Production Study Pipeline Execution — execStudyPipeline • picard

Executes the full study pipeline in production mode with full validation, version management, and reproducibility tracking. Creates a release branch, runs the complete pipeline, provides PR instructions, and saves reference to PENDING_PR.md.

Usage

execStudyPipeline(
  configBlock,
  updateType,
  skipRenv = FALSE,
  skipConnectivityCheck = TRUE,
  ignoreUncommittedPaths = NULL,
  skipCodeStateCheck = FALSE,
  env = rlang::caller_env()
)

Arguments

configBlock

Character or character vector. Name(s) of config block(s) to use.

updateType

Character. Type of version increment: 'major', 'minor', or 'patch'.

MAJOR: Breaking changes
MINOR: New features, backward compatible
PATCH: Bug fixes, no new features

skipRenv

Logical. If TRUE, skips renv validation. Defaults to FALSE. Useful for testing issues. Default: FALSE

skipConnectivityCheck

Logical. If TRUE (default), skips the optional database connectivity pre-flight check. Set to FALSE to attempt a test connection to each config block before execution begins.

ignoreUncommittedPaths

Character vector or NULL. Repo-relative paths (e.g. "inputs") whose uncommitted changes should not fail the code-state pre-flight check. Changes anywhere else — notably analysis/ — still fail it. When NULL (default) the list is read from ignoreUncommittedPaths in the default: block of config.yml, which itself defaults to ignoring nothing, so behaviour is unchanged unless a study opts in. Pass character(0) to force strict checking regardless of config.yml.

skipCodeStateCheck

Logical. If TRUE, skips the code-state check entirely — a last resort when ignoreUncommittedPaths cannot express the churn. The run proceeds, but the pre-flight checklist raises a warning and the run history records the tree as "unverified-skipped". Defaults to FALSE. Deliberately not settable from config.yml, so it cannot be baked permanently into a study.

env

The execution environment. Defaults to caller environment.

Value

Invisibly returns task results list

Examples

if (FALSE) {
# Run production pipeline with patch version increment
execStudyPipeline(configBlock = "myConfig", updateType = "patch")

# Tolerate manifest/ATLAS churn under inputs/ for this run only
execStudyPipeline(
  configBlock = "myConfig",
  updateType = "patch",
  ignoreUncommittedPaths = "inputs"
)

# Last resort: do not check the working tree at all
execStudyPipeline(
  configBlock = "myConfig",
  updateType = "patch",
  skipCodeStateCheck = TRUE
)
}