get_execution() exposes the manifest recorded by prep() without
recomputing contrasts, modifying results or rerunning any analysis. Samples
listed in dropped_samples were removed by the centralized model's
missing-value policy, not by method-specific filtering.
Arguments
- rec
A PrepRecipe object.
Value
A list containing the recorded contrasts, dropped_samples,
executed_steps and skipped_steps. The contrasts and skipped-step tables
retain their existing columns, including incompatibility reasons. No
centralized model also yields NULL contrasts in current recipes.
When supplied by an engine, an additional engine_details list records
its effective settings and package version. ALDEx3 records its seed,
scale-model basis, Monte Carlo settings, fitting engine and elapsed time.
Details
Older prepared recipes may lack all or part of this manifest. Unavailable
fields are returned as NULL: missing provenance does not imply that no
samples were removed or no steps were skipped.
Examples
data(metaHIV_phy)
prepared <- recipe(metaHIV_phy) |>
step_filter_by_prevalence(threshold = 0.1) |>
prep(parallel = FALSE)
get_execution(prepared)
#> $contrasts
#> NULL
#>
#> $dropped_samples
#> character(0)
#>
#> $executed_steps
#> character(0)
#>
#> $skipped_steps
#> # A tibble: 0 × 4
#> # ℹ 4 variables: step_id <chr>, engine <chr>, compatible <lgl>, reason <chr>
#>