Tools to support sensible statistics for functional response analysis.
cat(). This is the right thing to do.
bdII*family) in versions prior to 0.5 was not correctly specified.
Pwhich only "partitions" the fitted coefficients between multiple predators.
emdII. Hopefully this name better reflects its heritage.
real77*family was internally consistent, but not terribly well specified. It is probable that the old specification made curves difficult to fit and generated non-sensical 'q' parameters.
real77*family has now been fully deprecated and removed from FRAIR. Trying to use these functions will throw an error.
flexpnr) have been added as a replacement.
hassIIInr, which better reflects that it is a Non-Replacement model.
hassIIIr*function directly will pass through to
hassIIInr, with a warning.
hassIIIris no longer recognised by
frair_responses(), so attempting to fit this model in the FRAIR framework will throw an error.
bythotrephesdataset. It describes Bythotrephes spp. (water fleas) preying on prey items of different sizes.
print.frbootremoving an unnecessary extra line break. Yep. Big stuff!
drawpoly.frbootare TRUE, these functions try to plot to zero. If that results in an undefined value for the fitted response, then it draws to 1e-04 instead and a warning is thrown. Note this does not guarantee all bootstrapped fits will be defined at this value either.
getcalls are scoped to 'package:frair'. This means it's no longer possible (whether accidentally or deliberately) to 'overwrite' FRAIR functions with those in the global namespace. This "functionality"" might return in the future alongside some formal documentation for extending FRAIR.
bootdoes not use
::. The reason is that in the (fairly likely) scenario that users go digging into the structure of the frboot objects and call
boot::bootin the call confuses
print.bootand it reports some pretty crazy stuff!