cellmil.dataset¶

class cellmil.dataset.DatasetCreator(config: DatasetCreatorConfig)[source]¶

Bases: object

Class to create a dataset for MIL training based on the provided configuration.

__init__(config: DatasetCreatorConfig)[source]¶

_check_config()[source]¶: Check the configuration for any issues.

_read_excel()[source]¶: Read the Excel file specified in the configuration.

_preprocess_excel()[source]¶: Preprocess the Excel data to ensure paths are correctly formatted.

_get_magnification_openslide(path: str) → float | None[source]¶: Extract magnification from slide metadata using openslide.

_load_progress() → dict[str, list[str]][source]¶: Load previous progress from log.json if it exists.

_extract_features(patched_slide_path: Path, extractor: ExtractorType, wsi_path: Optional[Path] = None, graph_method: Optional[GraphCreatorType] = None, segmentation_model: Optional[ModelType] = None)[source]¶: Extract features from the segmented cells.

create()[source]¶: Create the dataset based on the configuration.

Modules

cellmil.dataset.dataset_creator