Changelog • supercells

supercells 1.9.5

Fixed integer overflow in vector indexing inside the C++ part of the code: C++ code now uses 64-bit arithmetic for array indices (std::size_t) instead of 32-bit integers when accessing raster data. This prevents silent data corruption or crashes when processing large rasters (>100M cells with multiple bands). The fix should have no performance impact. (Thanks to Lucas De Oto)

Added outcomes argument to sc_slic(), sc_slic_points(), and sc_slic_raster(); replaces metadata for controlling returned fields
Iteration diagnostics API redesigned: iter_diagnostics and sc_plot_iter_diagnostics() replaced by sc_slic_convergence() with a plot() method
Added sc_slic_get_params() and sc_slic_set_params() for reading/writing stored sc_slic() parameters
Added use_meters() for map-distance step values (replacing in_meters())
Added use_adaptive() for adaptive compactness mode (replacing compactness = "auto")
Updated metrics API (sc_metrics_pixels(), sc_metrics_supercells(), sc_metrics_global()) to better reuse sc_slic() metadata and improve scaling/compactness handling
Updated sc_tune_compactness() to align with the revised compactness/step workflows
Documentation and vignettes updated (pkgdown refresh, new articles, and revised examples)

compactness = "auto" enables ASLIC-style adaptive compactness
Metrics now support ASLIC-like adaptive compactness (pixels/supercells/global)
sc_tune_compactness() returns only step + one compactness, selected by metrics = "global" or "local"
Chunked raster IDs now use a consistent center-count offset strategy

Added sc_tune_compactness() to estimate compactness from a short SLIC run
Metrics API cleanup: renamed sc_metrics_clusters() to sc_metrics_supercells(), unified metric names, _scaled suffix for scaled outputs
Balance metrics now use absolute log ratio of scaled value/spatial distances
iter = 0 only works for point outputs; polygons/rasters error with guidance
sc_slic()/sc_slic_points()/sc_slic_raster() dropped the transform argument; legacy supercells() keeps transform = "to_LAB"
New step_unit supports map-unit step sizes
Chunking improvements: conservative memory estimates, options(supercells.chunk_mem_gb), chunk sizes align to step, deterministic ID offsets with file-backed merge
Verbose argument moved to the end in R and C++ APIs
Empty-cluster handling now consistent across cleanup
Removed future-based parallel chunking

Spatial distance uses precise (fractional) center positions; neighborhood search still uses rounded centers (minor output differences are expected vs 1.5)
More robust center handling (custom centers, empty centers) and safer cleanup for small regions
Local-minimum search improved
Tests standardized via tests/testthat/setup.R and updated expectations

Major C++ SLIC refactor (clearer data flow, diagnostics support) with supporting utility cleanup
Standardized sc_slic(), sc_slic_points(), sc_slic_raster() with consistent options/metadata/chunking; sc_slic() is now the main entry point
Added diagnostics API (sc_metrics_pixels(), sc_metrics_clusters(), sc_metrics_global()) and iter_diagnostics + sc_plot_iter_diagnostics()
Centralized internal helpers for validation, chunking, and normalization
Expanded regression and metrics tests

CRAN release: 2024-02-11