cellmil.datamodels.transforms.label_pipeline

Pipeline for composing multiple label transforms.

Classes

LabelTransformPipeline(transforms)

Pipeline for applying multiple label transforms in sequence.

class cellmil.datamodels.transforms.label_pipeline.LabelTransformPipeline(transforms: List[LabelTransform])

Bases: object

Pipeline for applying multiple label transforms in sequence.

This class manages a sequence of label transforms, ensuring that fittable transforms are properly fitted on training data before being applied.

__init__(transforms: List[LabelTransform])

Initialize the pipeline with a list of transforms.

Parameters:

transforms – List of LabelTransform instances to apply in sequence

fit(labels: Dict[str, Union[int, Tuple[float, int]]], **kwargs: Any) → LabelTransformPipeline

Fit all fittable transforms in the pipeline on training labels.

Parameters:

• labels – Training labels dictionary

• **kwargs – Additional keyword arguments passed to each transform’s fit method

Returns:

Self for method chaining

transform_labels(labels: Dict[str, Union[int, Tuple[float, int]]]) → Dict[str, Union[int, Tuple[float, int]]]

Apply all transforms in the pipeline sequentially.

Parameters:

labels – Labels dictionary to transform

Returns:

Transformed labels dictionary after applying all transforms

fit_transform(labels: Dict[str, Union[int, Tuple[float, int]]], **kwargs: Any) → Dict[str, Union[int, Tuple[float, int]]]

Fit the pipeline and apply it to labels.

Parameters:

• labels – Labels to fit and transform

• **kwargs – Additional keyword arguments for fitting

Returns:

Transformed labels

get_config() → Dict[str, Any]

Get configuration for all transforms in the pipeline.

Returns:

Dictionary containing pipeline configuration

save(directory: Path) → None

Save the pipeline and all its transforms to disk.

Parameters:

directory – Directory to save the pipeline configuration and transforms

classmethod load(directory: Path) → LabelTransformPipeline

Load a pipeline from disk.

Parameters:

directory – Directory containing the saved pipeline

Returns:

LabelTransformPipeline instance

__len__() → int

Return the number of transforms in the pipeline.

__getitem__(idx: int) → LabelTransform

Get a transform by index.