RDF Integration

The DDI Toolkit provides deep integration with RDF (Resource Description Framework) using the DataArtifex RDF Toolkit. This allows DDI metadata to be handled as high-level Python objects while remaining fully compatible with semantic web technologies.

Overview

The toolkit uses Pydantic models that are annotated with RDF property mappings. This approach provides several benefits:

  • Type Safety: Full Pydantic validation for all DDI attributes and associations.

  • Native RDF Mapping: Automatic conversion between Python fields and RDF predicates using the dartfx.rdf package.

  • Seamless Serialization: Built-in functionality to generate rdflib graphs for export to Turtle, JSON-LD, or other RDF formats.

  • Identifier Management: Automated handling of IRDI (International Registration Data Identifier) and URI generation.

The Definitive Model (v1.0.0)

The module dartfx.ddi.ddicdi.model_1_0_0 contains the definitive set of DDI-CDI 1.0 resources. These classes inherit from CDIResource, which provides the core RDF integration.

Core Resource Classes:

  • Agent: Represents individuals, organizations, or software systems.

  • InstanceVariable: Describes the use of a variable within a specific dataset.

  • DataStructure: Defines the logical and physical layout of data.

  • DataSet: A collection of data points organized by a data structure.

  • Activity: Records information about data collection and processing steps.

Direct Model Usage

While the DDI Cross Domain Integration (CDI) (Assistant framework) is the recommended way to build DDI-CDI objects, you can work with the models directly:

from dartfx.ddi.ddicdi import model_1_0_0 as model
from rdflib import URIRef

# Create a resource instance
var = model.InstanceVariable()

# Setting an ID is crucial to avoid blank nodes during serialization
var.id = "http://example.org/variables/age_1"

# Populate properties
var.set_simple_name("AGE")

# Generate an RDF Graph
graph = var.to_rdf_graph()
print(graph.serialize(format="turtle"))

SHACL Validation

The toolkit includes support for validating generated RDF graphs against the official DDI-CDI SHACL (Shapes Constraint Language) rules. This ensures that the generated metadata conforms to the specification’s structural and cardinality requirements.

Legacy RDF Implementation

Note

The initial implementation of RDF support used an earlier RDF object layer. Those legacy modules have been removed as the project migrated to the definitive model_1_0_0.py based on the newer RDF Toolkit.