dlup 0.9.2

dlup artifacts for 0.9.2. Python wheels published to PyPI (also attached here), with the sdist as the source fallback.

What's Changed

• feat: v0.9.1 by @jonasteuwen in #256
• feat: v0.9.3 by @jonasteuwen in #257
• fix(build): release script improved by @jonasteuwen in #258

Full Changelog: v0.8.0...0.9.2

v0.8.0

Import from internal monorepo, new versions will follow this paradigm.

v0.7.0

In this release the build tools have been modified. If you have issues installing dlup, please create an issue

What's Changed

• Allow annotation files with raster type (yet not supported) in #245
• Add a C++ TIFF writer #249

v0.6.1

What's Changed

• Annotation filter bugfix by @EricMarcus-ai in #246

New Contributors

• @EricMarcus-ai made their first contribution in #246

Full Changelog: v0.6.0...v0.6.1

v0.6.0

This release has performance and testing improvements. Backward incompatibility issues are not to be expected.

What's Changed

• Feature/speedup masking #242
• Increased test coverage in #242

Full Changelog: v0.5.3...v0.6.0

v0.5.3

What's Changed

• Fix annotations not being cropped #241

Full Changelog: v0.5.2...v0.5.3

v0.5.2

What's Changed

• Properly close open file handlers with PYVIPS image backend by @AjeyPaiK in #238

Full Changelog: v0.5.1...v0.5.2

v0.5.1

What's Changed

Two small, but important bugs have been fixed

• Bugfix/multiclass flag by @AjeyPaiK in #234
• Implement proper serialization of the dlup.annotations.Polygon and dlup.annotations.Polygon objects to ensure correct pickling by @AjeyPaiK in #236

Full Changelog: v0.5.0...v0.5.1

v0.5.0

What's Changed

• Support multiclass in convert_annotations (#228)
• Support colors and z_indices in Annotations (#228)
• Several performance improvements and bugfixes

Full Changelog: v0.3.38...v0.5.0

v0.4.0

Update with breaking changes

Next to several bugfixes, a major change is that the SlideImage and the dataset classes now output a pyvips.Image. For a simple fix, you can do the following:

pil_image = PIL.Image.fromarray(np.asarray(sample["image"]))

Note that calling np.asarray() on the image returns the same data as was previously the case with PIL. This changes has led to several performance improvements

Bugfixes

• TiffWriter did not work properly when using a pyramid. This has now been fixed.

Full Changelog: v0.3.38...v0.4.0