[
https://issues.apache.org/jira/browse/AIRFLOW-1359?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16107911#comment-16107911
]
ASF subversion and git services commented on AIRFLOW-1359:
----------------------------------------------------------
Commit 1932ccc881d10d220d1d06efaa477373a08596bb in incubator-airflow's branch
refs/heads/master from [~pdwryst]
[ https://git-wip-us.apache.org/repos/asf?p=incubator-airflow.git;h=1932ccc ]
[AIRFLOW-1359] Use default_args in Cloud ML eval
This change makes the create_evaluate_ops utility
method make use of the default_args parameters of
the DAG when possible. This simplifies the usage
of the create_evaluate_ops method, and improves
the usefulness of a variety of default_args.
To further the usefulness of default_args for
Cloud ML Operators, this change also introduces
version_name to the CloudMLVersionOperator,
allowing model_name and version_name to be
specified across an entire pipeline.
This change also resolves a small TODO by making
the DataFlowPythonOperator's `options` and
`dataflow_default_options` variables templatized.
Closes #2445 from
peterjdolan/eval_ops_arguments_from_default_args
> Provide GoogleCloudML operator for model evaluation via Cloud APIs
> ------------------------------------------------------------------
>
> Key: AIRFLOW-1359
> URL: https://issues.apache.org/jira/browse/AIRFLOW-1359
> Project: Apache Airflow
> Issue Type: New Feature
> Components: gcp
> Reporter: Younghee Kwon
> Assignee: Younghee Kwon
> Fix For: 1.9.0
>
>
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)
