[
https://issues.apache.org/jira/browse/COMDEV-462?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Jiajie Zhong updated COMDEV-462:
--------------------------------
difficulty-level: Medium
> GSoC: Python API CLI enhancement
> --------------------------------
>
> Key: COMDEV-462
> URL: https://issues.apache.org/jira/browse/COMDEV-462
> Project: Community Development
> Issue Type: Task
> Components: GSoC/Mentoring ideas
> Reporter: Jiajie Zhong
> Priority: Major
> Labels: DolphinScheduler, full-time, gsoc2022
>
> h1. About pydolphinscheduler
> *PyDolphinScheduler* is Python API for [Apache
> DolphinScheduler|https://dolphinscheduler.apache.org/], which allows you to
> define your workflow by Python code, aka workflow-as-codes. You could see
> more detail about PyDolphinScheduler in its document[4]. And all the source
> code hold as the submodule in DolphinScheduler main codebase[5].
> h1. The goal
> Make pydolphinscheduler's CLI more powerful, make it can operate the model of
> DolphinScheduler, run pydolphinscheduler's code, visualize its DAG graph in
> the terminal.
> h1. Detail
> Up to now, Apache DolphinScheduler Python API has CLI only with limited
> command supported and our community wishes it to become a more powerful tool
> and support as much command as possible(unless command has security issue).
> It only supports `version` and `config` for now, which you could see more
> detail in [1]
> Basically, we think the following command is helpful for CLI and you could
> add another command if it should be added(but may sure after discussing in
> the community):
> * `run <DAG name> [--example]`: Run local workflow DAG file or examples
> build-in
> * `users`: User's operation, CURD
> * `projects`: Project's operation, CURD, grant to other users
> * `tenants`: Tenant's operation, CURD
> * `workflow`: Workflow's operation, CURD, name change, should also change
> the local Python file name
> * `visualize`: Show task graph in the terminal.
> * etc...
> Besides the functional addition, we should also consider the output part of
> CLI which makes our output more clear and cool. We *may* consider using (we
> should also find other interesting packages to do it):
> * rich: For highlight, our output, or using some existing rich plugin like
> `click-rich`
> * tabulate: For the tables visualization in terminal
> h1. What Can You Learn
> We wish everyone join GSoC could learn some things from the project. When you
> finish this project, you could lean:
> * How to write production-level Python codes and docs, you could improve
> your Python syntax, how to write tests with `pytest` and `tox`, how to write
> document with `sphnix` and it related plugin, how to format your Python code
> and the linter inside
> * Adding knowledge about task scheduling system, what is it and what it
> focus, how it could be run
> h1. If You Interested in It
> If you want to take this ticket, you should
> * (Must) Python skill, especially packages click, pytest and etc.
> * Have a little knowledge of task scheduling systems.
> * (Optional) Basic Java knowledge is better because Apache DolphinScheduler
> core is written with Java and you may add some functional code to it.
>
> [1]: [https://dolphinscheduler.apache.org/python/cli.html]
> [2]: [https://github.com/Textualize/rich]
> [3]: [https://github.com/astanin/python-tabulate]
> [4]: [https://dolphinscheduler.apache.org/python/index.html]
> [5]:
> [https://github.com/apache/dolphinscheduler/tree/dev/dolphinscheduler-python/pydolphinscheduler]
--
This message was sent by Atlassian Jira
(v8.20.1#820001)
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscr...@community.apache.org
For additional commands, e-mail: dev-h...@community.apache.org
