pelagos_py.steps.base_step

pelagos_py.steps.base_step.REGISTERED_STEPS

Registry of explicitly registered step classes.

pelagos_py.steps.base_step.register_step(cls)

Decorator to mark a step class for inclusion in the pipeline.

class pelagos_py.steps.base_step.BaseStep(name, parameters=None, diagnostics=False, context=None)

Bases: pelagos_py.utils.config_mirror.ConfigMirrorMixin

Base class for pipeline steps with config-mirroring support. Every concrete subclass (registered via @register_step) inherits this.

classmethod describe_parameters()

Return a JSON-serialisable description of this step’s parameters.

Introspection surface for external tools (e.g. a dashboard) that need to render a parameter form without instantiating the step. See pelagos_py.utils.parameter_spec.describe().

log(message)

Log an info-level message with step name prefix.

log_warn(message, warning_type=UserWarning)

Log a warning-level message with step name prefix.

check_data()

Check for data in context for transformer steps.

halt(message)

Stop the pipeline cleanly for an unrecoverable misconfiguration.

Logs message (the detail and any fix/install hint) at ERROR level, then a terminal STOP line marking the deliberate halt, and exits. Use this instead of raising for config problems discovered at run time, so the user sees one actionable error rather than a wrapped traceback.

update_parameters(**kwargs)

Update parameter values both in attributes and in private store.

Example:

self.update_parameters(file_path='newfile.nc', add_meta=False)

generate_config()

Return this step’s config dict (suitable for saving to YAML).

save_config(path: str | None = None)

Save this step’s config to YAML (for standalone debugging).