples
 
+### Sklearn to OWL Ontology
+
+<details><summary> Click me! </summary>
+
+```python
+from owlapy.owl_ontology_manager import SyncOntologyManager
+from owlapy.util_owl_static_funcs import csv_to_rdf_kg
+import pandas as pd
+from sklearn.datasets import load_iris
+data = load_iris()
+df = pd.DataFrame(data.data, columns=data.feature_names)
+df.to_csv("iris_dataset.csv", index=False)
+path_kg = "iris_kg.owl"
+# Construct an RDF Knowledge Graph from a CSV file
+csv_to_rdf_kg(path_csv="iris_dataset.csv", path_kg=path_kg, namespace="http://owlapy.com/iris")
+onto = SyncOntologyManager().load_ontology(path_kg)
+assert len(onto.get_abox_axioms()) == 750
+
+```
+
+</details>
+
+
 ### Exploring OWL Ontology
 
 <details><summary> Click me! </summary>

owlapy/owl_literal.py

@@ -257,7 +257,7 @@ def parse_integer(self) -> int:
         Returns:
             An integer value that is represented by this literal.
         """
-        raise ValueError
+        return int(self._v)
 
     def is_string(self) -> bool:
         """Whether this literal is typed as string."""

owlapy/util_owl_static_funcs.py

@@ -62,15 +62,14 @@ def save_owl_class_expressions(expressions: OWLClassExpression | List[OWLClassEx
         ontology.add_axiom(equivalent_classes_axiom)
     ontology.save(path=path, inplace=False, rdf_format=rdf_format)
 
-def csv_to_rdf_kg(path_csv:str=None,path_kg:str=None,namespace:str=None,rdf_format:str=None):
+def csv_to_rdf_kg(path_csv:str=None,path_kg:str=None,namespace:str=None):
     """
     Transfroms a CSV file to an RDF Knowledge Graph in RDF/XML format.
 
     Args:
         path_csv (str): X
         path_kg (str): X
         namespace (str): X
-        rdf_format(str):X
 
     Raises:
         AssertionError:
@@ -94,10 +93,6 @@ def csv_to_rdf_kg(path_csv:str=None,path_kg:str=None,namespace:str=None,rdf_form
     assert path_kg is not None, f"path_kg cannot be None.Currently {path_kg}"
     assert namespace is not None, "namespace cannot be None"
     assert namespace[:7]=="http://", "First characters of namespace must be 'http://'"
-    if rdf_format is None:
-        rdf_format="rdfxml"
-    else:
-        assert rdf_format in ["ntriples", "turtle"]
 
     # Initialize an Ontology Manager.
     manager = SyncOntologyManager()
@@ -106,22 +101,27 @@ def csv_to_rdf_kg(path_csv:str=None,path_kg:str=None,namespace:str=None,rdf_form
 
     # Read the CSV file
     df = pd.read_csv(path_csv)
+
     # () Iterate over rows
-    for index, row in tqdm(df.iterrows()):
+    for index, row in (tqdm_bar := tqdm(df.iterrows()) ):
         individual=OWLNamedIndividual(f"{namespace}#{str(index)}".replace(" ","_"))
+        tqdm_bar.set_description_str(f"Creating RDF Graph from Row:{index}")
+        # column_name is considered as a predicate
+        # value is considered as a data property
         for column_name, value in row.to_dict().items():
-            if isinstance(value, float):
-                # Create an IRI for the predicate
-                str_property_iri=f"{namespace}#{column_name}".replace(" ","_")
-                str_property_iri=str_property_iri.replace("(","/")
-                str_property_iri = str_property_iri.replace(")", "")
+            # Create an IRI for the predicate
+            str_property_iri = f"{namespace}#{column_name}".replace(" ", "_")
+            str_property_iri = str_property_iri.replace("(", "/")
+            str_property_iri = str_property_iri.replace(")", "")
 
+            if isinstance(value, float) or isinstance(value, int) or isinstance(value, str):
                 axiom = OWLDataPropertyAssertionAxiom(subject=individual,
                                                       property_=OWLDataProperty(iri=str_property_iri),
                                                       object_=OWLLiteral(value=value))
                 ontology.add_axiom(axiom)
-
             else:
-                raise NotImplementedError(f"How to represent value={value} has not been decided")
-
+                raise NotImplementedError(f"How to represent\n"
+                                          f"predicate=**{str_property_iri}**\n"
+                                          f"value=**{value}**\n"
+                                          f"has not been decided")
     ontology.save(path=path_kg)

tests/test_owl_static_funcs.py

@@ -10,8 +10,6 @@
 import rdflib
 
 from owlapy.owl_ontology_manager import SyncOntologyManager
-from sklearn.datasets import load_iris
-import pandas as pd
 
 class TestRunningExamples:
     def test_readme(self):
@@ -35,7 +33,6 @@ def test_csv_to_kg(self):
         df = pd.DataFrame(data.data, columns=data.feature_names)
         df['target'] = data.target
         df.to_csv("iris_dataset.csv", index=False)
-
         assert len(df) == 150
         path_kg = "iris_kg.owl"
         csv_to_rdf_kg(path_csv="iris_dataset.csv", path_kg=path_kg, namespace="http://example.com/society")