datamodules.RolfFormat.datasets package

Submodules

datamodules.RolfFormat.datasets.dataset module

Load a dataset of historic documents by specifying the folder where its located.

class DatasetRolfFormat(dataset_specs: List[DatasetSpecs], image_dims: ImageDimensions, is_test: bool = False, image_transform: Optional[callable] = None, target_transform: Optional[callable] = None, twin_transform: Optional[callable] = None)[source]

Bases: Dataset

Dataset that loads the data in the Rolf format. Each file name has a fixed structure of name_{file_number}.jpg. The file number is a number between 0 and 9999. The different splits are defined by giving a range and a root folder for each split.

Parameters:

  • dataset_specs (List[DatasetSpecs]) – The dataset specs that specify the location of the data and ground truth files.

  • image_dims (ImageDimensions) – The dimensions of the images.

  • is_test (bool) – Is it the test dataset?

  • image_transform (callable) – Transformations that should be applied to the image.

  • target_transform (callable) – Transformations that should be applied to the ground truth.

  • twin_transform (callable) – Transformations that should be applied to both the image and the ground truth.

static get_img_gt_path_list(list_specs: List[DatasetSpecs]) → List[Tuple[Path, Path]][source]

Returns a list of tuples that contain the path to the gt and image that belong together.

Parameters:

list_specs (List[DatasetSpecs]) – The dataset specs that specify the location of the data and ground truth files.

Returns:

A list of tuples that contain the path to the gt and image that belong together.

Return type:

List[Tuple[Path, Path]]

class DatasetSpecs(data_root: str, doc_dir: str, doc_names: str, gt_dir: str, gt_names: str, range_from: int, range_to: int)[source]

Bases: object

This class is used to specify the location of the data and ground truth files. It can also be used to specify a range of files that should be used. This is useful if you want to split the data into train/val/test and want to use the same data root for all three splits.

data_root: str
doc_dir: str
doc_names: str
gt_dir: str
gt_names: str
range_from: int
range_to: int

Module contents