ZarrDataset update for inference

Description

New capabilities were added to PatchSampler and ImageBase to handle padded samples and incomplete patches.
This allows to use ZarrDataset in an inference workflow and not only for training.

Additionally, patch sampling is now based on multiples of the patch_size instead of the chunk size from input files. This allow to sample patches from "incomplete" chunks (those smaller than the file's chunk size).

Changes

• Added allow_incomplete_patches parameter to PatchSampler class. This allows to retrieve patches from chunks that are smaller than the input file's chunk size. Previously, these chunks and all patches that could potentially be extracted were discarded.
• Added pad parameter to PatchSampler to add padding into all spatial axes when generating the selection slices.
• Added stride parameter to PatchSampler to introduce spacing between sampled patches.
• Updated ImageBase to handle indices that request pixels from outside the actual image by adding padding to cached chunks.

ZarrDataset available at PyPI with complete documentation

Documentation ready at https://thejacksonlaboratory.github.io/zarrdataset/index.html and the package for installing from PyPI

Release testing publication of ZarrDataset to PyPI

This release is for python 3.9 and should be able to publish ZarrDataset to PyPI

First release

First release of ZarrDataset for testing with TestPyPI