torchcell.verification.verify_morphology_dataset

verify_morphology_dataset(records: Sequence[dict[str, Any]], *, dataset_name: str, provenance: Provenance, expected_count: int, expected_incomplete: int = 0) → VerificationReport[source]

Run the L0-L3 record-level gate for the Ohya2005 CalMorph dataset.

Parameters:

  • records – the per-LMDB-entry dicts (experiment/reference/publication).

  • dataset_name – dataset identity for the report.

  • provenance – where the records came from.

  • expected_count – the mutant record-count oracle (Ohya2005 = 4718).

  • expected_incomplete – how many records the loader declares short of the full CalMorph vocabulary (Ohnuki 2022’s HRD1 row, two blank cells key-absent).

Returns:

A VerificationReport carrying the L0-L3 results. L4 (cross-source perturbed-gene overlap) is asserted across datasets by the caller.