running aws pipeline with tar_dir

[yonicd]

I tried running the basic example for aws with a self contained tar_dir and got this error message even though my rprofile is set up with the relevant envars

library(targets)
tar_dir({ 
  tar_option_set(resources = list(bucket = "abc"))
  write_mean <- function(data) {
    tmp <- tempfile()
    writeLines(as.character(mean(data)), tmp)
    tmp
  }  
  tar_script(
    list(
      tar_target(data, rnorm(5), format = "aws_qs"),
      tar_target(mean_file, write_mean(data), format = "aws_file")
    ), 
    ask = FALSE
  )
  tar_make()
})

* run target data
* end pipeline
Error : S3 bucket name cannot be empty.
 Error: callr subprocess failed: S3 bucket name cannot be empty.
Visit https://books.ropensci.org/targets/debugging.html for debugging advice.

[yonicd]

figured it out:

tar_dir({ 
  tar_script({
    write_mean <- function(data) {
      tmp <- tempfile()
      writeLines(as.character(mean(data)), tmp)
      tmp
    }
    tar_option_set(resources = list(bucket = "abc"))
    list(
      tar_target(data, rnorm(5),format = 'aws_qs'),
      tar_target(mean_file, write_mean(data),format = 'aws_file')
    )
  }, 
    ask = FALSE,
    library_targets = TRUE
  )
  tar_make()
})
* run target data
* run target mean_file
* end pipeline