optimuskg 0.1.0

Client library for loading the OptimusKG biomedical knowledge graph from Harvard Dataverse.

optimuskg

Python client for loading the OptimusKG biomedical knowledge graph from Harvard Dataverse.

Install

pip install optimuskg

Usage

import optimuskg

# Fetch any file from the gold parquet folder by relative path.
# First call downloads from Dataverse and caches locally; subsequent calls reuse the cache.
path = optimuskg.get_file("nodes/gene.parquet")

# Load a single file as a Polars DataFrame.
drugs = optimuskg.load_parquet("nodes/drug.parquet")

# Load the full graph (or the largest connected component) as two Polars DataFrames.
nodes, edges = optimuskg.load_graph(lcc=True)

# Load the graph as a NetworkX MultiDiGraph with JSON properties parsed onto nodes/edges.
G = optimuskg.load_networkx(lcc=True)

Configuration

Downloads are cached in platformdirs.user_cache_dir("optimuskg") by default (~/Library/Caches/optimuskg on macOS, ~/.cache/optimuskg on Linux). Override with $OPTIMUSKG_CACHE_DIR or optimuskg.set_cache_dir(path).

To point at a different dataset (e.g. a pre-release), set $OPTIMUSKG_DOI or call optimuskg.set_doi("doi:10.xxxx/XXXX").

TODO: the baked-in DOI (doi:10.7910/DVN/IXA7BM) is a stub. Update it to the published DOI after the Dataverse release is live, and verify that optimuskg.load_graph(lcc=True) resolves without a 404.

License

MIT — see LICENSE. Use of the OptimusKG dataset itself is subject to the licenses of the constituent data sources; see the project license docs.