This package helps you in generating ontology-related objects and lets you easily create JSON-LD files.
Install the package:
pip install ontolutilsImagine you want to describe a prov:Person with a first name, last name and an email address but writing
the JSON-LD file yourself is too cumbersome and you want validation of the parsed parameters. The package
lets you design classes, which describe ontology classes like this:
from pydantic import EmailStr, Field
from pydantic import HttpUrl, model_validator
from ontolutils import Thing, urirefs, namespaces, as_id
@namespaces(prov="https://www.w3.org/ns/prov#",
foaf="https://xmlns.com/foaf/0.1/",
m4i='http://w3id.org/nfdi4ing/metadata4ing#')
@urirefs(Person='prov:Person',
firstName='foaf:firstName',
lastName='foaf:lastName',
mbox='foaf:mbox',
orcidId='m4i:orcidId')
class Person(Thing):
firstName: str
lastName: str = Field(default=None, alias="last_name") # you may provide an alias
mbox: EmailStr = Field(default=None, alias="email")
orcidId: HttpUrl = Field(default=None, alias="orcid_id")
# the following will ensure, that if orcidId is set, it will be used as the id
@model_validator(mode="before")
def _change_id(self):
return as_id(self, "orcidId")
p = Person(id="https://orcid.org/0000-0001-8729-0482",
firstName='Matthias', last_name='Probst')
# as we have set an alias, we can also use "lastName":
p = Person(id="https://orcid.org/0000-0001-8729-0482",
firstName='Matthias', lastName='Probst')
# The jsonld representation of the object will be the same in both cases:
json_ld_serialization = p.model_dump_jsonld()
# Alternatively use
serialized_str = p.serialize(format="json-ld") # or "ttl", "n3", "nt", "xml"The result of the serialization is shown below:
{
"@context": {
"owl": "http://www.w3.org/2002/07/owl#",
"rdfs": "http://www.w3.org/2000/01/rdf-schema#",
"prov": "https://www.w3.org/ns/prov#",
"foaf": "https://xmlns.com/foaf/0.1/",
"m4i": "http://w3id.org/nfdi4ing/metadata4ing#"
},
"@id": "https://orcid.org/0000-0001-8729-0482",
"@type": "prov:Person",
"foaf:firstName": "Matthias",
"foaf:lastName": "Probst"
}If you cannot define the class statically as above, you can also define it dynamically:
from typing import List, Union
from ontolutils import build, Property, Thing
Event = build(
namespace="https://schema.org/",
namespace_prefix="schema",
class_name="Event",
properties=[Property(
name="about",
default=None,
property_type=Union[Thing, List[Thing]]
)]
)
conference = Event(label="my conference", about=[Thing(label='The thing it is about')])
ttl = conference.serialize(format="ttl")The serialization in turtle format looks like this:
@prefix owl: <http://www.w3.org/2002/07/owl#> .
@prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#> .
@prefix schema: <https://schema.org/> .
[] a schema:Event ;
rdfs:label "my conference" ;
schema:about [ a owl:Thing ;
rdfs:label "The thing it is about" ] .Please visit the documentation for more information.
