Config#
The Registry is the central registration
mechanism for all extensible components in MyoGestic. It is instantiated
once as CONFIG_REGISTRY and imported throughout the application.
Built-in component registrations (models, features, filters, VIs).
User-defined component registrations – add your own models, features, filters, VIs and output systems here.
Registration Workflow#
Default components (built-in models, features, VIs) are registered in
myogestic.default_config. User extensions are registered in
myogestic.user_config, which is loaded after the defaults and can
add new components or override existing ones.
The following elements can be registered:
Models – via
register_model()Features – via
register_feature()(must be capable of running in real-time)Post-prediction filters – via
register_real_time_filter()(mainly for smoothing regression output)Visual interfaces – via
register_visual_interface()Output systems – via
register_output_system()
See also
The Extend MyoGestic tutorials walk through adding each type of component.
Registry#
|
The registry class is used to store different components of a MyoGestic application pipeline. |
Configuration Parameters#
Changeable Parameter#
- myogestic.utils.config.ChangeableParameter#
Represent a PEP 604 union type
E.g. for int | str
Unchangeable Parameter#
- myogestic.utils.config.UnchangeableParameter#
Represent a PEP 604 union type
E.g. for int | str