release-0.3.1

Features

2. Flink job plugin supports the multiple Flink versions #236
3. Support stopping/resuming job scheduling #241
4. Support log view of job execution for frontend #251
5. Improve documentation

Bug Fixes

1. Airflow operator context does not set correctly #250
2. Failed to trigger workflow execution #260
3. Fix update the deployed model version multiple times #269
4. Fix register_model_version sending wrong event type bug #290

Welcome to use this release version and give us the feedback of the AIFlow.

release-0.3.0

Features

AIFlow

1. Introduces the command-line interface to help operation.
2. Support database version migration.
3. Supports the worflow development on the Jupyter Notebook #140

Notification Service

1. Introduces the command-line interface to help operation.
2. Support database version migration.

Bug Fixes

1. The remote logging supports the current log mechanism #172
2. Unittest failed cause by init_ai_flow_context #185
3. AIFlow Webserver cannot sort model version by version #207

Welcome to use this release version and give us the feedback of the AIFlow.

release-0.2.2

Features

AIFlow

1. Add the documents of the AIFlow. #117

Airflow

1. Allow LocalExecutor to run with SQLite. #41
2. Airflow webserver supports the Airflow and Notification databases. #49

Notification Service

1. Introduces the countEvents interface. #11

Bug Fixes

1. Fix oss blob manager download concurrently. #3
2. Deepcopy tasks in celery executor to avoid race condition. #12
3. Fix periodic workflow cannot run. #77
4. Fix HDFSBlobManager failed to download existed file. #87
5. Removes the uncompleted api action_on_dataset_event. #104
6. The workflow directory is set incorrect in AIFlow runtime. #111

Welcome to use this release version and give us the feedback of the AIFlow.

release-0.2.1

[Flink AI Flow] Bump version to 0.2.1 (#650)

release-0.2.0

Features

AIFlow

1. Add workflow execution on event and ContextExtractor API #476
2. Add task execution restful api #478
3. AIFlow add WorkflowEventManager to listen and handle events #492
4. Support start new workflow execution with context #479
5. Introduce the workflow frontend of the AIFlow UI #509
6. Add FlinkSqlProcessor #527
7. Support job execution label #529
8. Frontend support metadata ui #533
9. Notification service supports the idempotence #553
10. Add read-only job plugin #555

Airflow

1. Support celery executor on event based scheduler #482
2. Add AirFlowScheduler with airflow restful api #486

Bug Fixes

1. Make AI Flow be able to use Notification Service with HA enabled #510
2. Duplicated entry when create dagrun #570
3. EventBaseScheduler catches and prints exceptions #586
4. Scheduler should find schedulable tasks once dagrun finished #587
5. EventBaseScheduler would trigger task multiple times incorrectly #598

release-0.1.0

Features

1. Define the machine learning workflow including batch/stream jobs.
2. Manage metadata(generated by the machine learning workflow) of data sets, models, artifacts, metrics, jobs, etc.
3. Run the machine learning workflow.
4. Publish and subscribe events.