extension vignette

README.qmd

@@ -1,3 +1,5 @@
+
+
 ![](man/figures/zoo_banner.png)
 
 Parameter estimates are often hard to interpret substantively, especially when they are generated by complex models with non-linear components, interactions, or transformations. Instead of reporting unintuitive parameters, data analysts would rather focus on simple quantities of interest, which have straightforward scientific interpretations. Unfortunately, meaningful estimands---and their standard errors---are often tedious to compute, and the terminology used to describe them varies tremendously across fields.
@@ -66,4 +68,4 @@ The `marginaleffects` package and the Marginal Effects Zoo book will always be f
 <img src="https://img.shields.io/badge/license-GPLv3-blue">
 <a href = "https://marginaleffects.com" target = "_blank"><img src="https://img.shields.io/static/v1?label=Website&message=Visit&color=blue"></a>
 <br><br>
-</div>
+</div>

vignettes/extensions.qmd

@@ -150,6 +150,17 @@ predictions(model, newdata = mtcars) |> head()
 
 Note that, for custom model, we typically have to supply values for the `newdata` and `variables` arguments explicitly.
 
+
+### Merging your extension in `marginaleffects`
+
+If you write a working extension for a package on CRAN, please consider submitting your code for inclusion in the package. Other users could greatly benefit from your work!
+
+The steps to do so are pretty much the same as outlined above. However, 
+
+1. Instead of calling `options(marginaleffects_model_classes)`, you need to add your model class to the permanent list of supported models in `R/sanitize_model.R`
+2. Add the package to the supported models vignette by editing this file: `data-raw/supported_models.csv`
+3. Add the valid prediction `type`  values to `R/type_dictionary.R` Note that the first row you insert will be the default prediction type, so please choose something standard and intuitive like "response".
+
 ## Modify or extend supported models
 
 Let's say you want to estimate a model using the `mclogit::mblogit` function.  That package is already supported by `marginaleffects`, but you want to use a `type` (scale) of predictions that is not currently supported: a "centered link scale." 
@@ -205,4 +216,4 @@ Finally, we can call any `slopes` function and obtain results. Notice that our o
 avg_predictions(model)
 
 avg_predictions(model_custom)
-```
+```