torchcell.datasets.scerevisiae.bloom2019.Bloom2019Dataset

class Bloom2019Dataset(root: str = 'data/torchcell/bloom2019', io_workers: int = 0, genome: SCerevisiaeGenome | None = None, transform: Any | None = None, pre_transform: Any | None = None, **kwargs: Any)

Bases: ExperimentDataset

Bloom 2019: one record per (segregant, condition), 530,100 records.

organism: ClassVar[Organism] = 'scerevisiae'

each loader declares its own, the build’s organism switch reads it off the class (dataset_organism), and a loader that forgot it is refused by name instead of inheriting an organism it may not have.

Type:

The host organism every record describes. Annotated here WITHOUT a value

property experiment_class: type[Experiment]

Experiment schema class produced by this dataset.

property reference_class: type[ExperimentReference]

Experiment-reference schema class produced by this dataset.

property raw_file_names: list[str]

The data files (phenotypes, 16 genotype matrices, README, xls); code and paper are read from the mirror by the verifier, not symlinked here.

download() → None

Symlink the manifest-listed mirror files into raw/ and verify each sha256.

control_spec(spec: ConditionSpec) → ConditionSpec

The served condition whose environment is the control plate’s environment.

process() → None

Build 13,950 x 38 records, one cross at a time, interning each genotype.

static extract_systematic_gene_names(genotype: dict[str, Any]) → list[str]

Not applicable: a mosaic genotype has no gene-keyed perturbations.

compute_gene_set() → GeneSet

S288C genes overlapping at least one haplotype block of any cross.

Computed once from the sorted marker tables kept during process (or rebuilt from raw/ when loading a built store), so it never walks 530k records.

preprocess_raw(df: DataFrame, preprocess: dict[str, Any] | None = None) → DataFrame

Preprocessing happens inside process() for this dataset.

create_experiment() → None

Records are assembled inside process(); see _iter_cross_genotypes.