v1.0.0

What's Changed

• [DAR-2469][External] Fixed opencv version to 4.10.0.82 by @JBWilkie in #871
• [DAR-2640][External] Lock tenacity to 8.3.0 by @JBWilkie in #869
• Bump urllib3 from 2.2.1 to 2.2.2 by @dependabot in #870
• [DAR-2606][External] Strip leading & trailing whitespaces from class names when reading Darwin JSON annotations by @JBWilkie in #866
• Housekeeping: Version bump to v1.0.0 by @JBWilkie in #873

BREAKING CHANGES

• [DAR-2246][External] Unify darwin-py naming convention by @JBWilkie in #868
• [DAR-2639][External] Add data & tests for in-platform model training by @JBWilkie in #872

Previous versions used a non-unified combination of item_name and source_name when saving and loading files. In an effort to unify our approach to local files when pulling releases and loading files, this version makes the following changes:

1. All exports will be exported in folders mirroring the data structure of the dataset by default (previously a flat structure was the default)
2. The item_name in the JSON export will always be used as a default to save and read files with three notable exceptions:
   • Downloading video files as individual video frames: Each video will result in a folder named after the item_name, containing individual frames named after the frame number in the frame index
   • Downloading multi-slotted items: Multi-slotted items have one item name (for the overall item) but potentially many source_names (for the component files that make up the item). This will result in a folder named after the item_name, containing one file for each slot named after the source name
   • Downloading a slotted item consisting of multiple concatenated DICOM slices: These items can contain multiple source files in a single slot. Each item will result in a folder named after the item_name, containing one file for each slice named after the source name

Full Changelog: v0.8.62...v1.0.0