This is an automated email from the ASF dual-hosted git repository.

potiuk pushed a commit to branch main
in repository https://gitbox.apache.org/repos/asf/airflow.git


The following commit(s) were added to refs/heads/main by this push:
     new 1a3f32aa9e3 Add OSSRemoteLogIO.from_config and register oss remote 
logging scheme (#70682)
1a3f32aa9e3 is described below

commit 1a3f32aa9e3ae205bef2d6cecb758550dbdb396f
Author: Baha Bouali <[email protected]>
AuthorDate: Sat Aug 1 06:54:55 2026 +0100

    Add OSSRemoteLogIO.from_config and register oss remote logging scheme 
(#70682)
    
    Signed-off-by: baha-bouali <[email protected]>
---
 providers/alibaba/provider.yaml                    |  4 ++
 .../alibaba/cloud/log/oss_task_handler.py          | 26 +++++++
 .../airflow/providers/alibaba/get_provider_info.py |  6 ++
 .../alibaba/cloud/log/test_oss_task_handler.py     | 83 +++++++++++++++++++++-
 4 files changed, 118 insertions(+), 1 deletion(-)

diff --git a/providers/alibaba/provider.yaml b/providers/alibaba/provider.yaml
index acc7bfa8a4a..988bf96a269 100644
--- a/providers/alibaba/provider.yaml
+++ b/providers/alibaba/provider.yaml
@@ -193,5 +193,9 @@ connection-types:
 logging:
   - airflow.providers.alibaba.cloud.log.oss_task_handler.OSSTaskHandler
 
+remote-logging:
+  - classpath: 
airflow.providers.alibaba.cloud.log.oss_task_handler.OSSRemoteLogIO
+    scheme: oss
+
 extra-links:
   - airflow.providers.alibaba.cloud.links.maxcompute.MaxComputeLogViewLink
diff --git 
a/providers/alibaba/src/airflow/providers/alibaba/cloud/log/oss_task_handler.py 
b/providers/alibaba/src/airflow/providers/alibaba/cloud/log/oss_task_handler.py
index a000c6ff01f..d679f5cf251 100644
--- 
a/providers/alibaba/src/airflow/providers/alibaba/cloud/log/oss_task_handler.py
+++ 
b/providers/alibaba/src/airflow/providers/alibaba/cloud/log/oss_task_handler.py
@@ -18,6 +18,7 @@
 from __future__ import annotations
 
 import contextlib
+import inspect
 import os
 import shutil
 from functools import cached_property
@@ -44,6 +45,31 @@ class OSSRemoteLogIO(LoggingMixin):  # noqa: D101
 
     processors = ()
 
+    @classmethod
+    def from_config(cls) -> OSSRemoteLogIO:
+        """Build the remote log IO from Airflow logging configuration."""
+        remote_task_handler_kwargs = conf.getjson("logging", 
"remote_task_handler_kwargs", fallback={})
+        if not isinstance(remote_task_handler_kwargs, dict):
+            raise ValueError(
+                "logging/remote_task_handler_kwargs must be a JSON object (a 
python dict), we got "
+                f"{type(remote_task_handler_kwargs)}"
+            )
+        # remote_task_handler_kwargs mixes FileTaskHandler kwargs with IO 
kwargs; only the
+        #  latter belongs to this (same split as airflow_local_settings.py).
+        fth_params = 
frozenset(inspect.signature(FileTaskHandler.__init__).parameters) - {
+            "self",
+            "base_log_folder",
+        }
+        io_kwargs = {k: v for k, v in remote_task_handler_kwargs.items() if k 
not in fth_params}
+        return cls(
+            **{
+                "base_log_folder": 
os.path.expanduser(conf.get_mandatory_value("logging", "base_log_folder")),
+                "remote_base": conf.get_mandatory_value("logging", 
"remote_base_log_folder"),
+                "delete_local_copy": conf.getboolean("logging", 
"delete_local_logs"),
+            }
+            | io_kwargs,
+        )
+
     def upload(self, path: os.PathLike | str, ti: RuntimeTI | None = None) -> 
None:
         """Upload the given log path to the remote storage."""
         path = Path(path)
diff --git 
a/providers/alibaba/src/airflow/providers/alibaba/get_provider_info.py 
b/providers/alibaba/src/airflow/providers/alibaba/get_provider_info.py
index 27e01cd0e74..29cbc14aae5 100644
--- a/providers/alibaba/src/airflow/providers/alibaba/get_provider_info.py
+++ b/providers/alibaba/src/airflow/providers/alibaba/get_provider_info.py
@@ -137,5 +137,11 @@ def get_provider_info():
             },
         ],
         "logging": 
["airflow.providers.alibaba.cloud.log.oss_task_handler.OSSTaskHandler"],
+        "remote-logging": [
+            {
+                "classpath": 
"airflow.providers.alibaba.cloud.log.oss_task_handler.OSSRemoteLogIO",
+                "scheme": "oss",
+            }
+        ],
         "extra-links": 
["airflow.providers.alibaba.cloud.links.maxcompute.MaxComputeLogViewLink"],
     }
diff --git 
a/providers/alibaba/tests/unit/alibaba/cloud/log/test_oss_task_handler.py 
b/providers/alibaba/tests/unit/alibaba/cloud/log/test_oss_task_handler.py
index 24131d44346..677f1c0a2a4 100644
--- a/providers/alibaba/tests/unit/alibaba/cloud/log/test_oss_task_handler.py
+++ b/providers/alibaba/tests/unit/alibaba/cloud/log/test_oss_task_handler.py
@@ -18,13 +18,14 @@
 from __future__ import annotations
 
 import os
+from pathlib import Path
 from typing import TYPE_CHECKING
 from unittest import mock
 from unittest.mock import PropertyMock
 
 import pytest
 
-from airflow.providers.alibaba.cloud.log.oss_task_handler import OSSTaskHandler
+from airflow.providers.alibaba.cloud.log.oss_task_handler import 
OSSRemoteLogIO, OSSTaskHandler
 from airflow.utils.state import TaskInstanceState
 from airflow.utils.timezone import datetime
 
@@ -45,6 +46,86 @@ MOCK_CONTENT = "mock_content"
 MOCK_FILE_PATH = "mock_file_path"
 
 
+class TestOSSRemoteLogIOFromConfig:
+    @conf_vars(
+        {
+            ("logging", "base_log_folder"): "~/airflow/logs",
+            ("logging", "remote_base_log_folder"): 
"oss://bucket/remote/log/location",
+            ("logging", "delete_local_logs"): "True",
+        }
+    )
+    def test_from_config(self):
+        subject = OSSRemoteLogIO.from_config()
+
+        assert subject.remote_base == "oss://bucket/remote/log/location"
+        assert subject.base_log_folder == 
Path(os.path.expanduser("~/airflow/logs"))
+        assert subject.delete_local_copy is True
+
+    @conf_vars(
+        {
+            ("logging", "base_log_folder"): "/tmp/airflow/logs",
+            ("logging", "remote_base_log_folder"): 
"oss://bucket/remote/log/location",
+            ("logging", "delete_local_logs"): "False",
+            ("logging", "remote_task_handler_kwargs"): '{"delete_local_copy": 
true, "max_bytes": 1024}',
+        }
+    )
+    def 
test_from_config_applies_io_kwargs_and_filters_file_handler_kwargs(self):
+        subject = OSSRemoteLogIO.from_config()
+
+        assert subject.delete_local_copy is True
+        assert not hasattr(subject, "max_bytes")
+
+    @conf_vars({("logging", "remote_task_handler_kwargs"): '["not", "a", 
"dict"]'})
+    def test_from_config_rejects_non_dict_remote_task_handler_kwargs(self):
+        with pytest.raises(ValueError, match="remote_task_handler_kwargs"):
+            OSSRemoteLogIO.from_config()
+
+    def test_provider_registers_oss_scheme(self):
+        from airflow.providers_manager import ProvidersManager
+
+        manager = ProvidersManager()
+        if not hasattr(manager, "remote_logging_handler_by_scheme"):
+            pytest.skip("Airflow core does not support remote logging provider 
dispatch")
+
+        info = manager.remote_logging_handler_by_scheme("oss")
+
+        assert info is not None
+        assert info.classpath == 
"airflow.providers.alibaba.cloud.log.oss_task_handler.OSSRemoteLogIO"
+
+    @pytest.mark.parametrize(
+        "manager_classpath",
+        [
+            pytest.param("airflow.providers_manager.ProvidersManager", 
id="core"),
+            pytest.param(
+                
"airflow.sdk.providers_manager_runtime.ProvidersManagerTaskRuntime", 
id="task-runtime"
+            ),
+        ],
+    )
+    @conf_vars(
+        {
+            ("logging", "remote_logging"): "True",
+            ("logging", "remote_base_log_folder"): 
"oss://bucket/remote/log/location",
+            ("logging", "remote_log_conn_id"): "oss_default",
+        }
+    )
+    def 
test_resolve_remote_task_log_uses_provider_dispatch_not_local_settings(self, 
manager_classpath):
+        factory = pytest.importorskip("airflow._shared.logging.factory")
+        from airflow._shared.module_loading import import_string
+        from airflow.configuration import conf
+
+        with mock.patch.object(factory, "discover_remote_log_handler", 
autospec=True) as legacy_discover:
+            remote_task_log, conn_id = factory.resolve_remote_task_log(
+                conf=conf,
+                providers_manager=import_string(manager_classpath)(),
+                import_string=import_string,
+            )
+
+        assert isinstance(remote_task_log, OSSRemoteLogIO)
+        assert remote_task_log.remote_base == 
"oss://bucket/remote/log/location"
+        assert conn_id == "oss_default"
+        legacy_discover.assert_not_called()
+
+
 class TestOSSTaskHandler:
     def setup_method(self):
         self.base_log_folder = "local/airflow/logs/1.log"

Reply via email to