Add PipelineDecorator param section specification (#268)

This commit is contained in:
pollfly 2022-06-19 10:46:03 +03:00 committed by GitHub
parent 06cd0e4894
commit b62e540af3
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23

View File

@ -15,7 +15,10 @@ Using the [`@PipelineDecorator.pipeline`](../references/sdk/automation_controlle
decorator transforms the function which implements your pipeline's execution logic to a ClearML pipeline controller decorator transforms the function which implements your pipeline's execution logic to a ClearML pipeline controller
```python ```python
@PipelineDecorator.pipeline(name='pipeline', project='examples', version='0.1') @PipelineDecorator.pipeline(
name='pipeline', project='examples', version='0.1',
args_map={'General':['pickle_url'], 'Mock':['mock_parameter']}
)
def main(pickle_url, mock_parameter='mock'): def main(pickle_url, mock_parameter='mock'):
data_frame = step_one(pickle_url) data_frame = step_one(pickle_url)
X_train, X_test, y_train, y_test = step_two(data_frame) X_train, X_test, y_train, y_test = step_two(data_frame)
@ -31,6 +34,10 @@ def main(pickle_url, mock_parameter='mock'):
* `version` - Numbered version string (e.g. 1.2.3). If `auto_version_bump` is set to `True`, the version number is * `version` - Numbered version string (e.g. 1.2.3). If `auto_version_bump` is set to `True`, the version number is
automatically bumped if the same version already exists and the pipeline code has changed automatically bumped if the same version already exists and the pipeline code has changed
* `default_queue` - The default [ClearML Queue](../fundamentals/agents_and_queues.md#what-is-a-queue) in which to enqueue all pipeline steps (unless otherwise specified in the pipeline step). * `default_queue` - The default [ClearML Queue](../fundamentals/agents_and_queues.md#what-is-a-queue) in which to enqueue all pipeline steps (unless otherwise specified in the pipeline step).
* `args_map` - Map arguments to their [configuration section](../fundamentals/hyperparameters.md#webapp-interface) in
the following format: `{'section_name':['param_name']]}`. For example, the pipeline in the code above will store the
`pickle_url` parameter in the `General` section and `mock_parameter` in the `Mock` section. By default, arguments will
be stored in the `Args` section.
* `pool_frequency` - The pooling frequency (in minutes) for monitoring experiments / states. * `pool_frequency` - The pooling frequency (in minutes) for monitoring experiments / states.
* `add_pipeline_tags` - If `True`, add `pipe: <pipeline_task_id>` tag to all steps (Tasks) created by this pipeline * `add_pipeline_tags` - If `True`, add `pipe: <pipeline_task_id>` tag to all steps (Tasks) created by this pipeline
(this is useful to create better visibility in projects with multiple pipelines, and for easy selection) (default: (this is useful to create better visibility in projects with multiple pipelines, and for easy selection) (default: