ggpmisc

to wrong test outcomes in some edge cases. Detect formulas containing spline

base functions bs, ns or lspline as not being regular polynomials.

• Add stat_spikes() with similar functionality as stat_peaks() and

stat_valleys() but targeting very narrow peaks and valleys with unusually

abrupt transitions from the baseline.

• Support flipping of orientation and guess default orientation automatically

from the aesthetic mapping in stat_multcomp() (**breaking** returned

variables renamed: x.left.tip into xmin and x.right.tip into xmax).

• Support flipping of orientation and guess default orientation automatically

from the aesthetic mapping when possible in stat_fit_tb(),

stat_fit_glance(), stat_fit_tidy() and stat_fit_augment().

• Add stat_distrmix_area() as a wrapper on stat_distrmix_line() with

default for plotting the sum of component Normals as an area.

• Guess default orientation automatically from the aesthetic mapping in

stat_distrmix_line(), stat_distrmix_area() and stat_distrmix_eq().

• Identify regions of the fitted component Normal distributions and their

mixture delimited by arbitrary quantiles in stat_distrmix_line() and

stat_distrmix_area().

• Export function find_quantiles() for the identification of regions delimited

by quantiles when mapping aesthetics in other statistics from 'ggplot2' or its

extension packages.

• Improve in stat_distrmix_line() and stat_distrmix_area() the

behaviour with fullrange = TRUE so that it extends the prediction to cover the

scale range if broader, but never shrinks the range below that needed to

encompass >= 0.999 of the probability. Plots rendered with 'ggpmisc' (>= 1.0.0)

cover a range slightly broader than in previous versions.

• Support methods "nls" and "onls" in stat_poly_line() and

stat_poly_eq().

• Add parameter limit.to to stat_poly_line() and stat_ma_line() as an

enhancement of parameter fullrange providing full control on the newdata

used to compute the prediction line.

• Add parameters fullrange and limit.to to stat_quant_line() and

stat_quant_band() for consistency, adding previously lacking control over the

prediction range.

• **Breaking:** as robustness weights are just one type of posterior weights, the

returned variable robustness.weights is renamed posterior.weights.

• Add support for extraction of posterior weights for models fitted with methods

"gls", ("lme" and "nlme"). These weights are NOT scaled to values in 0..1!

• Change default from show.legend = FALSE to show.legend = TRUE in

stat_fit_residuals() and stat_fit_deviations().

• Fix **bug** in support of smatr::ma() and smatr::sma() in stat_poly_eq()

with - 1 or + 0 in formula: $x^2$ shown instead of $x$.

• Rewrite find_spikes() with changes in behaviour and formal parameters. Old

version did not work as documented, or even usefully.

• Streamline source of documentation using 'roxygen2' inheritance and expand

and clarify explanations. List supported methods in tables in help pages.

Merge help for closely related statistics.

• Update the _User Guide_ with new/better examples and explanations including

use of new features.

• Update the documentation website of 'ggpmisc' editing existing and adding

two new articles: _Non-linear model fits_ and _Multiple comparisons_.