How about adding usemodels_addin() to make it more tied to workflowsets?

[amazongodman]

This is a note about features I would like to see added in the future.

I listened to Max Kuhn's talk at LA R Users. https://www.youtube.com/watch?v=2OfTEakSFXQ

At around 35:00 of this talk, he was trying to create multiple models with parsnip_addin() in order to include multiple models in workflowsets.

usemodels will output both recipes and models. I was commenting that it would be ideal to have a function like parsnip_addin() where we can use shiny to make a visual selection and then the recipe and model are output.

I can't help you technically, but if you find the above useful in the future, I'd be happy to have it on the implementation list.

P.S. Do you have any recommended tutorials to help with the technical implementation?

Finally. Many thanks to the developers of tidymodels.

[juliasilge]

Thanks for the idea! We'll keep this issue open here to prioritize it with other work.

[topepo]

it would be ideal to have a function like parsnip_addin() where we can use shiny to make a visual selection and then the recipe and model are output.

That's a good idea but fairly out-of-scope for this project. I think that the details are all there. Maybe make a package to do that?

[jrosell]

Would you be open to accept a new PR with a new workflowset = TRUE argument that it doesn't generate the last *_tune lines when tune = TRUE? I also image a new use_all function that can generate all classification models or all regression models and the final workflowset.