PipelineExecutor¶
Module: flowerpower.pipeline.executor.PipelineExecutor
The PipelineExecutor handles pipeline execution with comprehensive parameter handling. It is responsible for merging runtime parameters with pipeline defaults and delegating to Hamilton's execution engine.
Initialization¶
| Parameter | Type | Description |
|---|---|---|
config_manager |
PipelineConfigManager |
Configuration manager for pipeline configs. |
registry |
PipelineRegistry |
Registry for accessing pipeline objects. |
project_context |
Any \| None |
Optional project context for execution. |
Methods¶
run¶
Execute a pipeline synchronously and return its results. This is the main method used internally by PipelineManager.run() and FlowerPowerProject.run().
| Parameter | Type | Description | Default |
|---|---|---|---|
name |
str |
Name of the pipeline. | — |
run_config |
RunConfig \| None |
Structured execution configuration. | None |
**kwargs |
Any |
Runtime overrides. | — |
Returns: dict[str, Any] — mapping output variables to computed values.
Raises:
ValueError: if configuration is invalid.ImportError: if the pipeline module cannot be imported.RuntimeError: if execution fails.
run_async¶
Execute a pipeline asynchronously using Hamilton's async driver.
Tip
In most code you will call manager.run() or manager.run_async() rather than invoking the executor directly.