Skip to contents

Run the artma command-line interface

Source: R/cli.R
cli_run.Rd

Scriptable entry point behind the artma launcher and Rscript -e 'artma::cli_run()'. It parses an argument vector, dispatches to the matching public function (artma(), methods_list(), options_validate() / options_create() / options_list(), or the package version) and returns an exit code. No analysis logic lives here: each subcommand is a thin translation to the existing API.

The function never calls quit(); it returns the exit code invisibly so it can be driven in-process by tests. The launcher script turns that code into a process exit status.

Exit codes:

  • 0 success (including --help).

  • 1 an R error was raised while dispatching (message printed to stderr).

  • 2 a usage error (unknown command, unknown flag, malformed value); usage is printed to stderr.

Subcommands: run, methods, options (with sub-actions validate, create, list) and version. Every subcommand accepts --help.

Flags for run become an in-session options() overlay applied around the artma() call, so the user's YAML options file is never mutated: --data sets artma.data.source_path, --output-dir sets artma.output.dir, --verbose sets artma.verbose, --no-cache sets artma.cache.use_cache to FALSE, and --report sets artma.output.report to TRUE. --options, --options-dir and --methods are forwarded as artma() arguments.

In --json mode stdout carries only a JSON run manifest (methods_run, methods_skipped with reasons, output_dir, exported_files, seed, package_version), read back from the run.json the run wrote; all human-readable output is routed to stderr.

Usage

cli_run(args = commandArgs(trailingOnly = TRUE))

Arguments

args

[character, optional] The argument vector to parse. Defaults to commandArgs(trailingOnly = TRUE).

Value

[integer] The exit code (invisibly).

Examples

if (FALSE) { # \dontrun{
# List available methods
artma::cli_run("methods")

# Run two methods against an options file, emitting a JSON manifest
artma::cli_run(c(
  "run", "--options", "my_analysis.yaml",
  "--methods", "funnel_plot,effect_summary_stats", "--json"
))
} # }