First FARO release

v1.0.0

First release of FARO

v0.2.0

Attention: Breaking Changes

This release introduces a new segmentation abstraction, significant refactoring of the image processing pipeline, and multiple performance and robustness improvements.

• Introduced a new SegmentationMethod class to encapsulate segmentation metadata, including method name, implementation class, and channel usage.
• Refactored ImageProcessingPipeline to robustly handle scenarios where acquisition outpaces image processing. The pipeline now uses a three-queue architecture: a storage queue, an in-memory analysis queue, and a deferred analysis queue that reloads images from disk as needed.
• Refactored feature extraction logic in FE_ErkKtr and added a new FE_ErkKtrReduced class for optimized, lower-overhead feature extraction.
• Improved error handling in SegmentatorImagingServerKit by adding automatic retry logic for connection failures.
• Enhanced error handling for Ti microscopes to better recover from hardware faults (e.g., Mönch + Jungfrau microscope).
• Extended stimulation classes with a use_labels attribute, enabling stimulation strategies that do not depend on current segmentation masks and reducing processing overhead.
• Performed general code cleanup and reorganization across multiple modules to improve readability, maintainability, and long-term extensibility.