pyLMD v2.0.0

[lucas-diedrich]

This work-in-progress PR aggregates all changes planned for py-LMD v2.0.0.

Goals of v2.0.0

The main objectives of this release are to:

• Modularize the codebase.
• Expose useful internal innovations (e.g. optimal path sorting of shapes).
• Increase test coverage and reliability.

In the longer term, these changes should make it easier to:

• Support additional XML formats used by the Leica LMD7.
• Reuse individual methods in other packages and contexts.

All changes are intended to preserve the current public API, so that downstream packages require minimal (ideally no) modifications.

Planned Steps

• Preserve py-LMD v1.5.0 behavior by increasing test coverage to >85% and adding tests for all public methods.
• Expose optimal path sorting (TPS algorithms) as standalone utilities (see make poly_compression and path optimisation for shape collections not generated by the segmentation loader #9).
• Refactor monolithic modules into smaller, more focused modules while maintaining the high-level API.
• (Optional) Refactor large classes (Shape, Collection, SegmentationLoader) into smaller components and extract reusable logic into helper classes or functions.