Added methods for nestedmodels::nested()
(@ashbythorpe, #256).
Updated methods for mgcv::gam()
to also remove the hat
and offset
components (@rdavis120, #255).
Clarified the messaging for butchering results, as well as when butchering may not work for survival::coxph()
(#261).
Fixed a bug in butchering BART models (#263).
Added butcher methods for mixOmics::pls()
, mixOmics::spls()
, and mixOmics::plsda()
(#249).
Added butcher methods for klaR::rda()
and klaR::NaiveBayes()
(#246).
Added butcher methods for ipred::bagging()
(#245).
Added butcher methods for MASS::lda()
and MASS::qda()
(#244).
Added butcher methods for survival::coxph()
(#243).
Added butcher methods for xrf::xrf()
(#242).
Added butcher methods for mda::fda()
(#241).
Added butcher methods for dbarts::bart()
(#240).
Added butcher methods for ClusterR::KMeans_rcpp()
(#236).
Added butcher methods for clustMixType::kproto()
(@galen-ft, #235).
Julia Silge is now the maintainer (#230).
Updated printing for memory_released()
(#229).
Added butcher methods for mgcv::gam()
(#228).
Added an axe_fitted()
method to butcher the template
slot for prepped recipes (@AshesITR, #207).
Added butcher methods for glm()
(#212).
Removed axe_fitted()
and axe_ctrl()
for xgboost, because these methods caused problems for prediction (#218).
Moved usethis and fs to Suggests (#222).
Removed fastICA and NMF from Suggests. fastICA requires R >= 4.0.0 now, and NMF is often hard to install and was only used for one test (#201).
Preemptively fixed a test related to a recipes change in step_hyperbolic()
(#220).
Transitioned unit tests to make use of modeldata::Sacramento
rather than modeldata::okc
in anticipation of okc
’s deprecation in an upcoming release of modeldata (@simonpcouch, #219).
axe_env()
method to remove the terms
environment for recipe steps. This covers most recipe steps, but certain steps still need more specific methods (@juliasilge, #193).Ensure butcher is compatible with recipes 0.1.16, where a few steps have been renamed.
Fixed issue with survival 3.2-10, where butcher was using frailty terms incorrectly (#184).
Fixed an issue where axing a parsnip ‘model_fit’ would return the underlying model object rather than the altered ‘model_fit’.
Fixed a few test failures related to changes in parsnip (#157).
C50
objects were updated to reflect data files were now located in modeldata
.modeldata
was added as a dependency since the data files required for testing axe methods on models objects instantiated for testing were moved into this library.glmnet
was removed as a dependency since the new version depends on 3.6.0 or greater. Keeping it would constrain butcher
to that same requirement. All glmnet
tests are run locally.NEWS.md
file to track changes to the package.