tidysdm 1.0.4

• improve documentation of calib_class_threshold and implement collect_calib_thresh
• update saved objects to the new version of xgboost

tidysdm 1.0.3

• add mgcv to suggested packages to deal with upcoming changes in ggplot2 4.0.0

tidysdm 1.0.2

• when using predict_raster(), process the raster in chunks if it is too large to fit in memory.
• run tests relying on suggested packages on if those packages are already installed

tidysdm 1.0.0

• add interoperability with stars (thanks to @btupper)
• fixed inverted plots for explain_tidysdm when plotting for all workflows.
• when using explain_tidysdm with a recipe that contains steps, data have to be passed explicitly to the data argument
• implement make_mask_from_presence to define the area of interest
• check time units of rasters for sample_background_time() and sample_pseudo_absences_time() (thanks to @zpmdal)
• show the use of map projections in the overview vignette, and make sure that projections are properly handled in functions such as thin_by_cell and thin_by_dist

tidysdm 0.9.5

• implement clamping and MESS to manage extrapolation
• clearly separate sampling of background vs pseudo-absences
• update vignettes

tidysdm 0.9.4

• fix a but in the predict* functions that prevented a fixed threshold to be used to assign classes
• ensure compatibility with upcoming changes in tune

tidysdm 0.9.3

• fix bug in filter_high_cor due to changes in terra 1.6.75
• implement collect_metrics for ensembles.

tidysdm 0.9.2

• Release on CRAN

tidysdm 0.9.1

• Add a spatial_recipe class. This is a BREAKING change that makes previously saved objects unusable, but old code will work as expected.
• Additional articles showing how to use additional tidymodels features, and how to debug errors.
• Integration of DALEX to explain models.
• More functions to select variables.

tidysdm 0.9.0

• Initial release on GitHub.