kfx 0.1.0a1

Extensions for kubeflow pipeline sdk.

kfx

kfx is a python package with the namespace kfx. Currently, it provides the following sub-packages: - kfx.lib.vis: Data models and helpers to generate ui metadata object for rendering vis in kubeflow pipeline UI. - kfx.lib.utils: Helpers to extend kubeflow pipeline tasks/containerOps.

Documentation: https://kfx.readthedocs.io. Repo: https://github.com/e2fyi/kfx

Quick start

Installation

pip install kfx

Usage

Generating ui metadata artifacts for kubeflow pipeline UI to render visualizations.

import kfx.lib.vis as kfxvis
import kfx.lib.utils as kfxutils


@kfp.components.func_to_container_op
def test_op(
    mlpipeline_ui_metadata_file: OutputTextFile(str), markdown_file: OutputTextFile(str)
):

    import kfx.lib.utils as kfxutils
    import kfx.lib.vis as kfxvis


    # note that artifact name is `markdown` instead of `markdown_file`
    # `_file` and `_path` suffix are removed.
    markdown_file.write("### hello world")
    markdown_src = kfxutils.get_artifact_uri("markdown")

    # creates the ui metadata object
    mlpipeline_ui_metadata = kfxvis.kfp_ui_metadata(
        [kfxvis.markdown(source=markdown_src)]
    )
    # note that artifact name is `mlpipeline-ui-metadata` and not
    # `mlpipeline_ui_metadata_file`.
    mlpipeline_ui_metadata_file.write(kfxvis.asjson(mlpipeline_ui_metadata))

    # prints the artifact uri that will be saved by kfp to the artifactory.
    print(mlpipeline_ui_metadata.outputs[0].source)


# helper to decorate the task so that `kfx.lib.utils.get_artifact_uri` can be
# used to infer the uri of the artifact.
helper = kfxutils.ArtifactLocationHelper(
    scheme="minio", bucket="mlpipeline", key_prefix="artifacts/"
)

@kfp.dsl.pipeline()
def test_pipeline():
    """Test pipeline."""

    op: kfp.dsl.ContainerOp = test_op()
    # setup the required image and env vars, so that `kfx.lib.utils.get_artifact_uri`
    # can be used to infer artifact uri.
    op.apply(helper.set_envs())

Developer guide

This project used:

• isort: to manage import order
• pylint: to manage general coding best practices
• flake8: to manage code complexity and coding best practices
• black: to manage formats and styles
• pydocstyle: to manage docstr style/format
• pytest/coverage: to manage unit tests and code coverage
• bandit: to find common security issues
• pyenv: to manage dev env: python version (3.6)
• pipenv: to manage dev env: python packages

Convention for unit tests are to suffix with _test and colocate with the actual python module - i.e. <module_name>_test.py.

The version of the package is read from version.txt - i.e. please update the appropriate semantic version (major -> breaking changes, minor -> new features, patch -> bug fix, postfix -> pre-release/post-release).

Makefile:

# autoformat codes with docformatter, isort, and black
make format

# check style, formats, and code complexity
make check

# check style, formats, code complexity, and run unit tests
make test

# test everything including building the package and check the sdist
make test-all

# run unit test only
make test-only

# generate and update the requirements.txt and requirements-dev.txt
make requirements

# generate the docs with sphinx and autoapi extension
make docs

# generate distributions
make dists

# publish to pypi with twine (twine must be configured)
make publish