Refactoring

[GuyliannEngels]

Here is a set of modifications proposed in this pull request.

🧹 Dependency Updates

Updated dependencies: {svBase} and {equatiomatic}.
Removed obsolete dependencies: {data.io} and {svMisc}.

🧱 Refactored tabularise_***() Methods

All tabularise_***() methods for objects such as lm, summary.lm, pls, summary.nls, glm, summary.glm, anova, and aov have been refactored.
This includes functions like tabularise_glance.glm(), tabularise_default.glm(), and tabularise_tidy.glm().

The refactoring improves internal consistency and lays the groundwork for supporting multiple table rendering formats via {flextable}, {tinytable}, and {gt}.

🆕 New Functions :

Introduced summary_(), anova_() functions:

These mirror the behavior of summary() and anova() but retain a reference to the original model object via an added attribute. This allows retrieval of metadata that would otherwise be lost.

Added experimental lm_(), glm_() and nls_() functions:

These extend {stats::lm()}, {stats::glm()} and {stats::nls()} by attaching metadata such as variable labels and units. Argument order now begins with data =, and all three functions (lm_(), glm_(), nls_()) use a data-dot mechanism inspired by svMisc::eval_data_dot().
Note: nls_() defaults to model = TRUE, unlike the base nls() which defaults to FALSE.

[phgrosjean]

A couple of comments:

• There is still bootstrap: 5 in _pkgdown.yml, hence the compilation error in the corresponding GitHub Action.
• Functions like summary_() or anova_() must be generics and more widely available (also required for other packages). So, they must be in svBase for me.