pytest-big-dill

The pytest plugin for Big Dill — tooling for pytest-bdd that treats the feature file as the specification.

pip install pytest-big-dill

It registers with pytest automatically. Install it into the environment your tests run in.

What it does

• Feature metadata — attaches the .feature path, scenario name, tags and scenario line to each collected pytest-bdd item, so a host can present tests grouped by feature file rather than by Python module.
• Custom scenario names — name outline rows something meaningful ([E01]) rather than a concatenation of parameter values.
• Custom statuses — report your own outcome strings (waiting, knownError) for a host to map onto its own test states.
• Linting — pytest --bdd-lint statically checks feature files: typed step parameter validation, plus any rules you add.
• Reporting — sends discovery and execution payloads to a host over a local pipe when one is listening, and is inert otherwise.

Where this fits

Useful on its own as a CI lint gate; the editor experience needs the extension too.

Hooks

Implement any of these in your conftest.py:

def pytest_big_dill_test_name(scenario_name, example_params, feature_name, feature_path):
    """A display name for a scenario, or None to keep the default."""
    if example_params.get("id"):
        return f"{scenario_name} [{example_params['id']}]"
    return None


def pytest_big_dill_custom_status(report, config):
    """A custom status string for a result, or None."""
    if getattr(report, "wasxfail", None):
        return "knownError"
    return None

Two more add your own lint rules, each returning a list of LintDiagnostic:

from pytest_big_dill.lint_types import LintDiagnostic

def pytest_big_dill_lint_outline(scenario, examples):
    """Once per Scenario Outline during --bdd-lint."""

def pytest_big_dill_lint_scenario(scenario):
    """Per scenario, and per interpolated Examples row."""

Full reference, including typed step parameters: developer guide.

Linting

pytest --bdd-lint              # every discovered feature file
pytest --bdd-lint path.feature

Exits non-zero on any error-severity diagnostic, so it works as a CI gate.

This package contributes the runtime half of the lint story: typed step parameter validation, undefined-step detection, and whatever your own hooks add. Structural rules about the Gherkin itself — duplicate scenario names, unused Examples columns, outlines with a single row — come from the engine and are listed with everything else in the lint rules reference.

Documentation

• Developer guide — hookspecs, typed steps, custom lint rules
• Lint rules — every diagnostic, across all three packages
• Repository

License

MIT. See LICENSE and THIRD-PARTY-NOTICES.md.