function to get prediction columns #1224
Conversation
| res$probabilities <- ".pred" | ||
| } | ||
| } else { | ||
| cli::cli_abort("Unsupported model mode {model_mode}.") |
There was a problem hiding this comment.
I don't know that this branch of the if/then will ever be encountered, given the error check above. However, we might hit is when we have a mode for quantile regression data so I'd err on the side of leaving it in, untested.
There was a problem hiding this comment.
i think that is reasonable
Co-authored-by: Emil Hvitfeldt <emil.hvitfeldt@posit.co>
| #' | ||
| #' .get_prediction_column_names(lr_fit) | ||
| #' .get_prediction_column_names(lr_fit, syms = TRUE) | ||
| #' @export |
There was a problem hiding this comment.
Could we make this function @keyword internal?
| if (inherits(x, "workflow")) { | ||
| x <- x %>% extract_fit_parsnip(x) | ||
| } |
There was a problem hiding this comment.
With an eye toward the release cascade and trying to keep things as modular as possible, I think it would be good to make this function only for model_spec objects. This call to extract_fit_parsnip() seems to be all that's necessary for workflow objects. I'd move that to workflows or wherever this is going to be needed.
| if (length(predict_types) == 0) { | ||
| cli::cli_abort("Prediction information could not be found for this | ||
| {.fn {model_type}} with engine {.val {model_engine}} and mode | ||
| {.val {model_mode}}. Does a parsnip extension package need to |
There was a problem hiding this comment.
That last question would be very nice in an i bullet rather than the error message itself.
| ) | ||
|
|
||
| unk_fit <- ols_fit | ||
| unk_fit$spec$mode <- "Depeche" |
|
This pull request has been automatically locked. If you believe you have found a related problem, please file a new issue (with a reprex: https://reprex.tidyverse.org) and link to this issue. |
In service of tidymodels/workflows#273
For a fitted model or workflow, return the column names that will contain predictions for the main types given by
augment().Tests for censored regression models and workflows will be in extratests