This release adds support for pipeline steps being run in non-linear order, specifically as a DAG. If you have four steps defined, you can return a dictionary of keys and list values, where each key is a step, and each list is the set of steps that depend on the key. Example:
# define the pipeline
@pipeline(
name='test-pipeline',
version='0.0.1',
description='a test tinypipeline',
)
def pipe():
# run the steps in the defined order of the graph
return {
step_one: [step_two, step_four],
step_two: [step_three, step_four]
}
What's Changed
- Add support for nonlinear pipelines by @freddiev4 in #3
Full Changelog: 0.2.0...0.3.0