Start a study folder for drafted report files

Writes what every drafted report file stands on, so that the AI writes only the reports: study.yml (the study's id and settings), spec/data.yml (each ADaM file as a dataset, and an analysis set for each population flag of ADSL – SAFFL, FASFL, ITTFL, PPROTFL ... – as <flag> == "Y") and spec/defaults.yml (the page every report starts from: the running header and footer, and the cell of a count). Each report file then says extends: [defaults] and only what differs. Check the analysis sets' ids and labels: the table prompt gives them to the AI.

Usage

tfl_ai_study_init(
  dir,
  adam = "adam",
  study_id = "STUDY-01",
  title = NULL,
  overwrite = FALSE
)

Arguments

dir

The study folder (made if need be).

adam

The ADaM folder, as the study's programs reach it from dir (relative, such as "adam", or absolute); its .xpt (or .sas7bdat, .rds) files become the datasets.

study_id, title

The study's id and title.

overwrite

Write over files already there?

Value

The paths written, invisibly.

Examples

d <- file.path(tempdir(), "study-init")
dir.create(file.path(d, "adam"), recursive = TRUE, showWarnings = FALSE)
saveRDS(tflspec::adsl, file.path(d, "adam", "adsl.rds"))
tfl_ai_study_init(d, adam = "adam", study_id = "ONCO-01", overwrite = TRUE)
cat(readLines(file.path(d, "spec", "data.yml")), sep = "\n")
#> datasets:
#>   - dataset: ADSL
#>     level: ADaM
#>     path: adam/adsl.rds
#> populations:
#>   - id: ENRL
#>     dataset: ADSL
#>     where: !r ENRLFL == "Y"
#>     label: Enrolled Population
#>   - id: ITT
#>     dataset: ADSL
#>     where: !r ITTFL == "Y"
#>     label: Intent-To-Treat Population
#>   - id: SAF
#>     dataset: ADSL
#>     where: !r SAFFL == "Y"
#>     label: Safety Population
#>   - id: EFF
#>     dataset: ADSL
#>     where: !r EFFFL == "Y"
#>     label: Efficacy Evaluable Population