NEWS.md
Dependency to svBase and equatiomatic updated, and dependencies to data.io and svMisc eliminated.
Refactored all tabularise_***() methods for glm, summary.glm, anova, aov objects (e.g., tabularise_glance.glm(), tabularise_default.glm(), tabularise_tidy.glm(), etc.) to improve internal consistency and prepare for multi-format table rendering using {flextable}, {tinytable}, and {gt}.
added summary_() and anova_() functions. These two functions provide the same information as the summary() and anova() functions. They add an attribute that preserves a link to the object used in these two functions. Adding this argument makes it possible to retrieve information from the original object that would otherwise be lost when using summary() or anova().
Added the experimental glm_() and nls_() functions, which extend [stats::glm()] and [stats::nls()] by attaching additional metadata such as variable labels and units. The order of arguments has been modified to start with data =. All three functions — lm_(), glm_(), and nls_() — use a data-dot mechanism inspired by svMisc::eval_data_dot(). Note: nls_() uses model = TRUE by default, whereas the base nls() function uses model = FALSE.
The equation=
argument now accepts NA
in addition to TRUE
, FALSE
, and character strings, offering more flexibility in how model equations are displayed and used:
TRUE
(default): The equation is automatically generated and added to the table header. Its parameters are also used in the “Term” column.
FALSE
: No equation is generated or displayed, and its parameters are not used in the “Term” column.
NA
: The equation is generated but not displayed in the table header. Its parameters are still used in the “Term” column.
Character string: A custom equation is provided directly and added to the table header.
This enhancement allows for finer control over equation display and ensures that model parameters remain accessible even when the equation is hidden.
Refactored all tabularise_***() methods for lm, summary.lm, nls, summary.nls objects (e.g., tabularise_glance.lm(), tabularise_default.lm(), tabularise_tidy.lm(), etc.) to improve internal consistency and prepare for multi-format table rendering using {flextable}, {tinytable}, and {gt}.
Labels are now fully integrated into tabularise_***.nls and tabularise_***.summary.nls, allowing consistent and readable output for nonlinear model summaries.
Added the experimental lm_() function, which extends [stats::lm()] by attaching additional metadata such as variable labels and units. Also introduced summary.lm_() and anova.lm_() methods to preserve the additional metadata.
Better selection of X and Y variables in chart.nls()
and autoplot.nls()
.
Bug correction in tabularise()
for lm and glm object. Managing a conflict between auto.labs=
and equation=
.
Argument kind =
added to all tabularise()
methods and dependency update to v0.6.0 of {tabularise}.
Enhanced vignette and README file.
License switched to MIT.
...
argument in equation.nl()
and equation.summary.nls()
was causing a warning in R CMD check
. Corrected now.autoplot.nls()
: X and Y axes were inverted. Also the deprecated aes_string()
function is replaced with aes()
with proper arguments.tabularise()
.Use of the generic equation()
. equation_nls()
is transformed into an equation()
method for nls and summary.nls objects. Implementation of ital_vars=
, use_coefs=
, coef_digits=
and fix_signs=
arguments in order to get parameterized as well as non parameterized equation, with or without italic for the variables.
Dependency to {equatiomatic} is dropped. It is now included in {tabularise}.
The default for title=
argument is now automatically set to FALSE
if the table is generated inside a chunk that has tbl-cap
YAML entry defined (thus replaced by that caption).
The tabularise_default()
methods for summary.lm and summary.glm were lacking important information (they were identical to tabularise_coef)
for the same objects. Corrected now.
tabularise_default()
and tabularise_tidy()
added for anova objects.
tabularise_tidy()
added for aov objects.
tabularise_default()
, tabularise_coef()
, tabularise_glance()
, tabularise_tidy()
for lm objects
tabularise_default()
, tabularise_coef()
for summary.lm objects
tabularise_coef()
, tabularise_glance()
, tabularise_tidy()
for glm objects
tabularise_coef()
for summary.glm objects
tabularise_default()
, tabularise_coef()
, tabularise_glance()
, tabularise_tidy()
for nls objects
tabularise_default()
, tabularise_coef()
for summary.nls objects
autoplot()
and chart()
for objects lm, glm and nls: both the plot of the model and several residual analyses, plus a composed 2x2 plot for residuals analysis.
as.function()
for lm and nls objects.
Reexportation of several functions from {modelr}: add_predictions()
, add_residuals()
, geom_ref_line()
, rmse()
, rsquare()
, mae()
and qae()
.
fit_model()
to easily fit a {parsnip} model, and a series of methods like summary()
, anova()
or chart()
directly implemented for these objects.
SciViews::R
p <- chart(data = mtcars, wt ~ mpg) + geom_point()
p <- p + geom_smooth(method = "lm", formula = y ~ I(x^2) + x)
p <- p + theme_void() + hexSticker::theme_transparent()
dir.create ("inst/figures", showWarnings = FALSE)
hexSticker::sticker(p, package = "modelit", p_size = 9, s_x = 1, s_y = .75,
s_width = 1.3, s_height = 0.9, h_fill="lightblue3", h_color="blue",
filename = "inst/figures/modelit.png")
NEWS.md
file to track changes to the package.