OptiWindNet v0.2.2 This release focuses on visualization, heuristic consolidation, LKH parity, and routing robustness. High-level plotting now defaults to the interactive vector SVG backend for cleaner notebook output, while Matplotlib remains available through explicit `ax` handling. SVG plotting also gains feature parity with `gplot()` for labels, boundary and obstacle tags, and legends. Routing improvements include the new unified `heuristics.constructor()` path, expanded `EWRouter` capabilities including radial and rootlust methods, and a feature-complete `lkh3()` entry point with multi-root support, clustering, warm starts, retries, crossing repair, and richer metadata. PathFinder detours are substantially more robust around boundaries and exclusion zones. The release also adds Shapely-based geometric crossing diagnostics, more informative `repr()` output for high-level API objects, shorter substation labels in packaged datasets, Python 3.11-3.14 support metadata, stricter deprecation testing, and several performance and maintenance improvements. Breaking change: root-assignment and link-blockage helper functions moved from `optiwindnet.geometric` to `optiwindnet.interarraylib`. Import `add_terminal_closest_root()`, `add_link_blockmap()`, and `add_link_cosines()` from `optiwindnet.interarraylib`.