ggplot2 to be compatible with version 4.0.0 (thanks to @teunbrand for the PR).plotCor()gains the argument text_size to tweak the size of the text in the plotblockCV version 3.x (#24)raster package has been deprecated (#23)SDMtune migrated to package terra to handle spatial data. Objects created with the package raster will not be supported in future releasesSDMtune has a new progress bar and messages/warnings/errors have a new lookmodelReport(), prepareSWD, reduceVariable(), thinData(), and varSel() gain the argument verbose to disable informative messagesgridSearch(), optimizeModel(), randomSearch(), reduceVar(), and varSel() gain the argument interactive to disable the interactive real-time chartdoJk(), gridSearch(), optimizeModel(), predict(), randomSearch(), thinData(), train(), varImp(), and varSel() gain the argument progress to disable the progress barthinData() prints a message with the number of removed locationsfactors in modelReport() is deprecated. See terra documentation to see how to use factors during predictionsBug fixes:
plotCor() plots a correct heatmap when there are more than 9 variablesdoJk() runs also without testing datasetResolve CRAN notes
Bug fixes:
ENMeval >= 2.0Bug fixes: Add the argument factors to the function modelReport() to allow predictions for raster objects including categorical variables
Include Java >=8 in system requirements
get_tunable_args().Main changes:
checkMaxentInstallation() to check if Maxent is correctly configured.get_tunable_args() has been replaced by getTunableArgs() to be consistent with the camel case function naming. At the moment is still possible to use both functions but get_tunable_args() will be deprecated in the next release.Bug fix: plotPA function now works also with new version of ggplot2.
Main changes:
thinData accept now a matrix or a dataframe with several columns, useful if the users has information related to the coordinates that doesn't want to lose with the thinning procedure.plotResponse plots the response for the full range of presences and backgrounds/absences when only_presence is TRUE, only_presence is unused only to compute the provided function to the range of presence locations when marginal = TRUEBug fix:
Bug fixes:
Main changes:
addSamplesToBg to add presence locations to background location.extra_args in Maxent models cannot be changed anymore.Bug fix: Fix CRAN error for r-oldrel
Main changes:
SDMmodelCV objectstrain functionSDMmodelCV objectmergeSWD functionBug fixes:
plotPA function now works also with large raster objectsThis release is the first step to enable more methods to train models. The main change is that the SDW object now bundles together the presence and the absence/background locations and is not necessary anymore to pass the presence and absence locations as separate arguments to the train function.
Old objects of class SWD, SDMmodel, SDMmodelCV and SDMtune created with version <= 0.1.1 must be converted into the new format using the dedicated help functions.
Main changes:
SWD object bundles together the presence and absence/background locationstrain accepts now folds partition generated with other packages (i.e. ENMeval and blockCV)plot function for SDMtune objectsvirtualSp with presence, absence and background locationsDeprecated function: getSubsample
VarImp function for SDMmodelCV objectsmaxentVarImp function available for SDMmodelCV objects