torch_adata

Subpackages

Package Contents

Classes

AnnDataset

AnnDataset Module for formatting AnnData as a pytorch Dataset.

LightningAnnDataModule

Attributes

__module_name__

__doc__

__author__

__email__

__version__

torch_adata.__module_name__ = __init__.py
torch_adata.__doc__ = Main __init__ module.
torch_adata.__author__
torch_adata.__email__
torch_adata.__version__ = 0.0.20
class torch_adata.AnnDataset(adata: anndata.AnnData, use_key: str, groupby: str = None, obs_keys: list([str, '...', str]) = None, attr_names: dict({'obs': [str, '...', str], 'aux': [str, '...', str]}) = {'obs': [], 'aux': []}, one_hot: list([bool, '...', bool]) = False, aux_keys: list([str, '...', str]) = None, silent: bool = False, sampling_weight_key=None)

Bases: torch.utils.data.Dataset

AnnDataset Module for formatting AnnData as a pytorch Dataset.

property n_dims
__len__()
__getitem__(idx)
__repr__() → str
class torch_adata.LightningAnnDataModule(adata=None, h5ad_path=None, batch_size=2000, num_workers=os.cpu_count(), train_val_split=[0.8, 0.2], n_predict=2000, use_key='X_pca', groupby='Time point', train_key='train', val_key='val', test_key='test', predict_key='predict', shuffle=True, silent=True, **kwargs)

Bases: pytorch_lightning.LightningDataModule

property properly_formatted_index
property adata
property cell_idx
property n_cells
property n_features
property n_dims
property data_keys
property init_train_adata
property train_adata
property val_adata
property test_adata
property predict_adata
property AnnDatasetKWARGS
property train_dataset
property val_dataset
property test_dataset
property predict_dataset
_format_adata_obs_index()
_configure_adata()

configures the property self.adata

configure_train_val_split()
subset_adata(key: str)
to_dataset(key: str) → torch.utils.data.Dataset

key funciton to transform adata -> torch.utils.data.Dataset

_return_loader(dataset_key)
prepare_data()
setup(stage)
train_dataloader()
val_dataloader()
test_dataloader()
predict_dataloader()
__repr__()