New features:
- Added tests (WIP)
- Addded documentation (WIP)
- Switched from creationism to dicfg confiuguration support, including command line interface and environment ways to configure the batch iterator
- File collection is now based on sourcelib
- ASAP<2.1 depreciation warnings
bug fixes:
- asap annotation writer scaling
Breaking changes
- renamed
- accessoires -> interoperability
- label_map -> labels
- parser hooks-> callbacks
- dataset.annotations_* -> annotation_counts_*
- dataset.pixel_counts -> dataset.pixels
- wsi.get_annotation(Annotation) -> wsi.get_region_from_annotations([Annotation])
- in the sample method of the samplers, point is now a tuple instead of a shapely.geometry.Point
- Creating of objects in user config, see dicfg documentation
- Shapely2.0 support: Annotations are now based on composition instead of inheritance, so the actual geometry should be accessed by annotation.geometry
- removed jsonschema
- dicom support
- improved sliding window hook annotations and index
- added path property to annotation
- ASAP v1.9 compatibilty fix (contributed by Leander van Eekelen)
- added ASAP interpolation argument (contributed by Witali Aswolinskiy)
- added save normal image as pyramidal tif (contributed by Witali Aswolinskiy)
bug fixes:
- install of 0.0.14 failed due to not ruamel import error.
new features:
- updated to ASAP2.1
- cucim image backend
- tiffslide image backend (Contributed by Rishi Jumani)
- s3 asap parser (Contributed by Rishi Jumani)
- switched to rtree
- updated patch iterator
- get shape_from_spacing method in wsi
- loading slides < 2.0 spacing gives warning
- added check coordinates for point
- method for annotations and sampling annotations per label
- grand-challenge annotations to asap format
- mask patch with annotations
- load empty files gives warning
- added writer factory functions
breaking changes
- removed redundant shift coordinates method
docs:
- included binder files
requirements:
- added
- jsonschema>=4.4.0
- matplotlib>=3.3.1
- rtree>=1.0.0
bug fixes:
- count for specific sample labels
- hed callback
- wsi file reference
- virtum parser
- areaannotationsampler
- plotting: fixed plot batch function
- plotting: scaling in base coordinates in plot annotations (Contributed by Robin Lomans)
refactoring
- renamed orginal_path to original_path in File
new features:
- added patch iterator
bug fixes:
- install of 0.0.12 failed due to not finding requirements.txt. Requirements are now again specified in the setup.py.
breaking changes
- specific code for external software has been moved to accessories
- label weights in WeightedLabelSampler are now specified via a dictionary
refactoring
- parsing of annotation files has been refactored and now includes a JSON schema
- enhance shape and spacing error messages (Contributed by Jakub Kaczmarzyk)
new features:
- accessories are loaded with package import
- WholeSlideAnnotationParser based on JSON schema
- JSON schema can be used to serialize annotations
- albumentations callback (contributation by Thijs Gelton)
- HedCallback augmentation
- 'none' labels are sorted below other labels
- associater classes StemSplitterAssociater and AnyOneAssociater
- exact match option for associate_files
- clip polygon and check for valid boxed in detection patch label sampler
- MaskPatchLabelSampler now accepts a spacing
- auxiliary method to create yaml from folders with data (contributation by Thijs Gelton)
bug fixes:
- fix for randomness in uniform point sampler
setup file
- version is read from version.py
- requirements are now read from the requirements file (contribution by Thijs Gelton)
- url now points to github repo (contribution by Jakub Kaczmarzyk)
new features:
- image/annotations are now copied to separate folders
- added source configuration
- added one-hot-encoding preset
- dataset and labels are now sorted such that the generator will produce the same output when the seed is set
- minor refactoring
requirements:
- increased NumPy version to make it compatible with ASAP2.0
bug fixes:
- fixed sampling bugs
new features:
- increase speed sampling
bug fixes:
- fixed detection sampling bugs
new features:
-
detection support
-
pyvips image backend support
-
multiple spacings support
-
update call from producers to commander
-
fixed: copy data failed on first trial when using multiple cpus
- No change log or this version
- No change log or this version
- No change log or this version
- No change log or this version
- No change log or this version
- No change log or this version