Getting started

ontouml-json2graph requires Python 3.10 or newer. Install the published package with:

pip install ontouml-json2graph

Convert one file

The command-line interface writes the transformed graph to the current working directory by default. The default serialization is Turtle (ttl).

python -m json2graph.decode -i my_ontology.json

Choose another output directory or serialization with -o and -f:

python -m json2graph.decode -i my_ontology.json -o results -f json-ld

The output filename uses the input stem and the selected format as its extension. The command above writes results/my_ontology.json-ld.

Decode in Python

Use the library interface when Python code should receive an RDFLib Graph:

from json2graph.library import decode_json_project

graph = decode_json_project("my_ontology.json")

Use decode_json_model instead when diagrammatic and project-level resources should be removed.

Run the canonical example

The repository contains one minimal project used by both the CLI and library examples. Run the tested commands from docs/examples. The Examples page includes this minimal input and a complete catalog model with further command-line and Python recipes.

Choose the output scope

  • Complete-project decoding retains model, project, and supported diagrammatic resources.

  • Model-only decoding retains domain-level model resources, including enumeration literals, and removes packages, project, and diagrammatic resources.

Both routes use the same transformation core and policy defaults.

Know the input boundary

The decoder expects the OntoUML JSON structure. It parses JSON but does not perform JSON Schema validation. UTF-8 is preferred; input that cannot be decoded as UTF-8 is retried as CP1252 and produces a warning.

Complete-project output is not necessarily lossless. Read Limitations and diagnostics before relying on the RDF to reconstruct a source document.

Continue with the command-line guide, Python library guide, or Policies and configuration.