apache-airflow-providers-openlineage 1.9.0rc1__tar.gz → 1.9.0rc2__tar.gz

{apache_airflow_providers_openlineage-1.9.0rc1 → apache_airflow_providers_openlineage-1.9.0rc2}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: apache-airflow-providers-openlineage
-Version: 1.9.0rc1
+Version: 1.9.0rc2
 Summary: Provider package apache-airflow-providers-openlineage for Apache Airflow
 Keywords: airflow-provider,openlineage,airflow,integration
 Author-email: Apache Software Foundation <dev@airflow.apache.org>
@@ -80,7 +80,7 @@ Provides-Extra: common.sql
 
 Package ``apache-airflow-providers-openlineage``
 
-Release: ``1.9.0.rc1``
+Release: ``1.9.0.rc2``
 
 
 `OpenLineage <https://openlineage.io/>`__

{apache_airflow_providers_openlineage-1.9.0rc1 → apache_airflow_providers_openlineage-1.9.0rc2}/README.rst

@@ -42,7 +42,7 @@
 
 Package ``apache-airflow-providers-openlineage``
 
-Release: ``1.9.0.rc1``
+Release: ``1.9.0.rc2``
 
 
 `OpenLineage <https://openlineage.io/>`__

{apache_airflow_providers_openlineage-1.9.0rc1 → apache_airflow_providers_openlineage-1.9.0rc2}/airflow/providers/openlineage/extractors/base.py

@@ -41,7 +41,8 @@ class OperatorLineage:
 
 
 class BaseExtractor(ABC, LoggingMixin):
-    """Abstract base extractor class.
+    """
+    Abstract base extractor class.
 
     This is used mostly to maintain support for custom extractors.
     """
@@ -55,7 +56,8 @@ class BaseExtractor(ABC, LoggingMixin):
     @classmethod
     @abstractmethod
     def get_operator_classnames(cls) -> list[str]:
-        """Get a list of operators that extractor works for.
+        """
+        Get a list of operators that extractor works for.
 
         This is an abstract method that subclasses should implement. There are
         operators that work very similarly and one extractor can cover.
@@ -77,7 +79,8 @@ class DefaultExtractor(BaseExtractor):
 
     @classmethod
     def get_operator_classnames(cls) -> list[str]:
-        """Assign this extractor to *no* operators.
+        """
+        Assign this extractor to *no* operators.
 
         Default extractor is chosen not on the classname basis, but
         by existence of get_openlineage_facets method on operator.

{apache_airflow_providers_openlineage-1.9.0rc1 → apache_airflow_providers_openlineage-1.9.0rc2}/airflow/providers/openlineage/plugins/adapter.py

@@ -139,7 +139,8 @@ class OpenLineageAdapter(LoggingMixin):
         )
 
     def emit(self, event: RunEvent):
-        """Emit OpenLineage event.
+        """
+        Emit OpenLineage event.
 
         :param event: Event to be emitted.
         :return: Redacted Event.
@@ -295,11 +296,17 @@ class OpenLineageAdapter(LoggingMixin):
         """
         error_facet = {}
         if error:
-            if isinstance(error, BaseException):
+            stack_trace = None
+            if isinstance(error, BaseException) and error.__traceback__:
                 import traceback
 
-                error = "\\n".join(traceback.format_exception(type(error), error, error.__traceback__))
-            error_facet = {"errorMessage": ErrorMessageRunFacet(message=error, programmingLanguage="python")}
+                stack_trace = "\\n".join(traceback.format_exception(type(error), error, error.__traceback__))
+
+            error_facet = {
+                "errorMessage": ErrorMessageRunFacet(
+                    message=str(error), programmingLanguage="python", stackTrace=stack_trace
+                )
+            }
 
         event = RunEvent(
             eventType=RunState.FAIL,

{apache_airflow_providers_openlineage-1.9.0rc1 → apache_airflow_providers_openlineage-1.9.0rc2}/airflow/providers/openlineage/plugins/facets.py

@@ -102,7 +102,8 @@ class AirflowRunFacet(BaseFacet):
 
 @define(slots=False)
 class UnknownOperatorInstance(RedactMixin):
-    """Describes an unknown operator.
+    """
+    Describes an unknown operator.
 
     This specifies the (class) name of the operator and its properties.
     """

{apache_airflow_providers_openlineage-1.9.0rc1 → apache_airflow_providers_openlineage-1.9.0rc2}/airflow/providers/openlineage/plugins/listener.py

@@ -367,7 +367,7 @@ class OpenLineageListener:
             except BaseException:
                 # Kill the process directly.
                 self._terminate_with_wait(process)
-            self.log.warning("Process with pid %s finished - parent", pid)
+            self.log.debug("Process with pid %s finished - parent", pid)
         else:
             setproctitle(getproctitle() + " - OpenLineage - " + callable_name)
             configure_orm(disable_connection_pool=True)
@@ -381,7 +381,7 @@ class OpenLineageListener:
         if not self._executor:
             self._executor = ProcessPoolExecutor(
                 max_workers=conf.dag_state_change_process_pool_size(),
-                initializer=_executor_initializer(),
+                initializer=_executor_initializer,
             )
         return self._executor
 

{apache_airflow_providers_openlineage-1.9.0rc1 → apache_airflow_providers_openlineage-1.9.0rc2}/airflow/providers/openlineage/sqlparser.py

@@ -118,7 +118,8 @@ def from_table_meta(
 
 
 class SQLParser(LoggingMixin):
-    """Interface for openlineage-sql.
+    """
+    Interface for openlineage-sql.
 
     :param dialect: dialect specific to the database
     :param default_schema: schema applied to each table with no schema parsed
@@ -244,7 +245,8 @@ class SQLParser(LoggingMixin):
         sqlalchemy_engine: Engine | None = None,
         use_connection: bool = True,
     ) -> OperatorLineage:
-        """Parse SQL statement(s) and generate OpenLineage metadata.
+        """
+        Parse SQL statement(s) and generate OpenLineage metadata.
 
         Generated OpenLineage metadata contains:
 

{apache_airflow_providers_openlineage-1.9.0rc1 → apache_airflow_providers_openlineage-1.9.0rc2}/airflow/providers/openlineage/utils/selective_enable.py

@@ -32,7 +32,8 @@ log = logging.getLogger(__name__)
 
 
 def enable_lineage(obj: T) -> T:
-    """Set selective enable OpenLineage parameter to True.
+    """
+    Set selective enable OpenLineage parameter to True.
 
     The method also propagates param to tasks if the object is DAG.
     """
@@ -48,7 +49,8 @@ def enable_lineage(obj: T) -> T:
 
 
 def disable_lineage(obj: T) -> T:
-    """Set selective enable OpenLineage parameter to False.
+    """
+    Set selective enable OpenLineage parameter to False.
 
     The method also propagates param to tasks if the object is DAG.
     """
@@ -73,7 +75,8 @@ def is_task_lineage_enabled(task: Operator) -> bool:
 
 
 def is_dag_lineage_enabled(dag: DAG) -> bool:
-    """Check if DAG is selectively enabled to emit OpenLineage events.
+    """
+    Check if DAG is selectively enabled to emit OpenLineage events.
 
     The method also checks if selective enable parameter is set to True
     or if any of the tasks in DAG is selectively enabled.

{apache_airflow_providers_openlineage-1.9.0rc1 → apache_airflow_providers_openlineage-1.9.0rc2}/airflow/providers/openlineage/utils/sql.py

@@ -85,7 +85,8 @@ def get_table_schemas(
     in_query: str | None,
     out_query: str | None,
 ) -> tuple[list[Dataset], list[Dataset]]:
-    """Query database for table schemas.
+    """
+    Query database for table schemas.
 
     Uses provided hook. Responsibility to provide queries for this function is on particular extractors.
     If query for input or output table isn't provided, the query is skipped.
@@ -111,7 +112,8 @@ def get_table_schemas(
 
 
 def parse_query_result(cursor) -> list[TableSchema]:
-    """Fetch results from DB-API 2.0 cursor and creates list of table schemas.
+    """
+    Fetch results from DB-API 2.0 cursor and creates list of table schemas.
 
     For each row it creates :class:`TableSchema`.
     """

{apache_airflow_providers_openlineage-1.9.0rc1 → apache_airflow_providers_openlineage-1.9.0rc2}/airflow/providers/openlineage/utils/utils.py

@@ -150,7 +150,7 @@ class InfoJsonEncodable(dict):
             return value.isoformat()
         if isinstance(value, datetime.timedelta):
             return f"{value.total_seconds()} seconds"
-        if isinstance(value, (set, list, tuple)):
+        if isinstance(value, (set, tuple)):
             return str(list(value))
         return value
 
@@ -214,6 +214,12 @@ class TaskInstanceInfo(InfoJsonEncodable):
     }
 
 
+class DatasetInfo(InfoJsonEncodable):
+    """Defines encoding Airflow Dataset object to JSON."""
+
+    includes = ["uri", "extra"]
+
+
 class TaskInfo(InfoJsonEncodable):
     """Defines encoding BaseOperator/AbstractOperator object to JSON."""
 
@@ -242,6 +248,9 @@ class TaskInfo(InfoJsonEncodable):
         "run_as_user",
         "sla",
         "task_id",
+        "trigger_dag_id",
+        "external_dag_id",
+        "external_task_id",
         "trigger_rule",
         "upstream_task_ids",
         "wait_for_downstream",
@@ -255,6 +264,8 @@ class TaskInfo(InfoJsonEncodable):
             if hasattr(task, "task_group") and getattr(task.task_group, "_group_id", None)
             else None
         ),
+        "inlets": lambda task: [DatasetInfo(inlet) for inlet in task.inlets],
+        "outlets": lambda task: [DatasetInfo(outlet) for outlet in task.outlets],
     }
 
 

{apache_airflow_providers_openlineage-1.9.0rc1 → apache_airflow_providers_openlineage-1.9.0rc2}/pyproject.toml

@@ -28,7 +28,7 @@ build-backend = "flit_core.buildapi"
 
 [project]
 name = "apache-airflow-providers-openlineage"
-version = "1.9.0.rc1"
+version = "1.9.0.rc2"
 description = "Provider package apache-airflow-providers-openlineage for Apache Airflow"
 readme = "README.rst"
 authors = [