OC GraphEnricher enriches OpenCitations Data Model (OCDM) compliant knowledge graphs by finding missing identifiers and deduplicating entities.
Documentation: https://opencitations.github.io/oc_graphenricher/
pip install oc-graphenricherfrom 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.
Distributed under the ISC License. See LICENSE.