kacpermuda commented on code in PR #65727:
URL: https://github.com/apache/airflow/pull/65727#discussion_r3160643532
##########
providers/openlineage/src/airflow/providers/openlineage/utils/utils.py:
##########
@@ -862,10 +878,81 @@ class TaskInstanceInfo(InfoJsonEncodable):
class AssetInfo(InfoJsonEncodable):
- """Defines encoding Airflow Asset object to JSON."""
+ """
+ Defines encoding Airflow Asset object to JSON.
+
+ Output shape: ``{"uri", "extra", "type"[, "source_alias"]}``.
+ ``type`` distinguishes how the asset showed up on the task:
+
+ * ``"asset"`` - statically declared on ``task.outlets`` / ``task.inlets``.
+ * ``"asset_event"`` — emitted dynamically via
``outlet_events[asset].extra``.
+ * ``"asset_event_from_alias"`` — resolved at runtime from an ``AssetAlias``
+ via ``outlet_events[alias].add(asset, ...)``; ``source_alias`` then
carries the alias name.
+
+ ``AssetAlias`` are intentionally NOT included - they are just helpers with
no URI of their own.
+ """
includes = ["uri", "extra"]
+ def __init__(self, obj, asset_type: str = "asset", source_alias: str |
None = None):
+ self.type = asset_type
+ self.source_alias = source_alias
+ super().__init__(obj)
+
+ def _extend_fields(self) -> None:
+ if self.type:
+ self._fields.append("type")
+ if self.source_alias:
+ self._fields.append("source_alias")
+
+
+def get_runtime_outlet_assets(task_instance) -> list[tuple[Asset, AssetAlias |
None]]:
+ """
+ Return ``(asset, alias)`` pairs for every runtime-attached outlet.
+
+ * ``outlet_events[AssetAlias].add(asset, ...)`` → ``(asset, alias)``.
+ * ``outlet_events[Asset].extra = {...}`` → ``(asset, None)``.
+
+ Empty list on AF2 or when ``task_instance`` is not an AF3
``RuntimeTaskInstance``
+ (scheduler / API-server path has no usable worker template context).
+ """
+ if not AIRFLOW_V_3_0_PLUS:
+ return []
+ try:
+ from airflow.sdk.execution_time.task_runner import RuntimeTaskInstance
+ except ImportError:
+ return []
+ if not isinstance(task_instance, RuntimeTaskInstance):
+ return []
+ if not hasattr(task_instance, "get_template_context"):
+ return []
+ outlet_events = task_instance.get_template_context().get("outlet_events")
+ if outlet_events is None:
+ return []
+ results: list[tuple[Asset, AssetAlias | None]] = []
+ for key in outlet_events:
+ accessor = outlet_events[key]
+ if isinstance(key, AssetAlias):
+ for event in getattr(accessor, "asset_alias_events", []) or []:
+ # ``dest_asset_extra`` was introduced in Airflow 3.1.4 ; on
+ # earlier versions the alias event carries only ``extra``.
+ dest_asset_extra = getattr(event, "dest_asset_extra", None) or
{}
+ results.append(
+ (
+ Asset(
+ name=event.dest_asset_key.name,
+ uri=event.dest_asset_key.uri,
+ # Per-event extra wins over asset-level extra.
Review Comment:
Removed !
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]