Skip to main content

CLI Reference

The entry point is datacrafter (Typer). Global style: --path / -p selects the project directory (default: current working directory). --verbose / -v raises log level.

CommandPurpose
initCreate a project directory and datacrafter.yml
runExecute the pipeline (--dry-run prints a plan)
checkValidate config and environment
statusLatest state.json stages
logTail datacrafter.log
cleanRemove temp files (--storage also clears storage)
schemaInfer field types from output JSONL
metricsRecord counts and field histograms
config validateValidate YAML only
config schemaPrint expected configuration schema
versionPrint package version

Planned (not implemented)​

These commands exist as stubs and print that they are not implemented yet:

  • datacrafter builds — manage builds
  • datacrafter push — push to remote storage
  • datacrafter ui — web UI