Explain a flow: DAG + distribution + sinks

Returns a list with two components: a tibble of stage metadata and a resolved distribution plan (see resolve_dist_plan()). Printing the result shows both the stage table and the full execution layout with all conditional logic (parallelly availability, SLURM env vars, callr heuristics) resolved for the current environment.

Usage

# S3 method for class 'parade_flow'
explain(x, ...)

Arguments

x

A flow().

...

Additional arguments passed to methods (unused).

Value

An explained_parade_flow list (with a print method) containing $stages (tibble) and $dist_plan (resolve_dist_plan() result).

Examples

grid <- data.frame(x = 1:3)
fl <- flow(grid) |>
  stage("sq", function(x) x^2, schema = returns(result = dbl()))
explain(fl)
#> Stages
#> ------
#> # A tibble: 1 × 17
#>   stage needs inputs io_mode fields outputs retries retry_backoff cpus  memory
#>   <chr> <chr> <chr>  <chr>   <chr>  <chr>   <chr>   <chr>         <chr> <chr> 
#> 1 sq    ""    x      off     result result  0       none          ""    ""    
#> # ℹ 7 more variables: time <chr>, cpus_source <chr>, memory_source <chr>,
#> #   time_source <chr>, sink <chr>, prefix <chr>, hoist <chr>
#> 
#> Distribution Plan
#> -----------------
#>   Backend : none
#>   Group by: (none) -- 3 row-level groups
#>   Jobs    : 1 (3 groups/job)
#>   Within  : sequential
#>   Workers : 1 -- No distribution configured