lodkit 0.1.0

No project description provided

LODKit

LODKit is a collection of Linked Open Data related Python functionalities.

LODkit (includes|will include)

• a custom rdflib.Graph subclass that is capable of
  • RDFS and OWL-RL inferencing
  • Bnode-safe graph merging [todo]
• a custom importer for loading RDF files as if they where Python modules
• [...]

Examples

lodkit.Graph

lodkit.Graph is an rdflib.Graph subclass that is cabable of RFDS and OWL-RL inferencing.

The default plugin for inferencing is the owlrl native Python inferencing engine. The deductive closure type used for lodkit.Graph is RDFS_OWLRL_Semantics which allows for RDFS and OWL-RL reasoning.

from lodkit import Graph
from rdflib import Namespace

ex = Namespace("http://example.org/")

graph = Graph()
graph.add((ex.subj, ex.pred, ex.obj))
graph.add((ex.inverse, OWL.inverseOf, ex.pred))

len(graph)                              # 2
(ex.obj, ex.inverse, ex.subj) in graph  # False

graph.inference(reasoner="owlrl") 

len(graph)                              # 359
(ex.obj, ex.inverse, ex.subj) in graph  # True

The reasoner parameter of lodkit.Graph.inference (so far) also takes

• "rdfs" for owlrl's RDFS deductive closure type,
• "reasonable" for the reasonable inference engine and
• "allegro" for the Allegrograph reasoner (using the RDFS++ and OWL-RL).

Also the reasoner parameter takes Reasoner objects directly, see reasoners.py.

lodkit.importer

lodkit.importer is a custom importer for importing RDF files as if they where regular Python modules. RDF files are parsed into rdflib.Graph instances and made available in the module namespace.

E.g. in a directory structure

├── dir/
│   ├── main.py
│   ├── some_rdf.ttl
│   ├── subdir/
│       └── some_more_rdf.xml

the following creates rdflib.Graph instances in the current module namespace:

# main.py
import lodkit.importer

import some_rdf
from subdir import some_more_rdf

print(type(some_rdf))       # <class 'rdflib.graph.Graph'>
print(type(some_more_rdf))  # <class 'rdflib.graph.Graph'>

I find this really convenient for bulk-parsing graphs (example).