bioimageloader

Root module of bioimageloader

Expose core classes and functions

class bioimageloader.BatchDataloader(dataset: bioimageloader.base.Dataset, batch_size: int = 16, shuffle: bool = False, drop_last: bool = False, num_workers: Optional[int] = None)

Batch loader with multi-processing

class bioimageloader.ConcatDataset(datasets: List[bioimageloader.base.Dataset])

Concatenate Datasets

References

1

https://pytorch.org/docs/stable/_modules/torch/utils/data/dataset.html#ConcatDataset

class bioimageloader.Config(filename=None)

Construct config from a yaml file

Parameters

filenamestr

Path to config file

load_datasets(transforms: Optional[Union[albumentations.core.composition.Compose, Dict[str, albumentations.core.composition.Compose]]] = None) → bioimageloader.types.DatasetList

Load multiple datasets from a yaml file

Note that when you provide a dictionray for transforms, keys should be the class names, not their acronyms.

Parameters

configconfiguration object

Config instance that contains acronyms and arguments to initialize each dataset

transformsalbumentations.Compose or dictionary, optional

Either apply a single composed transformations for every datasets or pass a dictionary that defines transformations for each dataset with keys being the class names of collections.

replace_commonpath(new: str)

Replace common path for all root_dir with a new one

All root_dir should have a commonpath. Do not put trailing ‘/’ at the end.

You made a config with root_dir being relative path. You do not need to replace them manually with this method.

set_training(val: bool)

Iterate config and set all training to given value

It only affects those that have training kwarg.

set_ouput(val: str)

Iterate config and set all output to given value

It only affects those that have output kwarg.

set_grayscale(val: bool)

Iterate config and set all grayscale to given value

It only affects those that have grayscale kwarg.