Pytest has poor support/no support for test CLI commands, so we need a testing harness for the CLI, that is capable of:

• Running a shell command
• Handling the output and returning:
  • Output to STDOUT, STDERR as applicable
  • The shell return code, a signed 8-bit integer in POSIX systems
• Making assertions based on this

E.g.

def test_it_creates_a_dataset():
    stdout, stderr, output_code = run_cli_command("darwin dataset create test-dataset")

    assert stdout.contains("test-dataset created")
    assert output_code == 0
    assert not stderr

This repo that I started work on some time ago may be of help, but is maybe a bit more than we need, but the file https://github.com/owencjones/cli_test/blob/main/cli_test/types.py contains a lot of what we need. It creates an expectation class, which we could use, but the functional approach above is just as good.