openlineage-python 1.33.0__tar.gz → 1.34.0__tar.gz

{openlineage_python-1.33.0 → openlineage_python-1.34.0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: openlineage-python
-Version: 1.33.0
+Version: 1.34.0
 Summary: OpenLineage Python Client
 Author-email: OpenLineage <info@openlineage.io>
 Keywords: openlineage

{openlineage_python-1.33.0 → openlineage_python-1.34.0}/openlineage/client/client.py

@@ -2,6 +2,7 @@
 # SPDX-License-Identifier: Apache-2.0
 from __future__ import annotations
 
+import contextlib
 import json
 import logging
 import os
@@ -10,19 +11,9 @@ from typing import TYPE_CHECKING, Any, TypeVar, Union, cast
 
 import attr
 import yaml
-from openlineage.client.filter import Filter, FilterConfig, create_filter
-from openlineage.client.serde import Serde
-from openlineage.client.tags import TagsConfig
-from openlineage.client.utils import deep_merge_dicts
-
-if TYPE_CHECKING:
-    from requests import Session
-    from requests.adapters import HTTPAdapter
-
-import contextlib
-
 from openlineage.client import event_v2
 from openlineage.client.facets import FacetsConfig
+from openlineage.client.filter import Filter, FilterConfig, create_filter
 from openlineage.client.generated.environment_variables_run import (
     EnvironmentVariable,
     EnvironmentVariablesRunFacet,
@@ -30,6 +21,8 @@ from openlineage.client.generated.environment_variables_run import (
 from openlineage.client.generated.tags_job import TagsJobFacet, TagsJobFacetFields
 from openlineage.client.generated.tags_run import TagsRunFacet, TagsRunFacetFields
 from openlineage.client.run import DatasetEvent, JobEvent, RunEvent
+from openlineage.client.serde import Serde
+from openlineage.client.tags import TagsConfig
 from openlineage.client.transport import (
     Transport,
     TransportFactory,
@@ -37,26 +30,32 @@ from openlineage.client.transport import (
 )
 from openlineage.client.transport.http import HttpConfig, HttpTransport, create_token_provider
 from openlineage.client.transport.noop import NoopConfig, NoopTransport
+from openlineage.client.utils import deep_merge_dicts
+
+if TYPE_CHECKING:
+    from requests import Session
+    from requests.adapters import HTTPAdapter
+
 
 Event_v1 = Union[RunEvent, DatasetEvent, JobEvent]
 Event_v2 = Union[event_v2.RunEvent, event_v2.DatasetEvent, event_v2.JobEvent]
 Event = Union[Event_v1, Event_v2]
 
 
-@attr.s
+@attr.define
 class OpenLineageClientOptions:
-    timeout: float = attr.ib(default=5.0)
-    verify: bool = attr.ib(default=True)
-    api_key: str = attr.ib(default=None)
-    adapter: HTTPAdapter = attr.ib(default=None)
+    timeout: float = 5.0
+    verify: bool = True
+    api_key: str | None = None
+    adapter: HTTPAdapter | None = None
 
 
-@attr.s
+@attr.define
 class OpenLineageConfig:
-    transport: dict[str, Any] | None = attr.ib(factory=dict)
-    facets: FacetsConfig = attr.ib(factory=FacetsConfig)
-    filters: list[FilterConfig] = attr.ib(factory=list)
-    tags: TagsConfig = attr.ib(factory=TagsConfig)
+    transport: dict[str, Any] | None = attr.field(factory=dict)
+    facets: FacetsConfig = attr.field(factory=FacetsConfig)
+    filters: list[FilterConfig] = attr.field(factory=list)
+    tags: TagsConfig = attr.field(factory=TagsConfig)
 
     @classmethod
     def from_dict(cls, params: dict[str, Any]) -> OpenLineageConfig:

{openlineage_python-1.33.0 → openlineage_python-1.34.0}/openlineage/client/constants.py

@@ -2,7 +2,7 @@
 # SPDX-License-Identifier: Apache-2.0
 from __future__ import annotations
 
-__version__ = "1.33.0"
+__version__ = "1.34.0"
 
 DEFAULT_TIMEOUT_MS = 5000
 DEFAULT_NAMESPACE_NAME = "default"

{openlineage_python-1.33.0 → openlineage_python-1.34.0}/openlineage/client/facets.py

@@ -5,6 +5,6 @@ from __future__ import annotations
 import attr
 
 
-@attr.s
+@attr.define
 class FacetsConfig:
-    environment_variables: list[str] = attr.ib(factory=list)
+    environment_variables: list[str] = attr.field(factory=list)

{openlineage_python-1.33.0 → openlineage_python-1.34.0}/openlineage/client/filter.py

@@ -14,11 +14,11 @@ log = logging.getLogger(__name__)
 RunEventType = typing.Union[RunEvent, RunEvent_v2]
 
 
-@attr.s
+@attr.define
 class FilterConfig:
-    type: str | None = attr.ib(default=None)
-    match: str | None = attr.ib(default=None)
-    regex: str | None = attr.ib(default=None)
+    type: str | None = None
+    match: str | None = None
+    regex: str | None = None
 
 
 class Filter:

{openlineage_python-1.33.0 → openlineage_python-1.34.0}/openlineage/client/generated/base.py

@@ -92,7 +92,7 @@ class BaseFacet(RedactMixin):
 
         new_class = attr.make_class(
             self.__class__.__name__,
-            {k: attr.field() for k in kwargs if k not in current_attrs},
+            {k: attr.field(default=None) for k in kwargs if k not in current_attrs},
             bases=(self.__class__,),
         )
         new_class.__module__ = self.__class__.__module__

{openlineage_python-1.33.0 → openlineage_python-1.34.0}/openlineage/client/generated/column_lineage_dataset.py

@@ -44,7 +44,7 @@ class Fields(RedactMixin):
 
         new_class = attr.make_class(
             self.__class__.__name__,
-            {k: attr.field() for k in kwargs if k not in current_attrs},
+            {k: attr.field(default=None) for k in kwargs if k not in current_attrs},
             bases=(self.__class__,),
         )
         new_class.__module__ = self.__class__.__module__
@@ -81,7 +81,7 @@ class InputField(RedactMixin):
 
         new_class = attr.make_class(
             self.__class__.__name__,
-            {k: attr.field() for k in kwargs if k not in current_attrs},
+            {k: attr.field(default=None) for k in kwargs if k not in current_attrs},
             bases=(self.__class__,),
         )
         new_class.__module__ = self.__class__.__module__
@@ -122,7 +122,7 @@ class Transformation(RedactMixin):
 
         new_class = attr.make_class(
             self.__class__.__name__,
-            {k: attr.field() for k in kwargs if k not in current_attrs},
+            {k: attr.field(default=None) for k in kwargs if k not in current_attrs},
             bases=(self.__class__,),
         )
         new_class.__module__ = self.__class__.__module__

{openlineage_python-1.33.0 → openlineage_python-1.34.0}/openlineage/client/generated/documentation_dataset.py

@@ -12,6 +12,9 @@ class DocumentationDatasetFacet(DatasetFacet):
     description: str
     """The description of the dataset."""
 
+    contentType: str | None = attr.field(default=None)  # noqa: N815
+    """MIME type of the description field content."""
+
     @staticmethod
     def _get_schema() -> str:
-        return "https://openlineage.io/spec/facets/1-0-1/DocumentationDatasetFacet.json#/$defs/DocumentationDatasetFacet"
+        return "https://openlineage.io/spec/facets/1-1-0/DocumentationDatasetFacet.json#/$defs/DocumentationDatasetFacet"

{openlineage_python-1.33.0 → openlineage_python-1.34.0}/openlineage/client/generated/documentation_job.py

@@ -12,8 +12,11 @@ class DocumentationJobFacet(JobFacet):
     description: str
     """The description of the job."""
 
+    contentType: str | None = attr.field(default=None)  # noqa: N815
+    """MIME type of the description field content."""
+
     @staticmethod
     def _get_schema() -> str:
         return (
-            "https://openlineage.io/spec/facets/1-0-1/DocumentationJobFacet.json#/$defs/DocumentationJobFacet"
+            "https://openlineage.io/spec/facets/1-1-0/DocumentationJobFacet.json#/$defs/DocumentationJobFacet"
         )

{openlineage_python-1.33.0 → openlineage_python-1.34.0}/openlineage/client/generator/templates/dataclass.jinja2

@@ -157,7 +157,7 @@ class {{ class_name }}:
 
         new_class = attr.make_class(
                 self.__class__.__name__,
-                {k: attr.field() for k in kwargs if k not in current_attrs},
+                {k: attr.field(default=None) for k in kwargs if k not in current_attrs},
                 bases=(self.__class__,),
             )
         new_class.__module__ = self.__class__.__module__

{openlineage_python-1.33.0 → openlineage_python-1.34.0}/openlineage/client/tags.py

@@ -11,7 +11,7 @@ if TYPE_CHECKING:
     from openlineage.client.generated.tags_run import TagsRunFacetFields
 
 
-@attr.s
+@attr.define
 class TagsConfig:
-    job: list[TagsJobFacetFields] = attr.ib(factory=list)
-    run: list[TagsRunFacetFields] = attr.ib(factory=list)
+    job: list[TagsJobFacetFields] = attr.field(factory=list)
+    run: list[TagsRunFacetFields] = attr.field(factory=list)

{openlineage_python-1.33.0 → openlineage_python-1.34.0}/openlineage/client/transport/composite.py

@@ -17,7 +17,7 @@ if TYPE_CHECKING:
 log = logging.getLogger(__name__)
 
 
-@attr.s
+@attr.define
 class CompositeConfig(Config):
     """
     CompositeConfig is a configuration class for CompositeTransport.
@@ -36,8 +36,8 @@ class CompositeConfig(Config):
             transport will halt the emission process for subsequent transports.
     """
 
-    transports: list[dict[str, Any]] | dict[str, dict[str, Any]] = attr.ib()
-    continue_on_failure: bool = attr.ib(default=True)
+    transports: list[dict[str, Any]] | dict[str, dict[str, Any]]
+    continue_on_failure: bool = True
 
     @classmethod
     def from_dict(cls, params: dict[str, Any]) -> CompositeConfig:
@@ -86,6 +86,7 @@ class CompositeTransport(Transport):
             except Exception as e:  # noqa: BLE001
                 if self.config.continue_on_failure:
                     log.warning("Transport %s failed to emit event with error: %s", transport, e)
+                    log.debug("OpenLineage emission failure details:", exc_info=True)
                 else:
                     msg = f"Transport {transport} failed to emit event"
                     raise RuntimeError(msg) from e

{openlineage_python-1.33.0 → openlineage_python-1.34.0}/openlineage/client/transport/http.py

@@ -76,23 +76,23 @@ def get_session() -> Session:
     return Session()
 
 
-@attr.s
+@attr.define
 class HttpConfig(Config):
-    url: str = attr.ib()
-    endpoint: str = attr.ib(default="api/v1/lineage")
-    timeout: float = attr.ib(default=5.0)
+    url: str
+    endpoint: str = "api/v1/lineage"
+    timeout: float = 5.0
     # check TLS certificates
-    verify: bool = attr.ib(default=True)
-    auth: TokenProvider = attr.ib(factory=lambda: TokenProvider({}))
-    compression: HttpCompression | None = attr.ib(default=None)
+    verify: bool = True
+    auth: TokenProvider = attr.field(factory=lambda: TokenProvider({}))
+    compression: HttpCompression | None = None
     # not set by TransportFactory
-    session: Session | None = attr.ib(default=None)
+    session: Session | None = None
     # not set by TransportFactory
-    adapter: HTTPAdapter | None = attr.ib(default=None)
+    adapter: HTTPAdapter | None = None
     # custom headers support
-    custom_headers: dict[str, str] = attr.ib(factory=dict)
+    custom_headers: dict[str, str] = attr.field(factory=dict)
     # retry settings
-    retry: dict[str, Any] = attr.ib(
+    retry: dict[str, Any] = attr.field(
         default={
             "total": 5,
             "read": 5,

{openlineage_python-1.33.0 → openlineage_python-1.34.0}/openlineage/client/transport/kafka.py

@@ -24,21 +24,21 @@ log = logging.getLogger(__name__)
 _T = TypeVar("_T", bound="KafkaConfig")
 
 
-@attr.s
+@attr.define
 class KafkaConfig(Config):
     # Kafka producer config
     # https://docs.confluent.io/platform/current/clients/confluent-kafka-python/html/index.html#kafka-client-configuration
-    config: dict[str, str] = attr.ib()
+    config: dict[str, str]
 
     # Topic on which we should send messages
-    topic: str = attr.ib()
+    topic: str
 
     # Explicit key for Kafka producer
-    messageKey: str | None = attr.ib(default=None)  # noqa: N815
+    messageKey: str | None = None  # noqa: N815
 
     # Set to true if Kafka should flush after each event. The process that emits can be killed in
     # some cases - for example in Airflow integration, so flushing is desirable there.
-    flush: bool = attr.ib(default=True)
+    flush: bool = True
 
     @classmethod
     def from_dict(cls: type[_T], params: dict[str, Any]) -> _T:

{openlineage_python-1.33.0 → openlineage_python-1.34.0}/openlineage/client/transport/msk_iam.py

@@ -32,15 +32,15 @@ def _detect_running_region() -> None | str:
     return None
 
 
-@attr.s
+@attr.define
 class MSKIAMConfig(KafkaConfig):
     # MSK producer config
     # https://github.com/aws/aws-msk-iam-sasl-signer-python
 
-    region: str = attr.ib(default=None)
-    aws_profile: None | str = attr.ib(default=None)
-    role_arn: None | str = attr.ib(default=None)
-    aws_debug_creds: bool = attr.ib(default=False)
+    region: str | None = None
+    aws_profile: None | str = None
+    role_arn: None | str = None
+    aws_debug_creds: bool = False
 
 
 def _oauth_cb(config: MSKIAMConfig, *_: Any) -> tuple[str, float]:

{openlineage_python-1.33.0 → openlineage_python-1.34.0}/openlineage/client/transport/transform/transform.py

@@ -46,7 +46,7 @@ class EventTransformer:
         return f"<{self.__class__.__name__}({self.properties})>"
 
 
-@attr.s
+@attr.define
 class TransformConfig(Config):
     """Configuration class for transform transport.
 
@@ -54,9 +54,9 @@ class TransformConfig(Config):
     transformer class to use, and optional transformer-specific properties.
     """
 
-    transport: dict[str, Any] = attr.ib()
-    transformer_class: str = attr.ib()
-    transformer_properties: dict[str, Any] = attr.ib(default={})
+    transport: dict[str, Any]
+    transformer_class: str
+    transformer_properties: dict[str, Any] | None = None
 
     @classmethod
     def from_dict(cls, params: dict[str, Any]) -> TransformConfig:
@@ -113,7 +113,7 @@ class TransformTransport(Transport):
         if not inspect.isclass(transformer_class) or not issubclass(transformer_class, EventTransformer):
             msg = f"Transformer `{transformer_class}` has to be class, and subclass of EventTransformer"
             raise TypeError(msg)
-        self.transformer = transformer_class(properties=config.transformer_properties)
+        self.transformer = transformer_class(properties=config.transformer_properties or {})
         self.transport = get_default_factory().create(config.transport)
 
     def emit(self, event: Event) -> Any:
@@ -128,14 +128,17 @@ class TransformTransport(Transport):
         Returns:
             The result of the inner transport's emit call, or None if the event is dropped.
         """
-        log.debug("Event before transformation: %s", Serde.to_json(event))
-        event = copy.deepcopy(event)  # Copy to make sure the transformer does not modify original event
-
+        log.debug("Event before copy: %s", Serde.to_json(event))
+        try:
+            event = copy.deepcopy(event)  # Copy to make sure the transformer does not modify original event
+            log.debug("Event after copy, before transformation: %s", Serde.to_json(event))
+        except AttributeError as e:
+            msg = "Error serializing copy of event."
+            raise ValueError(msg) from e
         log.debug("Transforming OpenLineage event with %s", self.transformer)
         transformed = self.transformer.transform(event)
         if transformed is None:
             log.info("Transformed OpenLineage event is None. Skipping emission.")
             return None
-
         log.debug("Event after transformation: %s", Serde.to_json(transformed))
         return self.transport.emit(transformed)

{openlineage_python-1.33.0 → openlineage_python-1.34.0}/openlineage/client/transport/transport.py

@@ -29,7 +29,7 @@ if TYPE_CHECKING:
 _T = TypeVar("_T", bound="Config")
 
 
-@attr.s
+@attr.define
 class Config:
     @classmethod
     def from_dict(cls: type[_T], params: dict[str, Any]) -> _T:  # noqa: ARG003

{openlineage_python-1.33.0 → openlineage_python-1.34.0}/pyproject.toml

@@ -6,7 +6,7 @@ requires = [
 
 [project]
 name = "openlineage-python"
-version = "1.33.0"
+version = "1.34.0"
 description = "OpenLineage Python Client"
 readme = "README.md"
 keywords = [

{openlineage_python-1.33.0 → openlineage_python-1.34.0}/tests/test_events.py

@@ -103,25 +103,25 @@ def test_schema_field_default() -> None:
     )
 
 
-@attr.s
+@attr.define
 class NestedObject:
-    value: int | None = attr.ib(default=None)
+    value: int | None = None
 
 
-@attr.s
+@attr.define
 class NestingObject:
-    nested: list[NestedObject] = attr.ib()
-    optional: int | None = attr.ib(default=None)
+    nested: list[NestedObject]
+    optional: int | None = None
 
 
-@attr.s
+@attr.define
 class ListOfStrings:
-    values: list[str] = attr.ib()
+    values: list[str]
 
 
-@attr.s
+@attr.define
 class NestedListOfStrings:
-    nested: list[ListOfStrings] = attr.ib()
+    nested: list[ListOfStrings]
 
 
 def test_serde_list_of_strings() -> None:

{openlineage_python-1.33.0 → openlineage_python-1.34.0}/tests/test_facet.py

@@ -10,6 +10,7 @@ from unittest.mock import MagicMock
 import pytest
 from openlineage.client.client import OpenLineageClient
 from openlineage.client.facet import (
+    BaseFacet,
     ColumnLineageDatasetFacet,
     ColumnLineageDatasetFacetFieldsAdditional,
     ColumnLineageDatasetFacetFieldsAdditionalInputFields,
@@ -27,6 +28,7 @@ from openlineage.client.facet import (
     SymlinksDatasetFacetIdentifiers,
 )
 from openlineage.client.run import SCHEMA_URL, Dataset, Job, Run, RunEvent, RunState
+from openlineage.client.serde import Serde
 
 
 @pytest.fixture()
@@ -465,3 +467,19 @@ def test_job_type_job_facet(event: dict[str, Any]) -> None:
     expected_event["job"]["facets"]["jobType"] = job_type_facet
 
     assert expected_event == event_sent
+
+
+def test_facet_copy_serialization_base_facet():
+    facet = BaseFacet()
+    facet_copy = copy.deepcopy(facet)
+    assert Serde.to_json(facet) == Serde.to_json(facet_copy)
+
+
+def test_facet_copy_serialization_parent_run_facet():
+    facet = OwnershipJobFacet(
+        owners=[
+            OwnershipJobFacetOwners("some-owner", "some-owner-type"),
+        ],
+    )
+    facet_copy = copy.deepcopy(facet)
+    assert Serde.to_json(facet) == Serde.to_json(facet_copy)

{openlineage_python-1.33.0 → openlineage_python-1.34.0}/tests/test_facet_v2.py

@@ -2,10 +2,12 @@
 # SPDX-License-Identifier: Apache-2.0
 from __future__ import annotations
 
+import copy
 import json
 import os
 from unittest import mock
 
+import pytest
 from attr import asdict, define
 from openlineage.client.client import OpenLineageClient
 from openlineage.client.event_v2 import (
@@ -29,6 +31,7 @@ from openlineage.client.facet_v2 import (
     schema_dataset,
     set_producer,
 )
+from openlineage.client.serde import Serde
 
 
 def test_set_producer():
@@ -102,8 +105,6 @@ def test_custom_facet() -> None:
     assert expected_event == event_sent
 
 
-#
-# def test_full_core_event_serializes_properly(facet_mocker, event_mocker) -> None:
 def test_full_core_event_serializes_properly() -> None:
     session = mock.MagicMock()
     client = OpenLineageClient(url="http://example.com", session=session)
@@ -233,9 +234,10 @@ def test_with_additional_properties_works_with_attr_asdict():
 
     assert asdict(changed_facet) == {
         "_producer": "https://github.com/OpenLineage/OpenLineage/blob/v1-0-0/client",
-        "_schemaURL": "https://openlineage.io/spec/facets/1-0-1/DocumentationJobFacet.json#/$defs/DocumentationJobFacet",
+        "_schemaURL": "https://openlineage.io/spec/facets/1-1-0/DocumentationJobFacet.json#/$defs/DocumentationJobFacet",
         "_deleted": None,
         "description": "desc",
+        "contentType": None,
         "new_prop": "new_value",
     }
 
@@ -246,3 +248,46 @@ def test_with_additional_properties_isinstance_works():
 
     assert isinstance(changed_facet, documentation_job.DocumentationJobFacet)
     assert isinstance(changed_facet, BaseFacet)
+
+
+def test_facet_copy_serialization_base_facet():
+    facet = BaseFacet(producer="producer")
+    facet_copy = copy.deepcopy(facet)
+    assert Serde.to_json(facet) == Serde.to_json(facet_copy)
+
+
+def test_facet_copy_serialization_parent_run_facet():
+    facet = parent_run.ParentRunFacet(
+        run=parent_run.Run(runId="3bb703d1-09c1-4a42-8da5-35a0b3216072"),
+        job=parent_run.Job(namespace="default", name="parent_job_name"),
+        root=parent_run.Root(
+            run=parent_run.RootRun("3bb703d1-09c1-4a42-8da5-35a0b3216071"),
+            job=parent_run.RootJob(namespace="root_job_namespace", name="root_job_name"),
+        ),
+    )
+    facet_copy = copy.deepcopy(facet)
+    assert Serde.to_json(facet) == Serde.to_json(facet_copy)
+
+
+def test_custom_facet_copy_serialization_success():
+    @define
+    class SomeFacet(BaseFacet):
+        version: str
+
+    facet = SomeFacet(version="1")
+    facet_copy = copy.deepcopy(facet)
+    assert Serde.to_json(facet) == Serde.to_json(facet_copy)
+
+
+def test_custom_facet_copy_serialization_fails_when_mixing_attr_classes():
+    """This will fail as BaseFacet class uses attr.define and SomeFacet attr.s"""
+    import attr
+
+    @attr.s
+    class SomeFacet(BaseFacet):
+        version: str = attr.ib()
+
+    facet = SomeFacet(version="1")
+    facet_copy = copy.deepcopy(facet)
+    with pytest.raises(AttributeError, match="'SomeFacet' object has no attribute 'version'"):
+        Serde.to_json(facet_copy)

openlineage_python-1.34.0/tests/test_uuid.py

@@ -0,0 +1,102 @@
+# Copyright 2018-2025 contributors to the OpenLineage project
+# SPDX-License-Identifier: Apache-2.0
+from __future__ import annotations
+
+from datetime import datetime, timedelta, timezone
+
+import pytest
+from openlineage.client.uuid import generate_new_uuid, generate_static_uuid
+
+
+def test_generate_new_uuid_returns_uuidv7():
+    assert generate_new_uuid().version == 7  # noqa: PLR2004
+
+
+def test_generate_new_uuid_returns_different_result_on_each_call():
+    result1 = generate_new_uuid()
+    result2 = generate_new_uuid()
+    assert result1 != result2
+
+    instant = datetime.now()
+    result1 = generate_new_uuid(instant)
+    result2 = generate_new_uuid(instant)
+    assert result1 != result2
+
+
+def test_generate_new_uuid_result_is_increasing_with_instant_increment():
+    result1 = generate_new_uuid(datetime.now())
+    result2 = generate_new_uuid(datetime.now() + timedelta(seconds=0.001))
+    assert result1 != result2
+    assert result1 < result2
+
+
+def test_generate_new_uuid_returns_prefix_based_on_instant_milliseconds():
+    instant_milliseconds = datetime(2025, 5, 20, 10, 52, 33, 881000, tzinfo=timezone.utc)
+    instant_microseconds = datetime(2025, 5, 20, 10, 52, 33, 881863, tzinfo=timezone.utc)
+
+    uuid1 = generate_new_uuid(instant_milliseconds)
+    uuid2 = generate_new_uuid(instant_microseconds)
+
+    assert str(uuid1).startswith("0196ed52-e0d9-7")
+    assert str(uuid2).startswith("0196ed52-e0d9-7")
+
+
+def test_generate_static_uuid_returns_uuidv7():
+    assert generate_static_uuid(datetime.now(), b"some").version == 7  # noqa: PLR2004
+
+
+def test_generate_static_uuid_returns_same_result_for_same_input():
+    instant = datetime.now()
+    data = b"some"
+    assert generate_static_uuid(instant, data) == generate_static_uuid(instant, data)
+
+
+@pytest.mark.parametrize(
+    "data",
+    [
+        b"some",
+        b"other",
+    ],
+)
+def test_generate_static_uuid_result_prefix_depends_on_instant_milliseconds(data):
+    instant_milliseconds = datetime(2025, 5, 20, 10, 52, 33, 881000, tzinfo=timezone.utc)
+    instant_microseconds = datetime(2025, 5, 20, 10, 52, 33, 881863, tzinfo=timezone.utc)
+
+    uuid1 = generate_static_uuid(instant_milliseconds, data)
+    uuid2 = generate_static_uuid(instant_microseconds, data)
+
+    assert str(uuid1).startswith("0196ed52-e0d9-7")
+    assert str(uuid2).startswith("0196ed52-e0d9-7")
+
+
+@pytest.mark.parametrize(
+    ("data1", "data2"),
+    [
+        (b"some", b"other"),
+        (b"some", b"some"),
+    ],
+)
+def test_generate_static_uuid_result_is_increasing_with_instant_increment(data1, data2):
+    instant = datetime.now()
+    result1 = generate_static_uuid(instant, data1)
+    result2 = generate_static_uuid(instant + timedelta(seconds=0.001), data2)
+    assert result1 != result2
+    assert result1 < result2
+
+    # both timestamp and random parts are different
+    assert result1.bytes != result2.bytes
+    assert result1.bytes_le != result2.bytes_le
+
+
+def test_generate_static_uuid_result_is_different_for_different_data():
+    timestamp = datetime.now()
+    result1 = generate_static_uuid(timestamp, b"some")
+    result2 = generate_static_uuid(timestamp, b"other")
+    assert result1 != result2
+
+    # timestamp part is the same
+    assert int.from_bytes(result1.bytes, "big") & 0x0000 == int.from_bytes(result2.bytes, "big") & 0x0000
+
+    # random part is different
+    assert int.from_bytes(result1.bytes, "big") & 0xFFFF != int.from_bytes(result2.bytes, "big") & 0xFFFF
+    assert result1.bytes_le != result2.bytes_le

{openlineage_python-1.33.0 → openlineage_python-1.34.0}/tests/transform/test_transform.py

@@ -4,12 +4,15 @@ from __future__ import annotations
 
 import datetime
 import os
+import warnings
 from typing import TYPE_CHECKING
 from unittest.mock import MagicMock, patch
 
+import attr
 import pytest
 from openlineage.client import OpenLineageClient
 from openlineage.client.event_v2 import BaseEvent, Job, Run, RunEvent, RunState
+from openlineage.client.facet_v2 import external_query_run
 from openlineage.client.serde import Serde
 from openlineage.client.transport.transform import EventTransformer, TransformConfig, TransformTransport
 from openlineage.client.uuid import generate_new_uuid
@@ -91,7 +94,7 @@ def test_transform_config_from_dict_minimal_config() -> None:
     )
     assert config.transport == {"type": "console"}
     assert config.transformer_class == "some.class.to.be.Imported"
-    assert config.transformer_properties == {}
+    assert config.transformer_properties is None
 
 
 def test_transform_config_from_dict_missing_config() -> None:
@@ -205,6 +208,90 @@ def test_client_with_transform_transport_emits_modified_event(mocker: MockerFixt
     )
 
 
+def test_client_with_transform_transport_emits_modified_event_with_older_facets(
+    mocker: MockerFixture,
+) -> None:
+    with warnings.catch_warnings():
+        warnings.simplefilter("ignore", DeprecationWarning)
+        from openlineage.client.facet import BaseFacet as OldBaseFacet
+
+        @attr.s
+        class SomeOldRunFacet(OldBaseFacet):
+            version: str = attr.ib()
+
+    session = mocker.patch("requests.Session")
+    config = TransformConfig.from_dict(
+        {
+            "transport": {
+                "type": "http",
+                "url": "http://backend:5000",
+                "session": session,
+            },
+            "transformer_class": "tests.transform.test_transform.SampleEventTransformer",
+        }
+    )
+    transport = TransformTransport(config)
+    client = OpenLineageClient(transport=transport)
+    now = datetime.datetime.now().isoformat()
+    run_id = str(generate_new_uuid())
+    event = RunEvent(
+        eventType=RunState.START,
+        eventTime=now,
+        run=Run(
+            runId=run_id,
+            facets={
+                "externalQuery": external_query_run.ExternalQueryRunFacet(
+                    externalQueryId="queryid", source="source"
+                ),
+                "old_facet": SomeOldRunFacet(version="2"),
+            },
+        ),
+        job=Job(namespace="http", name="test"),
+        producer="prod",
+    )
+    modified_event = RunEvent(
+        eventType=RunState.START,
+        eventTime=now,
+        run=Run(
+            runId=run_id,
+            facets={
+                "externalQuery": external_query_run.ExternalQueryRunFacet(
+                    externalQueryId="queryid", source="source"
+                ),
+                "old_facet": SomeOldRunFacet(version="2"),
+            },
+        ),
+        job=Job(namespace="new_value", name="test"),
+        producer="prod",
+    )
+
+    client.emit(event)
+    transport.transport.session.post.assert_called_once_with(
+        url="http://backend:5000/api/v1/lineage",
+        data=Serde.to_json(modified_event),
+        headers={"Content-Type": "application/json"},
+        timeout=5.0,
+        verify=True,
+    )
+
+    # Assert the original event is unchanged
+    assert event == RunEvent(
+        eventType=RunState.START,
+        eventTime=now,
+        run=Run(
+            runId=run_id,
+            facets={
+                "externalQuery": external_query_run.ExternalQueryRunFacet(
+                    externalQueryId="queryid", source="source"
+                ),
+                "old_facet": SomeOldRunFacet(version="2"),
+            },
+        ),
+        job=Job(namespace="http", name="test"),
+        producer="prod",
+    )
+
+
 def test_client_with_transform_transport_skips_emission_when_transformed_event_is_none(
     mocker: MockerFixture,
 ) -> None:

openlineage_python-1.33.0/tests/test_uuid.py

@@ -1,54 +0,0 @@
-# Copyright 2018-2025 contributors to the OpenLineage project
-# SPDX-License-Identifier: Apache-2.0
-from __future__ import annotations
-
-from datetime import datetime, timedelta
-
-import pytest
-from openlineage.client.uuid import generate_new_uuid, generate_static_uuid
-
-
-def test_generate_new_uuid_returns_uuidv7():
-    assert generate_new_uuid().version == 7  # noqa: PLR2004
-
-
-def test_generate_new_uuid_returns_different_result_on_each_call():
-    result1 = generate_new_uuid()
-    result2 = generate_new_uuid()
-    assert result1 != result2
-
-    instant = datetime.now()
-    result1 = generate_new_uuid(instant)
-    result2 = generate_new_uuid(instant)
-    assert result1 != result2
-
-
-def test_generate_new_uuid_result_is_increasing_with_instant_increment():
-    result1 = generate_new_uuid(datetime.now())
-    result2 = generate_new_uuid(datetime.now() + timedelta(seconds=0.001))
-    assert result1 != result2
-    assert result1 < result2
-
-
-def test_generate_static_uuid_returns_uuidv7():
-    assert generate_static_uuid(datetime.now(), b"some").version == 7  # noqa: PLR2004
-
-
-def test_generate_static_uuid_returns_same_result_for_same_input():
-    instant = datetime.now()
-    data = b"some"
-    assert generate_static_uuid(instant, data) == generate_static_uuid(instant, data)
-
-
-@pytest.mark.parametrize(
-    ("data1", "data2"),
-    [
-        (b"some", b"other"),
-        (b"some", b"some"),
-    ],
-)
-def test_generate_static_uuid_result_is_increasing_with_instant_increment(data1, data2):
-    result1 = generate_static_uuid(datetime.now(), data1)
-    result2 = generate_static_uuid(datetime.now() + timedelta(seconds=0.001), data2)
-    assert result1 != result2
-    assert result1 < result2