check()that was preventing use of the
initparameter for samc objects created directly from a matrix.
NAvalues in landscape data.
samc-classobjects from raster/map data; does not work with manually constructed transition matrices.
origininput. It does not support the
origininput for CRW is a matrix with the cell number and a direction
map()function averages the results for all directions. This could change in the future to allow for more flexibility.
samc-classobject creation by eliminating overhead associated with user-defined functions. Currently, only
"1/mean(x)"is supported, but others can be added in the future.
rasterize()function publicly available. Mainly useful for converting matrices to a SpatRaster that matches the structure used internally by the package.
visitation()function is feasible for samc objects with 50 Million cells with 32 GB of RAM. Details about changing the solver can be found in the help documentation for the
samc()function so that gdistance can be removed as a dependency.
symoption for creating the samc object is currently ignored.
map()function was updated so that the output matches the input types used in the
modelto reflect future anticipated support for different types of models. Current usage will not change and assumes a default random-walk model.
occparameter in metrics to
init(short for “initial state” or “initialize”)
cond_passage()function to match other metrics, but it is not currently used.
samc()function parameters. This is a breaking change that will make maintaining the package and adding new features a simpler process going forward, and that will hopefully only be a minor inconvenience for users. The warning message on package load introduced in v1.4.0 has been updated to reflect the new changes.
0for when i==j in the vectors. This fixes an issue associated with shifted indices in
cond_passage(samc, dest). It also technically breaks backward compatibility for when
cond_passage(samc, origin, dest). Previously,
cond_passage(samc, origin, dest)would return
dest, but this decision was arbitrary. The
cond_passage()documentation explains why.
dispersal(samc, origin/occ)function via the RcppThread package. See the Parallel Computing vignette for details.
samc(). CRS objects have a hidden field that can vary depending on system and software versions, and previous versions of the check would not account for this. This would to lead to false positives where perfectly compatible rasters were reported as incompatible. The corresponding error message was also fixed to report the correct issue; the code was initially copied and modified from another input check, but the error message wasn’t updated in the process.
absorptionparameter in samc() is treated as the total absorption (consistent with previous behavior). After the creation of the samc-class object, additional absorbing states can be attached to the samc-class object. See the samc-class documentation and the new Multiple Absorption tutorial for more details.