gridl/modelgenerics

Generic Methods Related to Model Fitting

modelgenerics is designed to alleviate the situation where multiple packages use the same class. For example, if two packages have methods for a fit class, there will be conflicts when both packages are loaded. One approach would be for one package to import the other but this may result in many additional package dependencies during installation.

modelgenerics is a simple, lightweight packages that contains S3 and S4 generic definitions that other packages can use to avoid this problem. Some examples are:

> library(modelgenerics)
> 
> fit
standardGeneric for "fit" defined from package "modelgenerics"

function (object, ...) 
standardGeneric("fit")
<bytecode: 0x10889d930>
<environment: 0x108899b98>
Methods may be defined for arguments: object
Use  showMethods("fit")  for currently available ones.
> 
> train
standardGeneric for "train" defined from package "modelgenerics"

function (x, ...) 
standardGeneric("train")
<bytecode: 0x102993990>
<environment: 0x10298cb80>
Methods may be defined for arguments: x
Use  showMethods("train")  for currently available ones.

To use this in your package, It is suggested that, to do so by importing the package and re-exporting the method of interest. For example, if a S3 fit method were being used, the roxygen2 code to do this would be

#' @importFrom modelgenerics explain
#' @export
modelgenerics::explain

As an example, the recipes package defines a number of tidy S3 methods by importing this package (whereas it previously depended on broom).

To install it modelgenerics, use:

require("devtools")
install_github("tidymodels/modelgenerics")