Parse JSON-LD as list?

[edsu]

👋 If I already have some JSON-LD as a dict and would like to create a rdflib Graph from it, is there a way to do that?

For example:

test_json = {
    "@context": {
        "dc": "http://purl.org/dc/terms/",
        "rdf": "http://www.w3.org/1999/02/22-rdf-syntax-ns#",
        "rdfs": "http://www.w3.org/2000/01/rdf-schema#"
    },
    "@id": "http://example.org/about",
    "dc:title": {
        "@language": "en",
        "@value": "Someone's Homepage"
    }
}

g = Graph().parse(data=test_json, format='json-ld')

[edsu]

Actually, my example seems to work just fine :-) closing!

[edsu]

FWIW I think the error message was confusing me when I happened to be passing in a list instead of a dict, which can be legit JSON-LD. For example:

import rdflib

# json is a list instead of a dict

test_json = [{
    "@context": {
        "dc": "http://purl.org/dc/terms/",
        "rdf": "http://www.w3.org/1999/02/22-rdf-syntax-ns#",
        "rdfs": "http://www.w3.org/2000/01/rdf-schema#"
    },
    "@id": "http://example.org/about",
    "dc:title": {
        "@language": "en",
        "@value": "Someone's Homepage"
    }
}]

g = rdflib.Graph().parse(data=test_json, format='json-ld')

which results in:

Traceback (most recent call last):
  File "/Users/edsu/Projects/rdflib/y.py", line 16, in <module>
    g = rdflib.Graph().parse(data=test_json, format='json-ld')
  File "/Users/edsu/Projects/rdflib/rdflib/graph.py", line 1658, in parse
    source = create_input_source(
        source=source,
    ...<4 lines>...
        format=format,
    )
  File "/Users/edsu/Projects/rdflib/rdflib/parser.py", line 756, in create_input_source
    raise RuntimeError(f"parse data can only str, or bytes. not: {type(data)}")
RuntimeError: parse data can only str, or bytes. not: <class 'list'>

The message RuntimeError: parse data can only str, or bytes. not: <class 'list'> is a bit misleading since you can also pass in a dict?

Whereas it works ok if I serialize the JSON as a str and parse that:

import json
import rdflib

test_json = [{
    "@context": {
        "dc": "http://purl.org/dc/terms/",
        "rdf": "http://www.w3.org/1999/02/22-rdf-syntax-ns#",
        "rdfs": "http://www.w3.org/2000/01/rdf-schema#"
    },
    "@id": "http://example.org/about",
    "dc:title": {
        "@language": "en",
        "@value": "Someone's Homepage"
    }
}]

g = rdflib.Graph().parse(data=json.dumps(test_json), format='json-ld')

Ideally I think you could pass in a list as well to avoid serializing and reparsing the JSON?

[multimeric]

As you have noted, the data argument accepts only a JSON string. This correctly gets picked up by a type checker like pyright.

There is an internal API to_rdf that I sometimes use to avoid having to re-parse the JSON, although I probably shouldn't. e.g. you can replace the last line of your code with:

from rdflib.plugins.parsers.jsonld import to_rdf

graph = to_rdf(test_json, rdflib.Graph())
print(list(graph))