rsm-schema

This repository defines the Research Software Management (RSM) Metadata Schema and its Python representation.

The package includes:

• the JSON Schema itself
• generated Pydantic v2 models for validating RSM documents
• a small inspection API for traversing the JSON Schema

The documentation is available at lumc-dcc.github.io/rsm-schema.

Quick start

pip install rsm-schema

Python 3.14 is required.

Validate a document

RSMMetadata is the document root. Only project_slug is required.

from rsm_schema import RSMMetadata, validate_document

document = {
    "project_slug": "my-project",
    "project_name": "My Project",
    "development_status": "wip",
    "topics": {
        "entries": [
            {
                "term": "Data analysis",
                "uri": "https://edamontology.org/topic_3474",
            }
        ]
    },
    "contributors": {
        "entries": [
            {
                "name": "Ada Lovelace",
                "email": "ada@example.org",
                "roles": ["Original author", "Maintainer"],
                "affiliations": [
                    {"name": "Leiden University Medical Center"},
                    {"name": "Leiden University"},
                ],
            }
        ]
    },
}

validate_document(document)
metadata = RSMMetadata.model_validate(document)

print(metadata.contributors.entries[0].affiliations[0].name)
print(metadata.topics.entries[0].term)

topics describes the project-level research domains using EDAM terms. Function-specific software_functions separately describes operations, inputs, outputs, commands, and notes.

validate_document() is the authoritative validation path and enforces JSON Schema conditionals that generated Pydantic models cannot represent.

Inspect the schema

from rsm_schema import schema

print(schema.title)

for field in schema.root.properties:
    print(field.name, field.types, field.required)

person = schema.definition("person")
for field in person.properties:
    print(field.pointer, field.types, field.required)

SchemaNode.resolve() follows local JSON Pointer $ref values, and Schema.walk() traverses properties, definitions, array items, and composition keywords such as oneOf, anyOf, and allOf. External references are never fetched.

Development

git clone https://github.com/lumc-dcc/rsm-schema.git
cd rsm-schema
poetry install --with dev,docs
poetry run pre-commit install

poetry run pre-commit run --all-files
poetry run pytest

The JSON Schema at src/rsm_schema/schema/rsm.schema.json is canonical. After changing it, regenerate the models and commit both together:

poetry run datamodel-codegen
poetry run pytest

Build the documentation locally with:

poetry run sphinx-build -b html -W --keep-going docs docs/_build/html

Released schemas are archived under schema-archive/, one immutable directory per version. The documentation site publishes every archived version so versioned $refs continue to resolve. See CONTRIBUTING.md and the schema workflow for the release steps.

License

This project is licensed under Apache 2.0. See the LICENSE file for details.