Package: ggRandomForests Version: 2.7.3
plot.gg_partial(),
plot.gg_partial_rfsrc(), and
plot.gg_partialpro() now always return a single
ggplot/patchwork object. Previously, when both
continuous and categorical predictors were present, they returned a
named list list(continuous=, categorical=), which surprised
users and made autoplot() dispatch ambiguous. The two
panels are now combined vertically via
patchwork::wrap_plots() (patchwork moved from
Suggests to Imports). Closes #77.autoplot() S3 methods for all 10 gg_*
classes, delegating to the corresponding plot.gg_*() method
so objects work in |> pipelines, patchwork,
and cowplot compositions via
ggplot2::autoplot().print() and summary() S3 methods for every
gg_* data object (gg_error, gg_vimp, gg_rfsrc, gg_variable,
gg_partial, gg_partial_rfsrc, gg_partialpro, gg_roc, gg_survival,
gg_brier). print() is header-only — use head()
for rows. summary() returns a printable
summary.gg object with per-class diagnostics. Each
gg_* constructor now attaches a "provenance"
attribute (source, family, ntree, n, xvar.names) consumed by the new
methods.gg_brier() extractor and
plot.gg_brier() method for time-resolved Brier scores and
CRPS on survival forests (issue #9). Wraps
randomForestSRC::get.brier.survival() and adds the
mortality-quartile decomposition, a 15-85 percent per-subject envelope,
and running CRPS via trapezoidal integration. Supports
cens.model = c("km", "rfsrc"),
type = c("brier", "crps"), and envelope
(overall line + 15-85% ribbon). Multi-model comparison is left to
dplyr::bind_rows() on multiple gg_brier
outputs — see ?gg_brier for an example..gg_ribbon_alpha = 0.2) and
a shared fill (.gg_ribbon_fill = "steelblue") for
single-series cases (KM/NA CIs, bootstrap CIs, gg_brier
envelope); group-stratified ribbons keep their group-coloured fill.
Statistical bounds unchanged — only styling. ggRandomForests v2.7.2
=====================DESCRIPTION (Breiman 2001 and
Ishwaran et al. 2008, with <doi:...> auto-links) per
CRAN cookbook.man/shift.Rd Rd file: shift() is
an internal utility and the example used
ggRandomForests:::shift(...). Marked the function
@noRd so it no longer generates a help page.cat() in surv_partial.rfsrc() with
message() so progress output is suppressible
(suppressMessages()) and plays nicely inside notebooks /
Shiny / quarto.par() settings in the
surv_partial.rfsrc() example via
oldpar <- par(no.readonly = TRUE); on.exit(par(oldpar)).gg_partial_rfsrc() for survival forests:
partial.rfsrc() was being called without
partial.type, causing a zero-length comparison
(if (partial.type == "rel.freq") ...) inside the C-level
prediction routine and aborting the call. Survival forests now pass
partial.type = "surv" (default; configurable via the new
partial.type argument accepting "surv",
"chf", or "mort"). This unblocks the
partial-dep chunk in the survival vignette.gg_partial_rfsrc() for survival forests with
multiple partial.time values:
get.partial.plot.data() returns yhat as an
[length(partial.values) x length(partial.time)] matrix, but
the previous code assumed a vector and crashed on column-mismatch when
assigning time. The result is now reshaped to long form so
each (x, time) pair is a single row.plot.gg_partial_rfsrc() survival layout:
predictor value is now on the x-axis with one curve per (rounded) time
point coloured by Time, faceted by variable name. The
previous default put time on the x-axis and one curve per predictor
value, producing a saturated legend with dozens of nearly-identical
lines.tests/testthat/test_plot_layer_data.R: regression
suite that uses ggplot2::layer_data() to verify each
plot.gg_*() method renders non-empty layers for every
supported forest family. Catches the empty-figure class of bug
(transform/plot column-name mismatch) without requiring visual
inspection.ggrandomforests.news() now reads NEWS.md
(the canonical change log R also surfaces via
utils::news()). The legacy hand-maintained
inst/NEWS has been removed — it had silently drifted to
v2.4.0 (June 2025) across three releases, so users running the helper
saw stale version info. One source of truth, no more drift window.plot.gg_vimp() legend duplication: the bar geom
mapped both fill and color to the
positive column, but only the fill legend was titled “VIMP
> 0”, leaving a redundant second legend titled “positive”. Both
aesthetics now share the “VIMP > 0” title so ggplot merges them into
a single legend by default.plot.gg_vimp() for forests with all-positive VIMP:
the bar geom previously mapped only color (no
fill), producing hollow / outline- only bars and an
“Ignoring unknown labels: fill” warning whenever
labs(fill = ...) was applied. Both fill and
color are now mapped unconditionally, so bars render filled
in every case.@examples blocks to
plot.gg_partial_rfsrc() and
plot.gg_partialpro(). The latter uses a self-contained mock
of the varpro::partialpro() output structure so the example
runs without pulling in varpro as a dependency.gg_partial(),
gg_partialpro(), and gg_partial_rfsrc() now
stamp their return values with S3 classes (gg_partial,
gg_partialpro, gg_partial_rfsrc respectively),
enabling plot() dispatch without any boilerplate.plot.gg_partial(),
plot.gg_partial_rfsrc(), and
plot.gg_partialpro() S3 methods; continuous predictors
render as line plots, categorical as bar charts, faceted by variable
name. Survival forests produce curves over time; two-variable surface
plots group by xvar2.name.gg_survival() to an S3 generic dispatching on
the class of its first argument. New gg_survival.rfsrc()
method extracts the survival response directly from the fitted forest
(no separate data argument needed); gg_survival.default()
preserves the existing interface.plot.gg_survival() auto-coercion: previously called
gg_survival(rfsrc_obj) treating the forest as the
interval string argument, causing a latent crash; replaced
with inherits() guard.surv_partial.rfsrc() via
.Deprecated() with a pointer to
gg_partial_rfsrc(); all package tests updated to suppress
the warning.gg_partial_rfsrc() — make_eval_grid()
used unlist(dplyr::select()) which coerced factor columns
to integer codes; now uses newx[[xname]] to preserve column
class. Categorical detection extended to cover is.factor()
and is.character() in addition to the cardinality
check.gg_partial_rfsrc(): all-NA
xval after NA removal now emits a warning and skips the
variable; all-NA grouping variable (xvar2) calls
stop(); n_eval and cat_limit are
validated as single integers >= 2 near function entry.gg_partial_rfsrc.R:
refactored into eight top-level unexported helpers
(validate_scalar_int, validate_partial_args,
snap_partial_time, make_eval_grid,
call_partial_rfsrc, partial_one_var,
partial_no_group, partial_with_group,
split_partial_result); all functions now score below the
cyclocomp_linter limit of 20.@param partial.time documentation: “see the section
above” corrected to “see the section below”.tidyr::gather() with
tidyr::pivot_longer() in plot.gg_vimp() and
plot.gg_partialpro().gg_survival.rfsrc,
gg_survival.default, plot.gg_partial,
plot.gg_partial_rfsrc, and plot.gg_partialpro
to NAMESPACE; add corresponding @rdname /
@export roxygen tags.expect_s3_class() checks for all new
classes; add plot() smoke tests for
gg_partial, gg_partial_rfsrc,
gg_partialpro; add gg_survival.rfsrc tests for
KM extraction, by stratification, and error on non-survival
forest.plot.gg_partial,
plot.gg_partial_rfsrc, and plot.gg_partialpro
to _pkgdown.yml reference index.plot.gg_rfsrc: all
aes() calls used bare string literals instead of
.data[[col]], causing every aesthetic to map to a constant
string rather than the underlying data column. All plot types
(regression, classification, survival) were affected.aes() bare-string literals in
plot.gg_roc multi-class branch; remove unreachable
if (crv < 2) dead-code branch.bootstrap_survival CI-band indexing in
gg_rfsrc: negative index computed via
colnames() was a no-op on large datasets and a latent crash
for data with ≤ 2 unique event times.gg_rfsrc.rfsrc: is.null(df[, col])
does not detect missing columns; replaced with
!col %in% colnames() guard.gg_rfsrc.randomForest: method used non-existent
object$xvar; now recovers the training frame via
.rf_recover_model_frame().plot.gg_error for
single-outcome forests where the data frame has no variable
column.gg_vimp and plot.gg_vimp:
1:nvar replaced with seq_len(nvar) in both S3
methods; 1:0 silently returned c(1, 0) instead
of integer(0) when nvar == 0.expect_is
→ expect_s3_class / expect_type /
expect_true(is.*()); expect_equivalent →
expect_equal(ignore_attr = TRUE); all
context() calls removed; testthat 1.x
expect_that / is_identical_to removed..lintr package-level linter configuration; fix
lintr spacing in gg_partial.lint.yaml now fails CI on any
lint issue; R-CMD-check.yaml treats warnings as errors and
uses Rtools 44; test-coverage.yaml duplicate codecov upload
removed.covr and vdiffr to
Suggests.gg_partial for
categorical variable datagg_partial and gg_partial_rfsrc
to improve factor-level normalisation and categorical data handlinghvtiRutilities internal dependency; clean up
associated importsgg_partial_rfsrc to use .data
pronoun for all dplyr callsgg_partial_rfsrc function: computes partial
dependence data directly from an rfsrc model via
randomForestSRC::partial.rfsrc, without requiring a
separate plot.variable callxvar2.name) in
gg_partial_rfsrcTo pull this out of archive on randomForestSRC 3.1 build release. Fixed a plot bug for gg_error to show the actual curve (issue 35)
combine.gg_partial bug when giving a single variable
plot.variable object.
Remove dplyr depends to transitions from “Imports”
to “Suggests”.
Argument for single outcome gg_vimp plot for
classification forests.
Improvements to gg_vimp arguments for
consistency.
Add bootstrap confidence intervals to gg_rfsrc
function.
Initial partial.rfsrc function to replace the
randomForestSRC::plot.variable function.
Move cache data to randomForestSRC v1.6.1 to take
advantage of rfsrc version checking between function
calls.
Vignette updates for JSS submission of “ggRandomForests: Exploring Random Forest Survival”.
Vignette updates for arXiv submission of ggRandomForests: Random Forests for Regression
Some optimizations to reduce package size.
Remove all tests from CRAN build to optimise R CMD CHECK times.
Remove pdf vignette figure from CRAN build.
Return S3method calls to NAMESPACE for “S3 methods exported but not registered” for R V3.2+.
Misc Bug Fixes.
gg_ functions into S3 methods to allow future
implementation for other random forest packages.Maintenance release, mostly to fix gg_survival and gg_partial plots. * Fix the gg_survival functions to plot kaplan-meier estimates. * Fix the gg_partial functions for categorical variables. * Add some more S3 print functions. * Try to make gg_functions more consistent. * Further development of package vignette “Survival with Random Forests” * Modify the example cached datasets for better estimates and examples. * Improve help files. * Misc Bug Fixes.
Need a high-speed mirror for your open-source project?
Contact our mirror admin team at info@clientvps.com.
This archive is provided as a free public service to the community.
Proudly supported by infrastructure from VPSPulse , RxServers , BuyNumber , UnitVPS , OffshoreName and secure payment technology by ArionPay.