How to declare function dependencies in R Markdown documents?

[svraka]

tar_render() checks for upstream target dependencies in tar_read() and tar_load() calls. How are function dependencies handled? Obviously functions also used to produce upstream targets loaded in the R Markdown file become indirect dependencies. However, my understanding is that all the global functions and objects are available during knitting the document but I don't see any declared dependencies (e.g. in tar_visnetwork()).

A quick reprex shows that modifying a function called in the R Markdown doesn't invalidate the the report's output. Is there a way to somehow automatically declare these dependencies, or only with custom cues?

targets::tar_dir({
  library(targets)
  
  report <- c("---",
              "title: title",
              "---",
              "```{r}",
              "fun(1)",
              "```")
  writeLines(report, "report.Rmd")

  tar_script({
    library(tarchetypes)
    fun <- function(x) x + 1
    list(tar_render(report, "report.Rmd"))
  }, ask = FALSE)

  tar_make()
  # Modify the function
  writeLines(sub("x + 1", "x + 2", readLines("_targets.R"), fixed = TRUE),
             "_targets.R")
  tar_make()
})
#> • start target report
#> • built target report
#> • end pipeline
#> ✔ skip target report
#> ✔ skip pipeline

^{Created on 2021-09-16 by the reprex package (v2.0.1)}

Session info

sessioninfo::session_info()
#> ─ Session info ───────────────────────────────────────────────────────────────
#>  setting  value                       
#>  version  R version 4.1.1 (2021-08-10)
#>  os       Ubuntu 20.04.3 LTS          
#>  system   x86_64, linux-gnu           
#>  ui       X11                         
#>  language (EN)                        
#>  collate  en_US.UTF-8                 
#>  ctype    en_US.UTF-8                 
#>  tz       Etc/UTC                     
#>  date     2021-09-16                  
#> 
#> ─ Packages ───────────────────────────────────────────────────────────────────
#>  package     * version    date       lib source                           
#>  callr         3.7.0      2021-04-20 [1] RSPM (R 4.0.4)                   
#>  cli           3.0.1      2021-07-17 [1] RSPM (R 4.1.0)                   
#>  codetools     0.2-18     2020-11-04 [3] CRAN (R 4.0.3)                   
#>  crayon        1.4.1      2021-02-08 [1] RSPM (R 4.0.3)                   
#>  data.table    1.14.0     2021-02-21 [1] RSPM (R 4.0.3)                   
#>  digest        0.6.27     2020-10-24 [1] RSPM (R 4.0.3)                   
#>  ellipsis      0.3.2      2021-04-29 [1] RSPM (R 4.0.4)                   
#>  evaluate      0.14       2019-05-28 [1] RSPM (R 4.0.0)                   
#>  fansi         0.5.0      2021-05-25 [1] RSPM (R 4.1.0)                   
#>  fastmap       1.1.0      2021-01-25 [1] RSPM (R 4.0.3)                   
#>  fs            1.5.0      2020-07-31 [1] RSPM (R 4.0.2)                   
#>  glue          1.4.2      2020-08-27 [1] RSPM (R 4.0.2)                   
#>  highr         0.9        2021-04-16 [1] CRAN (R 4.0.5)                   
#>  htmltools     0.5.2      2021-08-25 [1] CRAN (R 4.1.1)                   
#>  igraph        1.2.6      2020-10-06 [1] RSPM (R 4.1.0)                   
#>  knitr         1.34       2021-09-09 [1] RSPM (R 4.1.0)                   
#>  lifecycle     1.0.0      2021-02-15 [1] RSPM (R 4.0.3)                   
#>  magrittr      2.0.1      2020-11-17 [1] RSPM (R 4.0.3)                   
#>  pillar        1.6.2      2021-07-29 [1] RSPM (R 4.1.0)                   
#>  pkgconfig     2.0.3      2019-09-22 [1] RSPM (R 4.0.0)                   
#>  processx      3.5.2      2021-04-30 [1] RSPM (R 4.0.4)                   
#>  ps            1.6.0      2021-02-28 [1] RSPM (R 4.0.3)                   
#>  purrr         0.3.4      2020-04-17 [1] RSPM (R 4.0.0)                   
#>  R6            2.5.1      2021-08-19 [1] RSPM (R 4.1.0)                   
#>  reprex        2.0.1      2021-08-05 [1] CRAN (R 4.1.0)                   
#>  rlang         0.4.11     2021-04-30 [1] RSPM (R 4.0.4)                   
#>  rmarkdown     2.11       2021-09-14 [1] CRAN (R 4.1.1)                   
#>  sessioninfo   1.1.1      2018-11-05 [1] RSPM (R 4.0.0)                   
#>  stringi       1.7.4      2021-08-25 [1] CRAN (R 4.1.1)                   
#>  stringr       1.4.0      2019-02-10 [1] RSPM (R 4.0.0)                   
#>  targets     * 0.7.0.9001 2021-09-16 [1] Github (ropensci/targets@d972f61)
#>  tibble        3.1.4      2021-08-25 [1] CRAN (R 4.1.1)                   
#>  tidyselect    1.1.1      2021-04-30 [1] RSPM (R 4.0.4)                   
#>  utf8          1.2.2      2021-07-24 [1] RSPM (R 4.1.0)                   
#>  vctrs         0.3.8      2021-04-29 [1] RSPM (R 4.0.4)                   
#>  withr         2.4.2      2021-04-18 [1] CRAN (R 4.0.5)                   
#>  xfun          0.26       2021-09-14 [1] CRAN (R 4.1.1)                   
#>  yaml          2.2.1      2020-02-01 [1] RSPM (R 4.0.0)                   
#> 
#> [1] /usr/local/lib/R/site-library
#> [2] /usr/lib/R/site-library
#> [3] /usr/lib/R/library

[wlandau]

You can use the ... arguments in tar_render(). That will change the command of the target to include some mentions of functions. Sketch:

library(targets)

report <- c("---",
            "title: title",
            "---",
            "```{r}",
            "fun(1)",
            "```")
writeLines(report, "report.Rmd")

tar_script({
  library(tarchetypes)
  fun <- function(x) x + 1
  list(
    tar_render(
      report,
      "report.Rmd",
      params = list(fun = fun) # example: function as a parameter
    )
  )
}, ask = FALSE)

tar_make()
#> • start target report
#> • built target report
#> • end pipeline
writeLines(sub("x + 1", "x + 2", readLines("_targets.R"), fixed = TRUE),
           "_targets.R")
tar_make()
#> • start target report
#> • built target report
#> • end pipeline

^{Created on 2021-09-20 by the reprex package (v2.0.1)}

[svraka]

Thanks! It would be cool if targets could discover all global functions automatically in an R Markdown file, so the user won't forget to declare dependencies manually. Although I can see how this might be a niche use case.

[wlandau]

I can think of 1 or 2 ideas, but they already get into tricky territory in terms of maintaining a waterproof abstraction that users can follow.

Function dependencies of R Markdown also run a bit counter to the intended usage pattern. In most situations, any object produced by a function should just be an upstream target loaded into the report.

[svraka]

Function dependencies of R Markdown also run a bit counter to the intended usage pattern. In most situations, any object produced by a function should just be an upstream target loaded into the report.

Thanks for making it clear, I agree. Coming from using Makefiles (and probably a lot a bad practices), I need to adjust to a function based pipeline.