Skip to content

Latest commit

 

History

History
50 lines (33 loc) · 2.61 KB

README.md

File metadata and controls

50 lines (33 loc) · 2.61 KB
Raw

JSON-LD in Python

There are two major Python libraries with different strengths and limitations.

rdflib

rdflib is the go-to library for working with RDF in Python. It supports parsing and serializing RDF in many formats (including JSON-LD via the rdflib-jsonld extension), supports in memory generation and manipulation of RDF graphs, and has connectors for SPARQL etc.. However, even with the rdflib-jsonld extension, support for JSON-LD is limited, and the functions of a JSON-LD processor are not implemented.

PyLD

PyLD is described as an "implementation of the JSON-LD specification in Python". It uses native JSON-LD data stuctures and implements the functions of a JSON-LD processor (compaction, expansion, framing, etc.). However, it has no facilities for working with data as RDF triples, or taking data from RDF sources except by reparsing a JSON-LD serialization.

Using rdflib and PyLD together?

Because rdflib and PyLD use different in-memory data structures it is not trivial to use the powerful RDF support of rdflib in conjuction with the JSON-LD processor implementation of PyLD. A small compatibility library rdflib_pyld_compat.py provides a bridge from rdflib to PyLD formats without the need to serialize and then re-parse the data.

Warning: rdflib_pyld_compat.py was written while preparing this demonstration as has not been thoroughly exercised or tested.

A simple example program select_anno.py illustrates operations that one might want to do in both libraries: first building and manipulating an RDF graph with rdflib, then processing and tweaking the JSON with PyLD. The program take one command line argument which is the URI of the the annotation target used to select annotations. Running:

jsonld-in-python> python select_anno.py http://example.org/hand2

produces:

{
  "@context": "http://example.org/context",
  "id": "http://example.org/anno1",
  "motivation": "http://example.org/Drawing",
  "on": "http://example.org/hand2",
  "resource": "http://example.org/hand1",
  "type": "oa:Annotation"
}

See the select_anno.py code to see use of pyld_json_from_rdflib_graph(...) to get PyLD data from an rdflib graph:

from rdflib_pyld_compat import pyld_json_from_rdflib_graph
...

# Get JSON-LD object in PyLD form
jld = pyld_json_from_rdflib_graph(g2)

| Index | Next: JSON-LD and Framing |