Releases: tidymodels/broom
broom 1.0.7
-
Corrected
nobs
entries inglance.coxph()
output: the package used to
returnlength(object$linear.predictors)
(equal ton
rather thannevent
)
and now uses survival's method (#1224). -
Corrected confidence interval values in
tidy.boot()
and addressed errors
when bootstrapping confidence intervals for multiple terms (#1212). -
Reverted deprecation of tidiers for objects from the margins package
now that the package is back on CRAN (#1220). -
Addressed failure in
tidy.anova()
ahead of upcoming car
release (#1215). -
Clarified documentation for a number of cases where dots were
documented as unused but actually passed to internal functions
(#1214). -
Addressed error in
augment.betareg()
andaugment.betamfx()
withdata = NULL
and non-nullnewdata
(#1216, #1218). -
glance.lm()
now returns non-NA
values forstatistic
,p.value
, anddf
for models fitted with a single predictor and no intercept (@jrob95, #1209).
broom 1.0.6
New Features
-
Added support for
conf.level
inaugment.lm()
(#1191 by@zietzm
). -
Added support for columns
adj.r.squared
andnpar
inglance()
method for objects outputted frommgcv::gam()
(#1172).
Deprecations
-
Soft-deprecated tidiers for margins objects, as the package was archived from CRAN in April 2024. In the case that the package is back on CRAN before the next package release, broom will once again Suggest and test support for the package (#1200).
-
Moved forward with deprecation of tidiers for objects from the sp package. See resources linked in tidymodels/broom#1142 for more information on migration from retiring spatial packages.
Bug Fixes
-
While this broom release contains no changes to the
tidy.survfit()
method for objects from the survival package, the package has bumped the minimum required version for survival. Before survival 3.6-4,tidy.survfit()
propagated "inconsistent"n.censor
values from survival for multi-state models (#1195). -
Corrected confidence interval values for precision components in
tidy.betareg()
output (#1169). -
Fixed bug in tidier for
car::linearHypothesis()
output with long formulas (#1171). -
Corrected coefficient values in
tidy.varest()
output (#1174).
broom 1.0.5
-
tidy.coxph()
will now pass its ellipses...
tosummary()
internally (#1151 by@ste-tuf
). -
Transitioned the deprecation of the
region
argument totidy.SpatialPolygonsDataFrame
from a warn- to a hard-deprecation (#1142). -
Removed maptools and rgeos as Suggested packages ahead of their retirement. sp tidiers will be removed from a future release of the package (#1142).
-
Addressed bug in mlogit tidiers where
augment.mlogit()
would fail if supplied a model fitted with a non-defaultdfidx()
(#1156 by@gregmacfarlane
). -
Addressed bug in ANOVA tidiers where
tidy.anova()
would fail if passed a model with many predictors (#1159 by@jwilliman
). -
Addressed warnings in ANOVA tidiers for unrecognized column names
Resid..Df
,Resid..Dev
, andDeviance
; those columns will be renameddf.residual
,residual.deviance
, anddeviance
, respectively (#1159 by@jwilliman
).
broom 1.0.4
- Added an
intercept
argument totidy.aov()
, a logical indicating whether to include information on the intercept as the first row of results (#1144 by@victor-vscn
). - Moved forward with soft-deprecation of tidiers for objects from the sp package ahead of the retirement of the rgeos and maptools packages later this year. sp tidiers will be removed from a future release of the package (#1142).
- Fixed bug in
augment.glm()
where the.std.resid
column always contained standardized deviance residuals regardless of the value passed to thetype.residuals
argument (#1147).
broom 1.0.3
- Addressed test failures on R-devel.
- Fixed bug in
tidy.multinom()
where theconf.level
argument would be ignored.
broom 1.0.2
- Fixed a bug where
augment()
results would not include residuals when the response term included a function call (#1121, #946, #937, #124). - The default
data
argument foraugment.coxph()
andaugment.survreg()
has been transitioned fromNULL
tomodel.frame(x)
(#1126 by@capnrefsmmat
). - Migrated 'ggplot2' from strong to weak dependency, i.e. moved from
Imports
toSuggests
.
broom 1.0.1
- Improves performance of
tidy.lm()
andtidy.glm()
for full-rank fits (#1112 by@capnrefsmmat
). In some contexts, execution time is halved. With broom 1.0.0:
library(broom)
system.time(
replicate(1000, tidy(lm(dist ~ speed, data = cars)))
)
#> user system elapsed
#> 1.385 0.010 1.396
With this release:
system.time(
replicate(1000, tidy(lm(dist ~ speed, data = cars)))
)
#> user system elapsed
#> 0.626 0.011 0.638
Created on 2022-08-30 by the reprex package (v2.0.1)
- Moves forward with deprecation of tidiers for sparse matrices outputted from the Matrix package, initially soft-deprecated in broom 0.5.0. The Matrix tidiers were light wrappers around coercion methods that will now be deprecated from Matrix itself in the upcoming 1.4-2 release. The affected methods are
tidy.sparseMatrix()
,tidy.dgCMatrix()
, andtidy.dgTMatrix()
. Note thattidy.confusionMatrix()
, for relevant objects outputted from the caret package, is unaffected (#1113). tidy.anova()
works again withanova
objects from thelme4
package (broken by addition of theterms
column in the previous release)
broom 1.0.0
broom 1.0.0 is the first "production" release of the package, and includes a number of notable changes to both functionality and governance.
As of this release, the broom team will be following a set of guidelines that clarify the scope of further development on the package. Given the package's wide use and long history, these guidelines prioritize backward compatibility over internal consistency and completeness. You can read those guidelines here!
We've also made notable changes to error handling in this release:
- Adds minimal ellipsis checking to warn on commonly misspecified arguments passed through ellipses. Notably:
tidy()
methods will now warn when supplied anexponentiate
argument if it will be ignored.augment()
methods will now warn when supplied anewdata
argument if it will be ignored.
- The warning regarding tidiers only maintained via dispatch to
lm
andglm
is now displayed only once per session, per unique dispatch. That is,
if aclass_a
object is tidied using a(g)lm
method, broom will not
warn when tidyingclass_a
objects for the rest of the session, but if a
class_b
object is tidied using a(g)lm
method in the same session, broom
will warn again (#1101).
Other fixes and improvements:
- Add
exponentiate
argument totidy.boot()
(#1039). - Update in
tidy.htest()
converting matrix-columns to vector-columns (#1081). - Address failures in
tidy.glht()
withconf.int = TRUE
(#1103). - Address failures in
tidy.zoo()
when input data does not havecolnames
(#1080). - Transition tidiers for bivariate linear or spline-based interpolation---using
list tidiers to interface with objects from the akima package is now
considered off-label. See the interp package for a FOSS alternative. - Address failures in
tidy.svyolr()
whenp.values = TRUE
. Instead of aliasing
tidy.polr()
directly,tidy.svyolr()
lightly wraps that method and
warns ifp.values
is supplied (#1107). - Adds a
term
column and introduces support forcar::lht()
output in
tidy.anova()
(#1106 by@grantmcdermott
). - Adds a dedicated
glance.anova
method (which previously dispatched to the
deprecatedglance.data.frame()
tidier, #1106 by@grantmcdermott
).
As always, thanks to the broom community for their contributions and feedback! 🧹
broom 0.8.0
This update makes significant improvements to documentation, fixes a number of bugs, and brings the development flow of the package up to date with other packages in the tidymodels.
In the big picture, this release:
- Makes many improvements to documentation:
- All tidiers now have example code demonstrating usage in their documentation. Tidiers for base packages as well as selected others also include sample code for visualization of results with ggplot2.
- Code examples in the documentation largely now follow consistent style---these changes were made largely to reflect the tidyverse style guide, addressing spacing, object naming, and commenting, among other things.
- Examples previously marked with
\dontrun
or\donttest
have been workshopped to run reliably.
- Clarifies errors and warnings for deprecated and un-maintained tidiers.
- Ensures that tidiers are placed in files named according to the model-supplying package rather than the model object class for easier navigability of the source code.
Bug fixes and other improvements
- Fix
glance.fixest
error when model includes only fixed effects and no regressors (#1018
by@arcruz0
,#1088
by@vincentarelbundock
). - Address excessive messaging from
tidy.speedlm
(#1084
by@cgoo4
,#1087
by@vincentarelbundock
). - Add
nobs
column to the output ofglance.svyglm
(#1085
by@fschaffner
,#1086
by@vincentarelbundock
). - Ensure
tidy.prcomp
description entries use consistent punctuation (#1072
by@PursuitOfDataScience
). - Address breaking changes in
glance.fixest
andtidy.btergm
. - Simplify handling of
MASS::polr
output in the correspondingtidy
andaugment
methods. - Align continuous integration with current standards in tidymodels packages.
broom 0.7.11
- Addressed issue with the ordering of original observations in
augment.rqs
. Now function preserves the originaldata.frame
names also when the inputdata.frame
only has one column (#1052
by@ilapros
). - Addressed warning from
tidy.rma
whenx$ddf
has length greater than 1 (#1064
by@wviechtb
). - Fix errors in
glance.lavaan
in anticipation of upcomingtidyr
release (#1067
by@DavisVaughan
). - Corrected the confidence interval in
tidy.crr()
. Thetidy.crr(conf.level=)
argument was previously ignored (#1068
by@ddsjoberg
).