convert_to_kfp

Export a pipeline to an Kubeflow YAML file

Convert and save a Link pipeline from Python code to an Kubeflow YAML file.

📘

This is possible after all components of the pipeline have been successfully executed.

pipeline.convert_to_kfp(file="str")

• Parameters
  • file(str): The path and name of the Kubeflow yaml file you wish to convert and save
• Troubleshooting
  • RuntimeError: Please try again after running the pipeline successfully
    • This occurs when conversion to YAML encounters a problem during pipeline execution.
  • NameError: It must not be blank
    • This occurs when the file name is blank.
  • NameError: It must not include [: * " ? < > |]
    • This occurs when : * " ? < > | is included in file.
  • NameError: The file name must end with .yaml or .yml
    • This occurs when the name of file does not end with .yaml or .yml.

Example

Python

from mrx_link.sdk.utils import *

# Code cell
code1 = """
x = 1
"""
code2 = """
print(f"{x=}, {y=}")
"""

if __name__ == "__main__":
    # Create components
    component1 = create_link_component(identifier="111-1", name="test", code=code1)
    component2 = create_link_component(identifier="111-2", name="test2", code=code2)
    components = [component1, component2]

    # Create edges
    edge1 = create_link_edge(parent_id="111-1", child_id="111-2")
    edges = [edge1]

    # Create pipeline parameters
    parameter1 = create_link_parameter(name="x", value="123")
    parameter2 = create_link_parameter(name="y", value="baregasdv")

    parameters = [parameter1, parameter2]

    # Create pipeline
    pipeline = create_link_pipeline(components=components, parameters=parameters, edges=edges)

    # Run the pipeline
    pipeline.execute_all()

    # Convert the pipeline to `sample.yaml` file
    pipeline.convert_to_kfp(file="sample.yaml")