Changelog
Source:NEWS.md
BiometryTraining 1.0.0
Final release of BiometryTraining
This package will no longer be developed. All future development work will be undertaken in the biometryassist package.
Major changes
- Created a variogram function that does not rely on ASreml (#33)
- Implemented
autoplot()
function fordesign()
andmct.out()
. (#60) - Coloured plots (
autoplot.design()
andvariogram()
) now gain apalette
argument to enable colour blind friendly colours in the plots if desired (as well as some other palettes). (#85)
Bug fixes
- More sanity checking of inputs in
design()
(#69) - Various bug fixes in
logl.test()
(#31 and #51)
BiometryTraining 0.9.0
Minor changes
- Fixed some bugs in
logl.test()
. - Unknown arguments in
design()
are no longer silently ignored (#77) - Enabled a way for split plot designs to be arranged row-wise or column-wise (#68)
- Added split plot treatment labels to the output plot (#78)
- Prevented misspelled arguments passed in the ellipsis from being silently ignored (#77)
BiometryTraining 0.8.0
Bug fixes
- Fixed a bug where factorial latin square designs were not generated properly (#62)
- Updated labels on design outputs so that they are consistent and prevent 3 way factorials (#65)
- Fixed up labelling in split plot designs (#64)
- Fixed an issue where incorrect significance values in
mct.out()
give a negative confidence interval (#73) - Fixed a bug where crossed designs didn’t produce spreadsheet output (#66)
BiometryTraining 0.7.0
- Added a
NEWS.md
file to track changes to the package.
Major changes
- Introduced the
design()
function to create designs and produce the output in one step. (#40) -
resplt()
now automatically produces multiple residual plots in situations like multisite analyses (#53) - Updated the look of
resplt()
and included output of the Shapiro-Wilk’s test for normality. - Bumped required minimum version of R to 3.6.0.
Bug fixes
- Fixed a bug where aliased treatment levels would cause
mct.out()
to crash (#47) - Fixed a bug where
logltest()
could infinitely cycle with models that don’t converge (#50) - Fixed a bug where
logltest()
would print all the terms on the boundary. - Reduced the amount of warnings and messages printed by
logltest()
(#17)