Components

This section explains the key components provided by ml-orchestrator with examples.

MetaComponent and Variants

The MetaComponent is a base class designed for implementing custom components for machine learning workflows.

Example:

from ml_orchestrator.meta_comp import MetaComponent

class DummyComponent(MetaComponent):
    @property
    def env(self):
        return self.env_params()

    def execute(self):
        print("Hello, World!")

Component Parsers

The ComponentParser class extends FunctionParser to allow generating components as serialized Kubeflow Pipelines (KFP).

Example usage:

from ml_orchestrator.comp_parser import ComponentParser
from ml_orchestrator.meta_comp import MetaComponent

class MyComponent(MetaComponent):
    @property
    def env(self):
        return ...  # Define environment parameters