OC GraphEnricher

OC GraphEnricher enriches OpenCitations Data Model (OCDM) compliant knowledge graphs by finding missing identifiers and deduplicating entities.

Documentation: https://opencitations.github.io/oc_graphenricher/

Quick start

pip install oc-graphenricher

from oc_ocdm.graph.graph_set import GraphSet
from oc_ocdm.reader import Reader
from rdflib import Graph

from oc_graphenricher.enricher import GraphEnricher
from oc_graphenricher.deduplication import GraphDeduplicator
from oc_graphenricher.storage import single_file_storage

graph = Graph().parse("data/input.nt", format="nt11")

reader = Reader()
graph_set = GraphSet(base_iri="https://w3id.org/oc/meta/")
reader.import_entities_from_graph(
    graph_set,
    graph,
    enable_validation=False,
    resp_agent="https://w3id.org/oc/meta/prov/pa/2",
)

GraphEnricher(
    graph_set=graph_set,
    storage=single_file_storage(
        graph_path="enriched.json",
        provenance_path="provenance.json",
    ),
).enrich()
GraphDeduplicator(
    graph_set=graph_set,
    storage=single_file_storage(
        graph_path="deduplicated.json",
        provenance_path="provenance.json",
    ),
).deduplicate_and_save()

By default, GraphDeduplicator does not merge contributor roles only because author names are similar. To enable that opt-in behavior, pass merge_similar_named_contributors=True.

Use deduplicate() instead of deduplicate_and_save() when another application needs to manage storage or provenance output itself. Use preferred_survivors with a set of entity URIs to keep selected entities when duplicate clusters are merged. Without a preferred survivor, duplicate clusters keep the entity with more functional metadata. Ties use URI order. Use merge_clusters() when another application has already selected the merge clusters, for example from a reviewed CSV. The mapping key is the surviving entity URI and the values are the URIs to merge into it. This method does not discover or merge any extra duplicates outside the provided mapping.

For configuration options and usage details, see the documentation.

License

Distributed under the ISC License. See LICENSE.

To cite the latest version of this software (2.1.12), use this BibTeX entry:

@software{oc-graphenricher-2.1.12,
author = {Gabriele Pisciotta and Arcangelo Massari and Elia Rizzetto and Arianna Moretti and Ilaria De Dominicis and Silvio Peroni and Simone Persiani and Davide Brembilla},
title = {oc-graphenricher},
url = {https://archive.softwareheritage.org/swh:1:snp:6d8bfb5983216c17c071209ec59c48fe217b07ab;origin=https://github.com/opencitations/oc_graphenricher},
version = {2.1.12},
year = {2026}
}