Official runs of a study — run

run_batch() runs the study's official-run program – the ARD (programs/ard/autoexec_ard.R), the reports (programs/tfl/autoexec_report.R) or both (programs/autoexec_all.R) – which makes a batch folder runs/<date>_<time>_<what>/ with each program's log, what the run made, the R and package versions (session_info.txt, and the study's renv.lock when it has one), with code = TRUE the programs and definition workbooks it ran, and with data = TRUE the data files its programs read. list_batches() lists a study's batch folders.

Usage

run_batch(
  study,
  parts = c("ard", "tfl"),
  code = TRUE,
  data = FALSE,
  only = NULL,
  exclude = NULL,
  batch = NULL,
  wait = TRUE
)

list_batches(study)

Arguments

study: An rtfstudy.
parts: "ard", "tfl" or both.
code: Keep the code in the batch folder.
data: Keep the data files the programs read in the batch folder.
only: Programs or outputs to run; NULL for all.
exclude: Reports to leave out (their ARD program and their report program), by output id or program name; NULL for none. The batch folder's batch.txt lists the programs left out.
batch: A named batch (batch_sets()): only its reports run, and the batch folder (runs/<date>_<time>_<what>_<batch>/) and its batch.txt name it. NULL: every report (or only / exclude).
wait: FALSE returns the running processx::process at once. The run goes on by itself: closing the app, or R, does not stop it, and a study runs one official run at a time.

Value

run_batch(): with wait = TRUE, a list: ok, batch (the batch folder), result (its run.csv), output (what the run printed); otherwise the process. list_batches(): a data frame, one row a batch folder, the newest first, with its state – "done" (every program without error), "failed" (errors of programs), or "incomplete" (no record of its end yet: still running, or stopped) – and when it finished.

Details

The data kept are the files of the data catalog the run's programs read (batch_data_plan()): each ARD program its report's datasets, each report program its own (a listing's, a figure's) or the ARD taken in it reads. They go to the batch folder's data/ under the path they have in the study folder (data/data/adam/adsl.rds; a file outside it under data/external/), listed in data_manifest.csv with their size, md5 and time. With the code and the logs, that is everything the run read: a run that can be made again.

The programs are the saved ones: save the study first.