API lifecycle and canonical interfaces

A large scientific package needs a navigable public surface. The 0.9 lifecycle layer inventories exports, maps them to conceptual families, and carries explicit statuses such as core, workflow, advanced, experimental, gated, compatibility, and deprecated. Unknown future symbols remain unreviewed; maturity is never inferred from a name alone.

import eyeprocesspy as ep

registry = ep.eye_api_lifecycle()
inventory = ep.eye_api_inventory()  # frozen R 0.11.1 public reference surface
registry = ep.register_eye_api_status(
    registry,
    "run_eye_pipeline",
    "workflow",
    canonical="run_eye_pipeline",
)
audit = ep.audit_eye_api(inventory, registry)
summary = ep.api_surface_summary(audit.table)
recommendations = ep.eye_api_recommendation(audit)

Packaged lifecycle closure

The Python package ships the frozen 0.9 lifecycle registry copied from the eyeprocess 0.11.1 source. It contains all 1,182 R exports and the 108-row module policy. Calling eye_api_inventory() with its default package="eyeprocess" inventories that frozen reference surface. Use package="eyeprocesspy" when you specifically want the currently implemented Python namespace.

Lifecycle status describes software-interface maturity and governance. It does not establish construct validity, empirical adequacy, device equivalence, or the scientific interpretation of gaze, pupil, response-time, sequence, or psychometric measures.