Linked Data (LD)

Linked Data (LD)#

The linked data module enables RDF (Resource Description Framework) integration with HDF5 files, bringing semantic web capabilities to your scientific data.

Background#

While HDF5 files provide a hierarchical structure for storing data and metadata, they lack standardized semantics. RDF provides a framework for describing resources with unique identifiers and relationships, enabling:

  • Global uniqueness: Persistent identifiers (e.g., ORCID, DOI) for researchers and data

  • Interoperability: Common vocabularies and ontologies for domain-specific terms

  • Machine-readable metadata: Enable automated discovery and integration

  • Knowledge graphs: Build interconnected datasets that can be queried with SPARQL

The h5rdmtoolbox ld module bridges HDF5 and RDF by enabling:

  1. Structural RDF: Automatically converting HDF5 structure to RDF (groups, datasets, attributes)

  2. Contextual RDF: Mapping HDF5 attributes to semantic concepts via ontologies

  3. SHACL validation: Validating HDF5 data against RDF shapes

  4. Serialization: Exporting to JSON-LD, Turtle, and other RDF formats

Key Concepts#

Structural RDF

Automatically generated RDF from HDF5 structure (groups, datasets, datatypes).

Contextual RDF

User-defined mappings from HDF5 attributes to semantic concepts.

RDF Mappings

Define how HDF5 attributes map to RDF predicates and objects.

SHACL Shapes

RDF shapes that HDF5 metadata must conform to.

Command Line and Web Viewer#

The h5tbx command exposes the linked-data tools without writing Python code. Use ld dump to serialize a file directly from the shell. See Command Line Serialization for details.

h5tbx ld dump example.h5
h5tbx ld dump example.h5 --format json-ld --file-uri https://example.org/data/ --prefix ex
h5tbx ld dump example.h5 --structural=false
h5tbx ld dump example.h5 --contextual=false

By default, structural and contextual RDF are both included. Use --structural=false or --contextual=false to restrict the output. The --file-uri option defines stable subject IRIs; --prefix binds a compact prefix for that URI in serializations that support prefixes.

For interactive inspection, start the local web viewer. See Local Web Viewer for the available pages and graph controls.

h5tbx serve example.h5
h5tbx serve data/ --h5ext=.h5 --h5ext=.hdf5
h5tbx serve example.h5 --file-uri https://doi.org/10.5281/zenodo.17572275# --local-iri-pattern "https://doi.org/10.5281/zenodo.*"

h5tbx serve accepts files and folders. Folder inputs are searched for .h5, .hdf, and .hdf5 files by default; repeat --h5ext to limit folder discovery to specific extensions.

Graph Metrics from Python#

The metrics shown in the web viewer are also available from Python. Use h5rdmtoolbox.compute_metrics for a filename, or File.metrics for an opened file:

import h5rdmtoolbox as h5tbx

metrics = h5tbx.compute_metrics("example.h5")

with h5tbx.File("example.h5") as h5:
    metrics = h5.metrics()

The returned dictionary contains the same RDF knowledge-graph metrics as the web page, including graph size, predicate and class usage, connectivity, datatype distribution, label coverage, and external namespace usage.