* Use google.protobuf.Value in v2 for passing parameters.
* retest samples.
* Fix tests.
* Update release, more cleanup.
* Use github.com/kubeflow/pipelines/api from same repo.
* Run go mod tidy
* chore: go mod tidy
* fix v2 compile error and clean up unused code
* pr comments.
* update goldens
* Fix metadata recording.
* Update kfp mlmd client.
* fix test again
* another try.
* chore: migrate v2 DAG driver input parameters to protobuf.Value + small refactorings
* fix v2 launcher + clean up
* fix a compile error
* fix a few more tests
* fix number parsing
* clean up
* disable cache_v2 test.
Co-authored-by: Yuan Gong <gongyuan94@gmail.com>
* fix the issue:
the 'mlpipeline-ui-metadata','mlpipeline-metrics' path is replaced by the default path as '_components._generate_output_file_name(output.name)'
* reformat code
* Update _container_op.py
* Update RELEASE.md
* add description optionally
* add release notes
* add new line
* capture if description is not accepted by server
* add todo comment
* add better description api missing exception
* add disbale/delete run (job) functions to sdk client
* change run to job to accuratly reflect api
* add release note for change
* align docstrings with style guide requirements
* formatting
Co-authored-by: Chen Sun <chensun@users.noreply.github.com>
* Try alternatives for obtaining pip when `ensurepip` does not
exist in container.
This PR also fixes the tests that have golden files. Developers can
update golden files by setting UPDATE_GOLDENS=True before running
tests. E.g., run the following from sdk/python directory:
UPDATE_GOLDENS=True python3 -m unittest discover --verbose --pattern '*test*.py' --top-level-directory .
* Update RELEASE.
* Require base and target images in built containers to be unique.
Also add a test to verify, and fix formatting in cli/components.py.
* Update release notes.
* fix formatting.
* Refactor and move all v2 related code to under the v2 namespace.
Most of the changes are around imports and restructuring of the
codebase. While it looks like a lot of code was added, most of the code
already existed and was simply moved or copied over to v2. The only
exceptions are:
- under kfp/v2/components/component_factory.py: some helper functions
were copied with simplification from _python_op.py
- we no longer strip the `_path` suffix in v2 components.
Note: there is still some duplication of code (particularly between
component_factory.py and _python_op.py), but it's ok for now since we
intend to replace some of this with v2 ComponentSpec + BaseComponent.
* Update setup.py.
* update tests.
* revert accidental change of gcpc
* Fix component entrypoint.
* Update goldens.
* fix tests.
* fix merge conflict.
* revert gcpc change.
* fix tests.
* fix tests.
* Add type aliases for moved files.
* merge and update goldens.
* Add a CLI command to help containerize and build v2 components.
* Flesh out the CLI for Docker.
* update comments.
* Support multiple component files in the same directory.
* address pr comments.
* Let user specify just the directory containing components.
The CLI will search for components in all Python files by default. It
can also search for a specific filepattern (as supported by
pathlib.Path objects).
Also add a bunch of tests.
* Clean up.
* fix merge error.
* Fix docker import error for tests.
* Update release notes.
* address PR comments.
* fix(sdk): sanitize op name. Fix#6433
* Add release note: Fix the error that kfp v1 compiler failed to provide unique name for ops of the same component.
* Adds URI template support for ComponentStore
* Reverts some autoformatting
* Reverts some more autoformatting
* Addresses code review comments:
* order of constructor args preserved for compatibility
* added docstrings
* added explicit uritemplate dependency
* Adds description of change
* Adds missing #
* bump kfp-pipeline-spec to 0.1.11, tests for local runner will fail with kfp-pipeline-spec 0.1.9
* Local client supports additional docker options
* Format _local_client.py using yapf
* Add release note that local runner supports additional docker options
Chen Sun