Version 0.0.7

New calculation of downsample factors
Support for 32-bit floating point images
8-bit indexed color images no longer fail, but treated as regular 8-bit (necessary for some .lif files)
Revised parsing of z/t values
New dumpMetadata() method to access OME-XML to help with debugging
New getChannelName(channel) method
Introduced DummyMetadata() when creating secondary readers for multithreading
Removed autoscaling with use of non-RGB getBufferedThumbnail
Improved parsing of objective magnification
Added new code to test QuPath's use of Bio-Formats
Added Bio-Formats version number under 'Help → Installed extensions'
An error is shown when starting QuPath is 'bioformats_package.jar' is missing

Version 0.0.6

Switch to Gradle
Added workaround for VSI reading bug that meant channels & z-slices were sometimes confused (available in preferences, turned off by default)
Added fix where 4-channel 8-bit images could be converted to RGB, losing the 4th channel information for some commands
Added fix where some formats (e.g..ims) would have 8-bit RGB data, but this was treated as multichannel; now such cases are treated as packed (A)RGB, giving better performance & more consistent behavior

This was a major revision incorporating many bug-fixes and performance improvements, including:

Better support for a range of image types, especially CZI, NDPIS and VSI
Bio-Formats' memoization can now be used to improve performance when (re)opening images
Improved parallelization when reading image tiles - should be faster, require less memory, and turn itself off when it would do more harm than good
A new set of Bio-Formats options have been added to the QuPath preference panel, including the ability to always use/skip Bio-Formats for specific file formats

Fixed bugs that prevented reading non-RGB whole slide images
Added nearest-neighbor interpolation method to fix accidental intensity rescaling if resized image regions were requested