RDFLib and jsonld

Use this pattern when JSON-LD is your interchange format, but RDF tooling is the best place to query, update, serialize, or store the graph.

Install PyLD:

pip install PyLD

PyLD 4 installs and uses RDFLib for RDF datasets. The processing order is:

1. Start with JSON-LD that has a usable @context.
2. Convert it with jsonld.to_rdf().
3. Process the returned rdflib.Dataset with RDFLib.
4. Convert it back with jsonld.from_rdf() if the next layer expects JSON-LD.
5. Compact the result with jsonld.compact() if developers or APIs should see terms such as name instead of full IRIs.

Example rdflib_processing.py

import json

from rdflib import Namespace, URIRef

from pyld import jsonld

SCHEMA = Namespace("http://schema.org/")
subject = URIRef("http://dbpedia.org/resource/Earth")

doc = {
    "@context": {"name": str(SCHEMA.name)},
    "@id": str(subject),
    "name": "Earth",
}

dataset = jsonld.to_rdf(doc)
dataset.add((subject, SCHEMA.url, URIRef("https://example.com/earth")))

jsonld_doc = jsonld.from_rdf(dataset)
compacted = jsonld.compact(
    jsonld_doc,
    {
        "name": str(SCHEMA.name),
        "homepage": {"@id": str(SCHEMA.url), "@type": "@id"},
    },
)

print(json.dumps(compacted, indent=2))

Output

{
  "@context": {
    "name": "http://schema.org/name",
    "homepage": {
      "@id": "http://schema.org/url",
      "@type": "@id"
    }
  },
  "@id": "http://dbpedia.org/resource/Earth",
  "name": "Earth",
  "homepage": "https://example.com/earth"
}

For simple lookups, use RDFLib's graph methods on the dataset returned by jsonld.to_rdf():

Example rdflib_query.py

import json

from rdflib import Namespace, URIRef

from pyld import jsonld

SCHEMA = Namespace("http://schema.org/")
subject = URIRef("http://dbpedia.org/resource/Earth")

doc = {
    "@context": {"name": str(SCHEMA.name)},
    "@id": str(subject),
    "name": "Earth",
}

dataset = jsonld.to_rdf(doc)
names = [str(value) for value in dataset.objects(subject, SCHEMA.name)]

print(json.dumps(names, indent=2))

Output

[
  "Earth"
]

For RDF that starts outside PyLD, parse it with RDFLib first, then pass the rdflib.Dataset to jsonld.from_rdf():

Example rdflib_parse.py

import json

from rdflib import Dataset

from pyld import jsonld

nquads = (
    '<http://dbpedia.org/resource/Earth> '
    '<http://schema.org/name> "Earth" .\n'
)
dataset = Dataset().parse(data=nquads, format="nquads")

doc = jsonld.from_rdf(dataset)
compacted = jsonld.compact(doc, {"name": "http://schema.org/name"})

print(json.dumps(compacted, indent=2))

Output

{
  "@context": {
    "name": "http://schema.org/name"
  },
  "@id": "http://dbpedia.org/resource/Earth",
  "name": "Earth"
}

Use format only when you need a serialized RDF string instead of an in-memory rdflib.Dataset:

nquads = jsonld.to_rdf(doc, {"format": "application/n-quads"})

Use jsonld.from_rdf(dataset) directly when you already have an rdflib.Dataset from another RDFLib parser, store, or query result.