ComputeActionOutputPath
Constructs the full output directory for a task action, including a deterministic shard prefix for storage hot-spot avoidance and an attempt segment to isolate retries.
def ComputeActionOutputPath(
ctx: context.Context,
namespace: string,
name: string,
runOutputBase: string,
actionName: string,
attempt: uint32
) - > storage.DataReference, error
Constructs the full output directory for a task action, including a sharded path for hot-spot avoidance and an attempt segment for retries. This function is used to generate a unique and optimized storage path for each action's output.
Parameters
| Name | Type | Description |
|---|---|---|
| ctx | context.Context | The context for the operation, used for cancellation and timeouts. |
| namespace | string | The namespace of the TaskAction, used to deterministically derive the shard prefix. |
| name | string | The name of the TaskAction, used to deterministically derive the shard prefix. |
| runOutputBase | string | The base output path for the run, which will be augmented with sharding, action name, and attempt number. |
| actionName | string | The name of the specific action within the task, used as a segment in the output path. |
| attempt | uint32 | The attempt number for the action, used to isolate each retry in the output path. |
Returns
| Type | Description |
|---|---|
storage.DataReference, error | The full output path as a storage.DataReference and an error if parsing or sharding fails. |