Skip to content

CLI reference

The openauc command mirrors the Python API and adds nothing scientific. It loads, describes, validates, archives and generates — and makes no claim about scientific validity, data quality or suitability for analysis.

uv run openauc --help

Commands

Command Purpose
version Print the installed version
formats List the formats this build can read
generate Write an illustrative synthetic dataset
inspect Print an experiment's factual structural summary
validate Report archival, structural and readiness findings
convert Write an experiment to an AUCX archive

Conventions

INPUT is uniformly "an experiment directory, a data file, or an .aucx archive" — the same dispatch openauc.load performs. You never have to tell the CLI which kind of input you have.

--json is available wherever structured output is useful, so the CLI composes with jq and friends.

Exit codes are stable and documented, which is what makes the CLI scriptable.

Errors print one line. Expected domain errors produce a single error: ... message on stderr and a documented exit code — never a traceback.

Running it

From a clone, with the project environment:

uv run openauc <command>

From an installed wheel, with that environment active:

openauc <command>

Every example in these docs uses the uv run form and is run from the repository root.

What the CLI does not do

No sedimentation-velocity or equilibrium analysis, no fitting, no quality scoring, no plotting subcommand (plotting stays a Python API), no batch or glob input, and no AUCX → CSV conversion.