torchcell.knowledge_graphs.releases.content_hashes_from_csv

content_hashes_from_csv(out_dir: Path) → dict[str, str][source]

Dataset.id -> content_sha256 from a BioCypher output directory.

Reads the ExperimentMemberOf edge files: :START_ID is the experiment id, :END_ID the dataset id (the dataset class name). Every id is held in memory, grouped by dataset, so this belongs on the build node.