Merge pull request #108 from dice-group/csv_kg

From CSV to RDF Knowledge Graph

owlapy/abstracts/abstract_owl_ontology.py

@@ -174,6 +174,7 @@ def remove_axiom(self, axiom: Union[OWLAxiom, Iterable[OWLAxiom]]):
         """
         pass
 
+    @abstractmethod
     def save(self, document_iri: Optional[IRI] = None):
         """Saves this ontology, using its IRI to determine where/how the ontology should be
          saved.

owlapy/iri.py

@@ -1,7 +1,7 @@
 """OWL IRI"""
 import weakref
 from abc import ABCMeta
-from typing import Final, Union, overload
+from typing import Final, Union
 from weakref import WeakKeyDictionary
 
 from owlapy import namespaces
@@ -49,45 +49,21 @@ def __init__(self, namespace: Union[str, Namespaces], remainder: str=""):
         self._namespace = sys.intern(namespace)
         self._remainder = remainder
 
-    @overload
     @staticmethod
-    def create(namespace: Namespaces, remainder: str) -> 'IRI':
-        ...
-
-    @overload
-    @staticmethod
-    def create(namespace: str, remainder: str) -> 'IRI':
-        """Creates an IRI by concatenating two strings. The full IRI is an IRI that contains the characters in
-        namespace + remainder.
-
-        Args:
-            namespace: The first string.
-            remainder: The second string.
-
-        Returns:
-            An IRI whose characters consist of prefix + suffix.
-        """
-        ...
-
-    @overload
-    @staticmethod
-    def create(string: str) -> 'IRI':
-        """Creates an IRI from the specified String.
-
-        Args:
-            string: The String that specifies the IRI.
-
-        Returns:
-            The IRI that has the specified string representation.
-        """
-        ...
-
-    @staticmethod
-    def create(string, remainder=None) -> 'IRI':
+    def create(iri:str | Namespaces, remainder:str=None) -> 'IRI':
+        assert isinstance(iri, str) | isinstance(iri, Namespaces), f"Input must be a string or an instance of Namespaces. Currently, {type(iri)}"
         if remainder is not None:
-            return IRI(string, remainder)
-        index = 1 + max(string.rfind("/"), string.rfind(":"), string.rfind("#"))
-        return IRI(string[0:index], string[index:])
+            assert isinstance(remainder,str), f"Reminder must be string. Currently, {type(remainder)}"
+            return IRI(iri, remainder)
+        else:
+            assert isinstance(iri, str) and remainder is None, \
+                f"iri must be string if remainder is None. Currently, {type(iri)} and {type(remainder)}"
+            # Extract reminder from input string
+            assert "/" in iri, "Input must contain /"
+            # assert ":" in iri, "Input must contain :"
+            assert " " not in iri, f"Input must not contain whitespace. Currently:{iri}."
+            index = 1 + max(iri.rfind("/"), iri.rfind(":"), iri.rfind("#"))
+            return IRI(iri[0:index], iri[index:])
 
     def __repr__(self):
         return f"IRI({repr(self._namespace)}, {repr(self._remainder)})"

owlapy/owl_axiom.py

@@ -1037,6 +1037,9 @@ class OWLDataPropertyAssertionAxiom(OWLPropertyAssertionAxiom[OWLDataPropertyExp
 
     def __init__(self, subject: OWLIndividual, property_: OWLDataPropertyExpression, object_: OWLLiteral,
                  annotations: Optional[Iterable['OWLAnnotation']] = None):
+        assert isinstance(subject,OWLIndividual), f"subject must be an OWLIndividual. Currently, {subject} of {type(subject)}"
+        assert isinstance(property_,OWLDataPropertyExpression), f"property_ must be an OWLDataPropertyExpression. Currently, {type(property_)}"
+        assert isinstance(object_,OWLLiteral), f"object_ must be an OWLLiteral. Currently, {type(object_)}"
         super().__init__(subject, property_, object_, annotations)
 
 

owlapy/owl_ontology.py

@@ -53,7 +53,6 @@
 })
 
 
-
 _VERSION_IRI: Final = IRI.create(namespaces.OWL, "versionIRI")
 
 _M = TypeVar('_M', bound='OWLOntologyManager')  # noqa: F821
@@ -983,20 +982,40 @@ def __init__(self, manager: _SM, path: Union[IRI, str], new: bool = False):
         self.manager = manager
         self.path = path
         self.new = new
+
+
         if isinstance(path, IRI):
             file_path = path.str
         else:
             file_path = path
         if new:  # create new ontology
             if isinstance(path, IRI):
-                self.owlapi_ontology = manager.get_owlapi_manager().createOntology(Stream.empty(),
+                self.owlapi_ontology = self.manager.get_owlapi_manager().createOntology(Stream.empty(),
                                                                                    owlapi_IRI.create(path.str))
             else:
                 raise NotImplementedError("Cant initialize a new ontology using path. Use IRI instead")
         else:  # means we are loading an existing ontology
-            self.owlapi_ontology = manager.get_owlapi_manager().loadOntologyFromOntologyDocument(File(file_path))
+            self.owlapi_ontology = self.manager.get_owlapi_manager().loadOntologyFromOntologyDocument(File(file_path))
         self.mapper = OWLAPIMapper(self)
 
+    def __eq__(self, other):
+        if isinstance(other, SyncOntology):
+            return other.owlapi_ontology.getOntologyID().equals(other.owlapi_ontology.getOntologyID())
+        return False
+
+    def __hash__(self):
+        return int(self.owlapi_ontology.getOntologyID().hashCode())
+
+    def __repr__(self):
+        return (f'SyncOntology:'
+                f'\t|Tbox|={len(self.get_tbox_axioms())}'
+                f'\t|Abox|={len(self.get_abox_axioms())}'
+                f'\t|Individuals|={len(self.individuals_in_signature())}'
+                f'\t|Classes|={len(self.classes_in_signature())}'
+                f'\t|Object Properties|={len(self.object_properties_in_signature())}'
+                f'\t|Data Properties|={len(self.data_properties_in_signature())}'
+                f'\n{self.manager}\tPath:{self.path}\tNew:{self.new}')
+
     def classes_in_signature(self) -> Iterable[OWLClass]:
         return self.mapper.map_(self.owlapi_ontology.getClassesInSignature())
 
@@ -1089,24 +1108,22 @@ def remove_axiom(self, axiom: Union[OWLAxiom, Iterable[OWLAxiom]]):
         else:
             self.owlapi_ontology.removeAxioms(self.mapper.map_(axiom))
 
-    def __eq__(self, other):
-        if isinstance(other, SyncOntology):
-            return other.owlapi_ontology.getOntologyID().equals(other.owlapi_ontology.getOntologyID())
-        return False
-
-    def __hash__(self):
-        return int(self.owlapi_ontology.getOntologyID().hashCode())
-
-    def __repr__(self):
-        return (f'SyncOntology:'
-                f'\t|Tbox|={len(self.get_tbox_axioms())}'
-                f'\t|Abox|={len(self.get_abox_axioms())}'
-                f'\t|Individuals|={len(self.individuals_in_signature())}'
-                f'\t|Classes|={len(self.classes_in_signature())}'
-                f'\t|Object Properties|={len(self.object_properties_in_signature())}'
-                f'\t|Data Properties|={len(self.data_properties_in_signature())}'
-                f'\n{self.manager}\tPath:{self.path}\tNew:{self.new}')
-
+    def save(self, path:str=None, document_iri: Optional[IRI] = None):
+        """
+        https://github.com/phillord/owl-api/blob/b2a5bfb9a0c6730c8ff950776af8f9bf19c78eac/
+                contract/src/test/java/org/coode/owlapi/examples/Examples.java#L206
+        """
+        assert isinstance(path,str), "Path must be a string"
+        from java.io import File
+        import org.semanticweb.owlapi.model.IRI
+        # //Create a file for the new format
+        file = File(path)
+        print(f"Saving Ontology into {path}")
+        if document_iri is None:
+            document_iri = org.semanticweb.owlapi.model.IRI.create(file.toURI())
+        else:
+            raise NotImplementedError("document_iri must be None for the time being")
+        self.manager.saveOntology(self.owlapi_ontology, self.manager.getOntologyFormat(self.owlapi_ontology), document_iri)
 
 OWLREADY2_FACET_KEYS = MappingProxyType({
     OWLFacet.MIN_INCLUSIVE: "min_inclusive",

owlapy/owl_ontology_manager.py

@@ -1,3 +1,4 @@
+import os.path
 from typing import Union
 
 import jpype
@@ -131,11 +132,19 @@ def create_ontology(self, iri: Union[IRI, str]) -> SyncOntology:
             assert isinstance(iri, IRI), "iri either must be string or an instance of IRI Class"
         return SyncOntology(self, iri, new=True)
 
-    def load_ontology(self, iri: Union[IRI, str]) -> SyncOntology:
-        return SyncOntology(self, iri, new=False)
+    def load_ontology(self, path: str) -> SyncOntology:
+        assert isinstance(path, str), "path either must be string or an instance of IRI Class"
+        assert os.path.exists(path), "path does not lead to an RDF knowledge graph."
+        return SyncOntology(self, path, new=False)
 
     def get_owlapi_manager(self):
         return self.owlapi_manager
 
     def apply_change(self, change: AbstractOWLOntologyChange):
         raise NotImplementedError("A change cannot be applied at the moment.")
+
+    def getOntologyFormat(self,*args):
+        return self.owlapi_manager.getOntologyFormat(*args)
+
+    def saveOntology(self,*args)->None:
+        self.owlapi_manager.saveOntology(*args)

owlapy/owl_reasoner.py

@@ -1009,7 +1009,7 @@ def __init__(self, ontology: Union[SyncOntology, str], reasoner="HermiT"):
             # https://owlcs.github.io/owlapi/apidocs_5/org/semanticweb/owlapi/apibinding/OWLManager.html
             self.manager = SyncOntologyManager()
             # OWLOntology
-            self.ontology = self.manager.load_ontology(iri=ontology)
+            self.ontology = self.manager.load_ontology(ontology)
 
         self._owlapi_manager = self.manager.get_owlapi_manager()
         self._owlapi_ontology = self.ontology.get_owlapi_ontology()

owlapy/util_owl_static_funcs.py

@@ -1,9 +1,15 @@
-from .owl_ontology import Ontology
+from .owl_ontology import Ontology, SyncOntology
 from .owl_ontology_manager import OntologyManager
-from typing import List
 from .class_expression import OWLClassExpression, OWLClass
+from .owl_individual import OWLNamedIndividual
 from .iri import IRI
-from .owl_axiom import OWLEquivalentClassesAxiom
+from .owl_axiom import OWLEquivalentClassesAxiom, OWLDataPropertyAssertionAxiom
+from .owl_property import OWLDataProperty
+from .owl_literal import OWLLiteral
+import os
+from typing import List
+from tqdm import tqdm
+import pandas as pd
 
 def save_owl_class_expressions(expressions: OWLClassExpression | List[OWLClassExpression],
                                path: str = 'predictions',
@@ -45,13 +51,77 @@ def save_owl_class_expressions(expressions: OWLClassExpression | List[OWLClassEx
 
     namespace= 'https://dice-research.org/predictions#' if namespace is None else namespace
     assert "#" == namespace[-1], "namespace must end with #"
-    # ()
+    # Initialize an Ontology Manager.
     manager = OntologyManager()
-    # ()
+    # Create an ontology given an ontology manager.
     ontology:Ontology = manager.create_ontology(namespace)
     # () Iterate over concepts
     for th, i in enumerate(expressions):
         cls_a = OWLClass(IRI.create(namespace, str(th)))
         equivalent_classes_axiom = OWLEquivalentClassesAxiom([cls_a, i])
         ontology.add_axiom(equivalent_classes_axiom)
-    ontology.save(path=path, inplace=False, rdf_format=rdf_format)
+    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):
+    """
+    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:
+
+    Example:
+        >>> from sklearn.datasets import load_iris
+        >>> import pandas as pd
+        # Load the dataset
+        >>> data = load_iris()
+        # Convert to DataFrame
+        >>> df = pd.DataFrame(data.data, columns=data.feature_names)
+        >>> df['target'] = data.target
+        # Save as CSV
+        >>> df.to_csv("iris_dataset.csv", index=False)
+        >>> print("Dataset saved as iris_dataset.csv")
+        >>> csv_to_rdf_kg("iris_dataset.csv")
+    """
+    from owlapy.owl_ontology_manager import SyncOntologyManager
+    assert path_csv is not None, "path cannot be None"
+    assert os.path.exists(path_csv), f"path **{path_csv}**does not exist."
+    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()
+    # Create an ontology given an ontology manager.
+    ontology:SyncOntology = manager.create_ontology(namespace)
+
+    # Read the CSV file
+    df = pd.read_csv(path_csv)
+    # () Iterate over rows
+    for index, row in tqdm(df.iterrows()):
+        individual=OWLNamedIndividual(f"{namespace}#{str(index)}".replace(" ","_"))
+        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(")", "")
+
+                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")
+
+    ontology.save(path=path_kg)

setup.py

@@ -10,6 +10,7 @@
     include_package_data=True,
     package_data={'owlapy': ['jar_dependencies/*.jar'],},
     install_requires=[
+        "scikit-learn>=1.5.2",
         "pandas>=1.5.0",
         "requests>=2.32.3",
         "rdflib>=6.0.2",

tests/test_save_owl_expressions.py → tests/test_owl_static_funcs.py

@@ -1,12 +1,18 @@
-from owlapy.util_owl_static_funcs import save_owl_class_expressions
+from owlapy.util_owl_static_funcs import save_owl_class_expressions, csv_to_rdf_kg
 from owlapy.class_expression import OWLClass, OWLObjectIntersectionOf, OWLObjectSomeValuesFrom
 from owlapy.owl_property import OWLObjectProperty
 from owlapy import owl_expression_to_sparql, owl_expression_to_dl
 from owlapy.owl_ontology_manager import OntologyManager
 from owlapy.owl_axiom import OWLDeclarationAxiom, OWLClassAssertionAxiom
 from owlapy.owl_individual import OWLNamedIndividual, IRI
+from sklearn.datasets import load_iris
+import pandas as pd
 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):
         # Using owl classes to create a complex class expression
@@ -22,4 +28,16 @@ def test_readme(self):
                                    path="owl_class_expressions.owl",
                                    rdf_format= 'rdfxml')
         g=rdflib.Graph().parse("owl_class_expressions.owl")
-        assert len(g)==22
+        assert len(g)==22
+
+    def test_csv_to_kg(self):
+        data = load_iris()
+        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")
+        onto = SyncOntologyManager().load_ontology(path_kg)
+        assert len(onto.get_abox_axioms()) == 750