Skip to content

fit.epi_workflow silently drops epi_workflow class when built or fit on a non-epi_df or non-epi_recipe in most cases #363

Closed
@brookslogan

Description

@brookslogan

I assume this will cause issues down the line when users are unexpectedly being missing out on slathering, or even earlier in the operation any step_epi*.

library(recipes)
#> Loading required package: dplyr
#> 
#> Attaching package: 'dplyr'
#> The following objects are masked from 'package:stats':
#> 
#>     filter, lag
#> The following objects are masked from 'package:base':
#> 
#>     intersect, setdiff, setequal, union
#> 
#> Attaching package: 'recipes'
#> The following object is masked from 'package:stats':
#> 
#>     step
library(epipredict)
#> Loading required package: epiprocess
#> Registered S3 method overwritten by 'tsibble':
#>   method               from 
#>   as_tibble.grouped_df dplyr
#> 
#> Attaching package: 'epiprocess'
#> The following object is masked from 'package:stats':
#> 
#>     filter
#> Loading required package: parsnip
#> Registered S3 method overwritten by 'epipredict':
#>   method            from   
#>   print.step_naomit recipes
library(testthat)
#> 
#> Attaching package: 'testthat'
#> The following object is masked from 'package:dplyr':
#> 
#>     matches

tbl <- tibble::tibble(
  geo_value = 1,
  time_value = 1:100,
  x = 1:100,
  y = x + rnorm(100L)
)
edf <- as_epi_df(tbl)

rec_tbl <- recipe(y ~ x, data = tbl)
rec_edf <- recipe(y ~ x, data = edf)
erec_tbl <- epi_recipe(y ~ x, data = tbl) # warning, good; not testing further; why not error though?
#> Warning: epi_recipe has been called with a non-epi_df object, returning a
#> regular recipe. Various step_epi_* functions will not work.
erec_edf <- epi_recipe(y ~ x, data = edf)

ewf_rec_tbl <- epi_workflow(rec_tbl, linear_reg())
ewf_rec_edf <- epi_workflow(rec_edf, linear_reg())
ewf_erec_edf <- epi_workflow(erec_edf, linear_reg())

# above are all epi_workflows:

expect_s3_class(ewf_rec_tbl, "epi_workflow")
expect_s3_class(ewf_rec_edf, "epi_workflow")
expect_s3_class(ewf_erec_edf, "epi_workflow")

# but fitting drops the class or generates errors in many cases:

ewf_rec_tbl %>% fit(tbl) %>% class()
#> [1] "workflow"
ewf_rec_tbl %>% fit(edf) %>% class()
#> [1] "workflow"
ewf_rec_edf %>% fit(tbl) %>% class()
#> [1] "workflow"
ewf_rec_edf %>% fit(edf) %>% class()
#> [1] "workflow"
ewf_erec_edf %>% fit(tbl) %>% class()
#> Error in if (new_meta != old_meta) {: argument is of length zero
ewf_erec_edf %>% fit(edf) %>% class()
#> [1] "epi_workflow" "workflow"

Created on 2024-07-18 with reprex v2.1.1

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions