NEWS.md
In datalorax/equatiomatic: Transform Models into 'LaTeX' Equations

equatiomatic 0.3.7

Extracts equations for summary.lm PR #245.
Correction in the plotting-integration vignette: ggplot2 now requires a vector for text annotation, so, use as.character() and specify parse = TRUE () PR #244.

The penguins dataset is now explicitly loaded from {equatiomatic} : another version appears in the {datasets} package version 4.5.0 with different column names and it causes some tests to fail and confusion in the examples.
Correction of an URL (https instead of http) in the lme4-lmer vignette.

{tidymodels} partial compatibility: method extract_eq() added for model_fit objects of the {parsnip} package, and for workflow objects of the {workflows} package, see #237. Previously, the model object had to be extracted using parsnip::fit() before using extract_eq(). This is not necessary any more, but still valid (no breaking changes). Note that extract_eq() cannot handle yet all {tidymodels}. For instance, it can handle reg_linear(), but cannot handle reg_linear() |> set_engine("stan"). There is currently no plan to support all {tidymodels} models, but pull requests will be considered if you find and solve a problematic case by yourself.

Vignette 'intro-equatiomatic' renamed 'equatiomatic' to enable a "Getting Started" entry in the pkgdown site.
Documentation was cleaned up. In particular, methods for internal generics are not exported any more. All methods for extract_eq() are now properly documented in the extract_eq() help page.

New maintainer (Philippe Grosjean phgrosjean@sciviews.org).
Remotes to yonicd/texPreview eliminated because a suitable version is on CRAN.
Respect of the arguments of the generic functions for add_greek.forecast_ARIMA and add_coefs.forecast_ARIMA (first argument side= is renamed rhs=).

Minor bug fix related to lifecycle badges
New license: CC-BY
Bug fix: If names overlap, prior version had an error with the ordering and construction of the coefficients.
Bug fix: Prior versions did not escape characters in multilevel models when declaring the grouping factor (e.g., for census_division l = 1 is now rendered as for census\_division l = 1).
Feature addition: added se_subscripts argument, which allows the standard error for each coefficient to be included in parentheses below the coefficient when se_subscripts = TRUE. This is supported for lm and glm models.

Export new renderEq() and eqOutput() functions for working with equatiomatic with shiny.
Specific parts of the equation can now be colored with the new greek_colors, subscript_colors, var_colors and var_subscript_colors arguments.
Includes new swap_var_names and swap_subscript_names arguments to change the names of the variables or subscripts in the rendered equation
A new vignette documents the usage of colors and name swapping.
Now accommodates lme4::glmer() models.
Include new return_variances argument, which allows users to optionally return the variance/covariances in lme4::lmer() and lme4::glmer() models
Added new font_size argument, which takes any LaTeX font size (see here)
Added new label argument, which allows for cross-referencing equation in-text with PDF outputs.
Now includes rendering of poly(), log(), and exp() functions for in-line equations, and will drop the I() for generic in-line equation operation
Bug fix related to categorical variables and level parsing for lme4::lmer() and and lme4::glmer() models
Minor bug fix related to indexing of coefficients for lme4::lmer() models
Minor bug fix related to very large models, which would not render properly before
lme4::lmer() models now allow for in-line alterations to the code, e.g., I(n >5).
Minor bug fix related to SARIMA models

New models: This is the first version to support lme4::lmer() models and forecast::Arima() models.
New vignettes for each of the new models supported.
New vignette showing how to use the package with plotting
Switched the testing framework to use snapshot testing
Uses new print method so users no longer have to specify results = "asis" in the R Markdown chunk option
Fixed error in rendering logistic/probit regression equations by removing the epsilon (error term) at the end of the left-hand side
The epsilon (error term) is no longer shown when rendering the fitted model equation, i.e. use_coefs = TRUE, for lm models. The hat sign is also added to the response variable.

Initial CRAN Release
Extracts equations for lm models
Extracts equations for glm models with family = binomial(link = "logit") or family = binomial(link = "probit")
Extracts equations form ordered regression models using MASS::polr or ordered::clm for logit and probit link functions
glm models have an optional show_distribution argument to show the distributional assumptions
All equations can be displayed using Greek notation or the estimated coefficients
The raw_tex code can be used to supply custom TeX code for the intercept or the coefficients, through the intercept and greek arguments respectively
Long equations can be wrapped to multiple lines with the optional wrap argument. The length of the wrapping is controlled by terms_per_line.
Added a NEWS.md file to track subsequent changes to the package.