Autoload and Load¶

The two entry points for integrating behave-steplib with your environment.py.

`autoload(context)`¶

Discovers all installed steplib plugins via Python entry points and registers their steps with behave. Returns a SteplibState attached to context.steplib.

from steplib.behave import autoload

def before_all(context):
    context.steplib = autoload(context)

Optional filters¶

context.steplib = autoload(
    context,
    categories=["api"],           # only load API steps
    backends={"api": "httpx"},    # use httpx backend for API
)

categories: A list of category names to keep. Steps whose category is not in the list are removed from the registry before behave sees them.
backends: A mapping of category → backend. For each category present in the mapping, only steps whose backend matches the requested value are kept. Steps without a backend are always kept.

When both are None (the default), every installed plugin is loaded with every backend.

`load(context, *modules)`¶

Imports specific step modules by dotted path. No entry point discovery is performed — you explicitly name the modules to load.

from steplib.behave import load

def before_all(context):
    context.steplib = load(
        context,
        "steplib.modules.api.steps",
        "steplib.modules.db.steps",
    )

Each module must expose a register(registry) function. If it does not, an AttributeError is raised with a clear message.

Lifecycle hooks¶

The SteplibState object exposes two lifecycle methods that should be called from behave hooks:

from steplib.behave import after_scenario, autoload

def before_all(context):
    context.steplib = autoload(context)

def before_scenario(context, scenario):
    context.steplib.reset()    # reset per-scenario state

def after_scenario(context, scenario):
    after_scenario(context, scenario)  # close resources

reset(): Iterates over all module-level attributes on the state (non-underscore) and calls reset() on each that has it. This clears per-scenario data such as the last HTTP response or the last query result, while keeping configuration (base URL, default headers, connection strings).
cleanup(): Iterates over the same attributes and calls cleanup() on each that has it. This closes HTTP clients, browser drivers, database connections and Kafka producers/consumers.

You can also call context.steplib.cleanup() directly instead of importing after_scenario — both are equivalent.

How discovery works¶

autoload uses importlib.metadata.entry_points() to find all packages that register steps under the steplib.plugins entry point group:

[project.entry-points."steplib.plugins"]
api = "steplib.modules.api.steps:register"
web = "steplib.modules.web.steps:register"
db = "steplib.modules.db.steps:register"
kafka = "steplib.modules.kafka.steps:register"

Third-party packages can add their own entry points in the same group:

[project.entry-points."steplib.plugins"]
mycompany = "mycompany.steps:register"

When autoload runs, it loads every entry point in the group and calls each register(registry) function. This means any package installed in the environment that declares a steplib.plugins entry point is automatically discovered — no manual imports needed.

API reference¶

See Core API Reference for the full autodoc reference of steplib.behave and steplib.core.state.